Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females
(USC Thesis Other)
Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DETERMINING COMPETENT/RESILIENT OUTCOMES ACROSS MULTIPLE DOMAINS IN SEXUALLY ABUSED AND NONABUSED FEMALES by Dawn A. Kurtz A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (PSYCHOLOGY) May 2003 Copyright 2003 Dawn A. Kurtz Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3103923 Copyright 2003 by Kurtz, Dawn A. All rights reserved. ® UMI UMI Microform 3103923 Copyright 2003 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90089-1695 This dissertation, written by P<twn f t under the direction o f h dissertation committee, and approved by all its members, has been presented to and accepted by the Director o f Graduate and Professional Programs, in partial fulfillment o f the requirements for the degree of DOCTOR OF PHILOSOPHY Director Date May 1 6 , 2003 Dissertation Committee Chair . . A . . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ii Table of Contents List of Tables iii List of Figures v Abstract vi Chapter 1: Introduction 1 Chapter 2: The Present Study 42 Chapter 3: Method 45 Chapter 4: Results 53 Chapter 5: Discussion 69 References 85 Appendix: Intercoirelation tables 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Tables Table 1. Studies examining resilience to maltreatment Table 2. Outcomes used to assess resilience in prior studies of maltreatment Table 3. Means and standard deviations for demographic variables for abused an nonabused participants Table 4. Means, standard deviations, cutoff scores, and percent of participants scoring in the competent range for entire sample Table 5. Means, standard deviations, and percent of participants scoring in the competent range Table 6. Total number of domains in which participants received scores in the competent/resilient range Table 7. Cross tabulation of competent/resilient classifications for social competence and psychological competence Table 8. Cross tabulation of competent/resilient classifications for academic competence and psychological competence Table 9. Cross tabulation of competent/resilient classifications for academic competence and social competence Table 10. Classification results from the Discriminant Function Analysis Table 11. Results of multiple regression analyses using social network variables, mother-daughter closeness, and cognitive ability as predictors of domains o f competence Table 12. Significance tests of simple slopes o f regression lines predicting social competence from number of males in social network and academic competence from Time 1 PPVT scores at conditional values of abuse status Table 13. Results of multiple regression analyses using characteristics of abuse as predictors of domains of competence iii 16 31 47 55 55 56 57 58 58 59 61 65 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 14. Characteristics of abuse for resilient and nonresilient sexually abused participants in each domain o f competence Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. V List of Figures Figure 1. Regression lines predicting social competence from males in 63 social network at conditional values of abuse status Figure 2, Regression lines predicting academic competence from Time 64 1 PPVT scores at conditional values o f abuse status Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vi ABSTRACT The current study examines competent/resilient outcomes across three domains of functioning in 162 (87 sexually abused; 75 nonabused) females. Data are taken from a longitudinal study of the psychobiological impact of child sexual abuse on female development. At the outset of the study the participants met the following criteria: they ranged in age from 6 to 16 (M=l 1; SD=3); abuse disclosure had occurred within six months of participation; abuse involved genital contact and/or penetration; the perpetrator was a family member; and a non-abusing parent or guardian was willing to participate in the study. Data used for the current analyses were obtained during the first and fifth assessments. At the fifth data collection the mean age of the participants was 20.68 years. At the initial assessment participants reported on social support and social network variables and completed a measure of cognitive ability. For the sexually abused participants, information about the abuse experience was also obtained at Time 1. At the fifth assessment, participants completed measures of depressive symptomatology, social competence and academic ability. Data were examined two ways: first using cutoff scores to define resilient functioning and then using continuous variables. Results from this investigation revealed that sexually abused and nonabused females were equally competent on each of the three domains of functioning when cutoff scores were used to determine competent functioning. However, competent/resilient classifications varied from one domain to another. Participants classified as resilient in one domain were not Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vii necessarily classified as resilient in other domains. Results of a Discriminant Function Analysis indicated that the strongest predictor of overall resilience (across all three domains, using cutoff scores) was cognitive ability. There were two significant interaction effects when competent outcomes were examined using continuous variables, indicating that abuse status acts as a moderator in some circumstances. Specifically, results indicated a significant interaction effect for number of males in one’s social network by abuse status when predicting social competence. Additionally, there was a significant interaction of Time 1 cognitive ability scores by abuse status on academic competence. Findings are considered in relation to the existing body of research on resilience and implications for future research are discussed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 Chapter 1: Introduction The present study examines resilient outcomes in three distinct domains for a group of sexually abused individuals. One of the major aims of this study is to determine whether individuals who are classified as competent/resilient in one domain are also classified as competent/resilient in other domains. For the purposes of this investigation, resilience refers to successful adaptation, reflected by psychological competence, social competence, and academic achievement, despite exposure to childhood sexual abuse. This study adds to the existing body of research on resilient outcomes for sexually abused individuals in that few studies have previously examined whether resilience is, in fact, domain specific. Most studies that have examined resilience in sexually abused samples have classified participants based on one specific domain. Those studies that have looked at resilience across domains have yielded mixed findings. More specifically, individuals found to be resilient in one domain have not necessarily been resilient in other domains. Another problem is that all of the existing research on resilient outcomes for sexually abused individuals has consisted of cross-sectional studies. The data for this study comes from a longitudinal study that has followed the development of sexually abused females across 9 years and includes a sample of comparison females from similar demographic backgrounds. There are three specific objectives for this study. First, this study investigates resilient outcomes for sexually abused females utilizing three distinct domains— psychological adjustment, social competence, and academic achievement; and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 examines which of the three domains results In Identification of the largest number of resilient participants. Another goal of the current research is to determine whether individuals who are classified as resilient in one domain are also classified as resilient in other domains. Finally, this study seeks to determine whether factors assessed shortly after the abuse experience can be used to predict resilient functioning 9 years later. Research on child abuse and maltreatment Child abuse has been a significant problem in our society for several decades. Several forms of child abuse have been studied in the literature, including neglect, physical abuse, emotional abuse, and sexual abuse. Although research in recent history has provided some insight into the complexities and consequences of child abuse and neglect, there is still much to be discovered, especially pertaining to outcomes related specifically to sexual abuse. In 1996, child protective services agencies investigated more than 2 million reports regarding the maltreatment of more than 3 million children; one million of these cases were substantiated (NCCAN, 1996). Of these, nearly 220,000 children are recognized by public protective service agencies each year as having experienced significant “harm” via sexual abuse, while another 100,000 are recognized as “endangered” of being sexually abused (Noll, Trickett & Putnam, 2001). Sexual abuse appears to be a unique type of abuse that warrants examination in and of itself. A substantial proportion of people growing up in the United States experience what they, and/or authorities, describe as inappropriate, unwanted, or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 disturbing sexual experiences imposed by another during childhood (Noll, et ah, 2001). Furthermore, some researchers suggest that because of its explicit sexual nature, the impact of childhood sexual abuse is likely to differ from other forms of child maltreatment (Wyatt & Powell, 1988). Despite this, studies of outcomes specifically related to sexual abuse are rare. Instead sexual abuse is often included as one type of abuse in more general studies of child maltreatment (e.g., Moran & Eckenrode, 1992). Other studies examine the impact of sexual abuse without giving specific details about the nature of the abuse (Feinauer & Stuart, 1996). Research in this area has been hindered by several methodological problems. Specifically, many studies are retrospective, do not include meaningful comparison groups, and have small sample sizes. In addition, variability in definitions of abuse and variability in the abuse experience itself (i.e. variability in characteristics of the abuse) make comparisons from one study to the next difficult. A brief discussion of these methodological complexities follows. Methodological complexities in studying the impact of child abuse Use o f comparison groups. Many researchers suggest that inconsistent findings regarding the impact of childhood sexual abuse can be attributed, at least in part, to the fact that numerous studies fail to include comparison groups as part of their investigations. Furthermore, studies that have included comparison groups often choose comparison group members from clinical settings instead of community [non-clinical] settings. Perkins & Luster (1999) note the over-reliance of studies of the relationship between sexual abuse and eating disorders on clinical samples of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 girls who are actively involved in treatment for eating disorders or sexual abuse. It stands to reason that patients who are in treatment are not likely to be representative of all patients who have been sexually abused. While sexual abuse may have a negative impact on development, it is difficult to interpret findings from studies that use individuals from clinical populations as their controls. Variability in definitions o f abuse and maltreatment. Previous studies on the impact of child maltreatment have been complicated by inconsistent definitions and classifications of types of abuse (Trickett & McBride-Chang, 1995). Inconsistency in definitions of abuse makes it difficult to understand the child’s experience. It is often unclear whether an individual has experienced one specific type of abuse, or several types. Moreover, researchers note that criteria for determination of the occurrence of child abuse and neglect are often questionable, widely varied, and may include unsubstantiated cases. In fact, most studies do not distinguish between abuse and neglect, which may obscure important behavioral and/ or psychological distinctions (Widom, 1989). Furthermore, whereas each type of maltreatment poses a risk to a child’s functioning, there may be differential effects depending on the type (Heller, Larrieu, D’Imperio & Boris, 1999). Compounding the problem is the fact that various types of abuse often occur simultaneously, thus making it nearly impossible to report the individual effects of one type of abuse (Zingraff, Leiter, Myers & Johnson, 1993). Often studies that do differentiate between types of abuse (physical, neglect, sexual abuse and emotional maltreatment) later combine all of the participants and refer to them collectively as maltreated (Kaufman et al, 1994). In Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 addition to varying definitions of abuse, studies vary widely in the population from which they draw their subjects. Whereas some abuse groups include only children who have been abused by family members (Lipovsky, Saunders, & Murphy, 1989), others use mixed samples which include victims of sexual abuse perpetrated by family members as well as strangers, acquaintances, etc. (Mannarino, Cohen & Gregor, 1989). Characteristics o f abuse. Much of the extant research on child sexual abuse (CSA) has included investigations into the impact specific characteristics of the abuse experience have on outcomes. Aspects of sexual abuse which have been studied in terms of their impact on child adaptation include type of abuse, frequency, chronicity, and severity of maltreatment (Egeland, 1997). Several studies support the theory that the relationship of the perpetrator to the sexual abuse victim has a serious impact on the victim’s perception of the experience. Many researchers posit that abuse by someone close to the victim is more traumatic than abuse by a stranger or an acquaintance (Browne & Finkelhor, 1986; Sides, Adams-Tucker, 1982; McLeer, Deblinger, Atkins, Foa & Ralphae, 1988; Russell, Schurman, & Trocki, 1988). Likewise, numerous studies of type of abuse report that penetration has a more detrimental impact than less severe form of abuse (Bagley & Ramsey, 1986; Russell, 1986; Mennen & Meadow, 1995; Morrow & Sorrell, 1989). Others suggest that compound abuse (e.g. sexual abuse + physical abuse) results in more negative repercussions than sexual abuse that is not accompanied by another type of abuse (Kamsner & McCabe, 2000; Mullen, Martin, Anderson, Romans & Herbison, 1996). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 The findings regarding age at onset of abuse have been mixed. While some studies have reported a significant relationship between age at onset of abuse and the impact of the abuse experience (McClellan, McCrary, Ronnei, Adams, Eisner & Stork, 1996; Nash, Zivney, & Hulsey, 1993) still others have found no relationship between the two (Einbender & Friedrich, 1989; Russell, 1986). The methodological complexities of studying the impact of abuse add to other challenges present in resilience research, making inquiry into resilient outcomes for sexual abuse victims a difficult task. Although the major purpose of the current study is to explore competent/resilient outcomes in sexually abused girls, it is important to first review more general findings of the developmental impact of sexual abuse. Thus a general discussion of the existing research on the relationship between child abuse and psychological, social, and academic competence will precede a more specific discussion of competent/resilient outcomes for victims of child abuse and maltreatment. Research on the impact of sexual abuse Several recent reviews have documented the deleterious impact that childhood sexual abuse (CSA) can have on development (see Browne & Finkelhor, 1986; Beitchman et a!., 1992). Often included in these reviews are summaries of the relationship between childhood sexual abuse and depression, social competence and academic performance. Much o f the research done on these relationships has been marred by the aforementioned problems in the sexual abuse literature. Namely, many of the studies lack meaningful comparison groups, and either explore the short Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 term impact of abuse or are retrospective in nature. Only studies that include comparison groups will be considered here, given that they are most relevant to the design of the present study. Psychological competence. Poor psychological functioning has been one of the most commonly studied consequences of childhood sexual abuse however findings have been mixed. Using nonabused siblings as their comparison group Lipovsky, Saunders & Murphy (1989) explored the impact of CSA perpetrated by a father or other adult male caretaker. Their results indicated that abused children were significantly more depressed (i.e. in the clinical range on the CDI) than their nonabused siblings. Interpretation of these results is somewhat difficult given that the children were from the same households and therefore while one sibling was not sexually abused there are likely other environmental and/or household experiences that impact the development of both children. Mannariito, Cohen & Gregor (1989) compared sexually abused girls with two comparison groups: a clinical control group and a group of normal controls with similar demographic characteristics to the sexually abused girls. Their data did not reveal significant differences in depression between the groups. In two follow-ups (at six and twelve-months) of the same three groups, there were still no significant differences in depression (Mannarino, Cohen, Smith & Moore-Motily, 1991). McLeer, Dixon, & Henry (1998) compared non-clinically referred sexually abused children with nonabused clinically referred children as well as a group of non- clinically referred school children. Results of this study revealed no significant Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 differences between the nonclinicaily referred sexually abused children and the clinically referred nonabused children. However, there were significant differences between the sexually abused children and the nonabused children, with the sexually abused group scoring higher on the CDI Similar findings have come out of retrospective studies of sexual abuse using adult samples. Results of several retrospective studies have indicated a higher incidence of depression in adult women who report having been abused during childhood (Sedney & Brooks, 1984: Hunter, 1991; Mullen, et a!., 1996; Zuravin & Fontanella, 1999; Whiffen, Thompson, & Aube, 2000; Kamsner & McCabe, 2000). For example, Zuravin & Fontanella (1999) found that the rate of depression among the sexually abused women in their sample was nearly three times greater than the rate among the nonsexually abused women (45.7% vs. 16.4%, respectively). Whiffen et al. (2000) also found a significant relationship between CSA and depressive symptoms in abused participants. This finding held true for both men and women. Social competence. Social competence has not been studied as extensively as depression in terms of its relationship to CSA. Furthermore, social competence is often examined under the larger umbrella of interpersonal or adjustment problems. Many studies look at a compilation of interpersonal problems including variables such as intimacy problems and submissiveness (Whiffen, et al., 2000), negative self- regard and dissatisfaction in social situations (Abdulrehman & De Luca, 2001). Friedrich, Beilke & Urquiza (1988) examined the relationship between CSA and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 social competence in sexually abused nonabused children, as well as a clinical control group. Results indicated that sexually abused children were less socially competent than nonsexually abused children. However, the sexually abused children in this sample were more socially competent than the participants from the clinical comparison group. In their retrospective study of the relationship between CSA and adult social behavior, Abdulrehman & De Luca (2001) found that college students who reported a history of CSA scored significantly higher on overall social dysfunction and, more specifically, reported more dissatisfaction in social situations than participants with no history of CSA. Along a similar vein, Harter, Alexander & Neimeyer (1988) report that college females who reported having been sexually abused during childhood scored significantly poorer on social adjustment compared to their nonabused counterparts. However the relationship did not hold once family and cognitive characteristics were controlled for. Hunter (1991) compared adult males and females molested as children with a control group and found that males and females with a history of sexual abuse exhibited more dysfunction on a scale of relationship dysfunction than nonabused participants. Academic competence. Intellectual functioning and academic achievement appear to be among the least explored variables in terms of their relationship to childhood sexual abuse. Specifically, there is a paucity of research on the long-term consequences of CSA on intellectual and academic outcomes. There are, however, a few studies that have examined sexual abuse as one domain among children who Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 have experienced, child abuse and/or maltreatment (Perez & Widora, 1994; Rowe & Eckenrode, 1999). For example, Rowe & Eckenrode (1999) examined grade repetitions and poor English and Math grades as indicators of academic difficulties in maltreated and nonmaltreated children between the ages of 5-18. Participants in their study included children who had been neglected, physically abused and sexually abused. However, 56% of the participants in the maltreated group were victims of neglect whereas only 9.2% had been sexually abused. Type of abuse was not considered in their analyses due to the relatively small number of sexually abused participants. Results indicated a greater risk for maltreated children in repeating a grade and receiving poor English and Math grades. The few studies that specifically examine the link between sexual abuse and academic achievement have resulted in mixed findings, possibly due to the fact that they utilized different outcomes to assess achievement. Some have used standardized achievement tests to represent academic success (Einbender & Friedrich, 1989) whereas others have used parent or teacher reports of school performance (Tong, Oates, & McDowell, 1987; Erickson, Egeland & Pianta, 1989). Still others have used grades in school and grade retention in addition to achievement tests (Eckenrode, Laird & Doris, 1993). Using several outcome variables to assess academic success, Trickett, McBride-Chang, & Putnam (1994) found abuse status to be a significant predictor of academic performance in a group of sexually abused females between the ages of 6-16. Specifically, sexually abused girls were more likely to have lower ratings of classroom social competence, be more school Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11 avoidant and were less competent learners than their nonabused counterparts when controlling for ethnic minority status and age. In contrast, Himelein (1995) did not find a significant relationship between CSA and poor academic outcomes in her study examining three academic variables and their relationship to CSA in a sample of college women. The variables under investigation were student status, defined as either currently enrolled vs. no longer enrolled in school; overall grade point average; and total number of earned credit hours (only for enrolled students). The results did not indicate any significant group differences for the three academic variables. It should be noted that the data for this study came from retrospective accounts of sexual abuse. In one of the few studies of long-term intellectual and academic outcomes, Perez & Widom (1994) assessed IQ and reading ability in abused and neglected participants 20 years post-abuse, as well as in a matched comparison group. Their results indicated that more than 50% of the abused and/or neglected participants scored in the deficient range in reading ability, compared with 30% of the comparison participants. Abused and/or neglected participants also scored significantly lower than controls on IQ scores, even when age, race, sex and social class were held constant. In addition, abused and/or neglected children had completed significantly fewer years of school, reported being more truant, having repeated a grade, and having been suspended or expelled. It is important to note that many of the aforementioned studies examined the impact of maltreatment but did not focus specifically on sexual abuse. Furthermore, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 findings from these studies are clearly mixed and the majority of the studies that do suggest a link between maltreatment and negative outcomes in adulthood are not longitudinal and therefore do not examine the long-term relationship of maltreatment and negative outcomes over time. Instead the majority are retrospective and thus rely on an individual’s memory of past events. In fact, Zuravin & Fontanella (1999) admit that one of the limitations of their retrospective study is that reports of abuse may be biased due to selective recall and/or forgetting. They further suggest that people who are depressed may be more sensitive and therefore more likely to remember negative events than nondepressed women, which would lead to an overestimation in the relationship between CSA and depression. Perez & Widom (1994) assert that childhood maltreatment often occurs in the context of multi problem homes in which the maltreatment is only one of the problems occurring. The present study endeavors to combat this problem by using data from a prospective study of sexually abused and nonsexually abused women. Given the mixed findings in the extant sexual abuse literature it seems appropriate to focus on individuals who have experienced sexual abuse and yet are functioning well in adulthood. Such individuals are often referred to as being resilient. Research on resilience Research on resilient outcomes has proliferated in recent years. Resilience is typically defined in one of two ways; successful adaptation despite adversity (Masten & Coatsworth, 1998); or avoidance of undesirable outcomes in the face of adversity. One of the premises of resilience research holds that studying positive as well as Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 negative outcomes might reveal the processes that give rise to or avert psychopathology (Masten, 1999). Numerous risk factors have been studied, including: having a history of recent life stressors, having psychiatrically ill parents, parents of low status occupations, divorced parents, or an absent parent; being from a large family, being a minority; being a child with a history of perinatal stress; and suffering abuse, neglect, and/or poverty. Given that many individuals who are exposed to risk and/or adversity do not exhibit negative outcomes, it seems appropriate to examine factors that contribute to positive outcomes despite exposure to risk. This has traditionally been done by assessing characteristics of the individual (e.g. personality traits), as well as characteristics of the environment in which the child lives as possible mediators of risk exposure. Examples of characteristics of the individual that have been linked to successful adaptation include: high self-esteem (Cicchetti & Rogosch, 1997; Neighbors, Forehand, & Me Vicar, 1993; Moran & Eckenrode, 1992); ego resilience and ego control (Cicchetti & Rogosch, 1997); lack of self blame (Feinauer & Stuart, 1996; Liem, Janies, O'Toole & Boudewyn, 1997); adaptive coping styles (Spaccarelli & Kim, 1995; Tremblay, et al., 1999); internal locus of control (Himelein & McElrath, 1996; Liem, et al., 1997; Moran & Eckenrode, 1992); optimism (Himelein & McElrath, 1996); and intellectual capacity (Masten et al., 1999; Herrenkohl, et al., 1994; Lutfiar, 1991). Environmental factors include: parental support (Spaccarelli & Kim, 1995; Lynskey & Fergusson, 1997; Tremblay et al., 1999; Masten, et al., 1999); stable caretaking (Herrenkohl, et al., 1994); less Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 stressful family environments (Herman-Stahl, & Petersen, 1996; Liem, et al., 1997); social support (O’Grady & Metz, 1987) affiliation with non-deviant peers (Lynskey & Fergusson, 1997); and receipt o f therapy (Liem, et al., 1997). For the purposes of this study, emphasis will be placed on research that has examined resilient outcomes in people who have experienced child maltreatment. Research on resilient outcomes in victims o f child maltreatment As previously noted, child abuse is often thought to have severely deleterious effects on the majority of the individuals experiencing it. Past research has indicated a correlation between abuse and/or neglect and negative outcomes such as aggression, depression, social withdrawal, low academic performance, and psychopathology. What has often been overlooked is that studies that have found CSA to be associated with negative outcomes have also demonstrated that there is much variability in the impact of abuse. In fact, a recent review of the literature on consequences of child sexual abuse (CSA) revealed that approximately one-third of children with CSA histories show no sign of symptoms following abuse, and that a significant amount of the children who were symptomatic recovered quite quickly (Kendall-Tackett, et al., 1993). More specifically, it has been noted that even when an abuse group shows a significantly higher percentage of participants with specific problems or diagnoses compared with a comparison group, quite often the majority of both groups do not manifest that problem. For example, 40% of sexually abused girls in one study received a diagnosis of major depression, leaving 60% whom did not (Conte & Schuerman, 1987). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 Compared to research examining other risk factors, less research has been done regarding resilient outcomes for individuals with abuse and neglect experiences. In fact, an exhaustive review of the literature related to resilient outcomes for individuals exposed to maltreatment only resulted in identification of 14 studies (see Table 1). Three of the fourteen studies examined protective and mediating factors of abusive experiences without specific reference to the construct of resilience (Lynskey & Ferguson, 1997; Moran & Eckenrode, 1992; Tremblay, Hebert, & Piche, 1999) and one study examined resilience but never explicitly defined the term (Lam & Grossman, 1997). Thus, only ten of the fourteen studies were specifically aimed at investigating the construct of resilience. As can be seen in Table 1, eight of the fourteen studies of resilient outcomes for victims of maltreatment focused specifically on sexual abuse1 . Of the eight studies that examined sexual abuse, six utilized retrospective accounts of childhood sexual abuse (Feinauer & Stuart, 1996; Himelein & McElrath, 1996; Hyman & Williams, 2001; Lam & Grossman, 1997; Liem, et al. 1997; Lynskey & Fergusson, 1997). It should be noted that all of the participants in these six retrospective studies were adults. Only two studies were located which examined sexual abuse during childhood, and both used a cross-sectional design (Spaccarelli & Kim, 1995; 1 One additional study (Moran & Eckenrode, 1992) included sexually abused children in their sample but addressed sexual abuse as one type of abuse present in a group of maltreated children. Outcomes were not reported separately for the sexually abused children and as such this study was excluded from the subsequent discussion of studies that specifically examine resilient outcomes in sexually abused individuals. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1. Studies Examining Resilience to Maltreatment Study Definition of resilience Definition of Risk/Adversity Sample Study design Outcomes Cicchetti & Rogosch (1997); see also Cicchetti, Rogosch, Lynch, & Holt (1993) High adaptive functioning in four or more areas (out of a possible seven) Child MaltrealmenX- determined by DDS records 73% of maltreated children experienced multiple forms of maltreatment (mean = 2.11 types) N = 213 children attending a summer day camp for maltreated and other low- income, disadvantaged children (133 Maltreated; 80 nonmaltreated Age- mean age 8.03 for both groups Gender- 62% bovs Ethnicitv- 62% African- American, 13% Hispanic, 19% Caucasian SES- Mean Hollingshead = maltreated (21.09); nonmaltreated (22.53) Longitudinal Used a comparison group recruited from AFDC; no statistical differences except that there were slightly fewer minorities in maltreated group (76% vs. 89%) Maltreated children had fewer areas of adaptive functioning. Considerably more maltreated children represented at the low end (competent range), with many children exhibiting zero to one area of competence. Maltreated children exhibited significantly lower ego-resilience and lower intelligence. Farber & Egeland (1987) Competent functioning despite abusive experience(s) Competence measured by: Attachment (12 & 18 mos.); Autonomous functioning (24 Child Maltreatment included various forms of maltreatment Four groups: physically abusive (e.g. spanking, unprovoked, angry outbursts); hostile/verbally N= 129 (44 maltreatment group; 85 comparison group) A ge-12.18.24. & 42 months at each testing, respectively Gender- not given Ethnicitv- not given Longitudinal (infancy- preschool) Comparison group also chosen from at-risk sample, consisted of mothers who _ _ j Number of abused children who were competent decreased between 12 months and preschool Mothers of securely attached abused children reported receiving more emotional support than Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) months); self- awareness & socialization (48 months); peer relations & socialization (preschool) abusive (e.g. chronic, overly harsh criticism); psychologically unavailable (e.g. detached, emotionally uninvolved); neglectful (e.g. irresponsible in managing daily activities of child rearing) provided adequate care to their children mothers of anxiously attached infants More abused infants from two-parent families were securely attached (at 12 & 18 mos.) Feinauer & Stuart (1996) Current lack of trauma symptoms Sexual Abuse- no details given about nature of sexual abuse N= 276 Age- mean age 37 G ender-100% female Ethnicity- not reported SES- mean income $23,000 Retrospective Severity of abuse was related to level of functioning. Subjects who blamed self, fate, or both self and fate have much higher trauma symptom counts. Subjects were blamed the perpetrator experienced much less distress. Herrenkohl Herrenkohl , & Egolf (1994) Top 40% (N=88) on cognitive/academi c functioning, social functioning, and Child Abuse & Neglect children from families receiving services for abuse and neglect, not otherwise defined hfc 457 children in 1976; 345 at follow-up in 80-82 Age- 18 months to 6 years (time 1) (elementary school age) & 90-92 (aged 15-21). 105 abused, 86 Longitudinal Used non abused participants as comparison Graduation from high school was used to assess resilience (or relative success) at the adolescent assessment. Only 14 of 23 (61%) Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) emotional functioning were considered high functioning, bottom 40% (N=T16) considered low functioning, remainder (N=141) considered middle- functioning 25 of 88 high functioning also abused, so resilient neglected Gender- breakdown is not provided. SES- not reported group (N=154) had graduated from or were still attending high school. The 14 graduates were of average or above- average intelligence. Chronicity of abuse was related to success (all 6 of 23 resilient adolescents that lived chronically physically abusive homes dropped out o f school) Himelein & McElrath (1996). Healthy adjustment following a history o f CSA. Sexual Abuse sexual experience, involving contact, prior to age 15, with a perpetrator who was at least 5 years older N= 180 college freshman (45 reported CSA) at a Southeastern Public University G ender- 100% female Age- mean age 18.1 Ethnicity- 94% white SES- incomes ranged from below 15,000 (5%) to more than $60,000 (24%) Retrospective Includes a comparison group (the 129 non-sexually abused girls). There were no differences in age, family income, race, marital status for the two CSA victims and nonvictims didn’t differ in overall adjustment; greater perceptions of internal control and higher levels o f unrealistic optimism were strongly related to better adjustment Regardless of abuse history, believing in one’s personal ability Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) groups to control future events and believing that those events will be more positive than negative was highly adaptive (are cognitive illusions powerful enough to ameliorate negative effects of CSA?) Kaufman, Cook, Amy, Jones, & Pittinsky, (1994) Absence of pathology (using cutoff scores used in previous studies to discriminate between clinical and nonclincal populations). Child Maltreatment (no other information given about type of abuse, etc.) 59% Single-parent household; 43% welfare recipients IN N 56 elementary ( l st-6th grade) school children Age- Mean age 9 years, 7 months (7-12) Gender- 29 girls. 27 bovs Ethnicity- 58% (32) Caucasian, 14% (8) African-American, 28% (16) Hispanic SES- mean Hollingshead- 1.5 (43% o f the children’s families were welfare recipients) Cross sectional No comparison group Cutoff norms were used to determine pathology Wide variation for number o f children classified as resilient, when comparing individual measures within and across domains (range 21- 64%, depending on measure and informant). Hyman & Williams (2001) Functional competence in 5 spheres (mental health, physical health, interpersonal relationships, Child Sexual Abuse Sexual contact against child’s wishes, involving force or coercion, and perpetrated by a male 5 or more years older NM 136 Age- 18-31 G ender- 100% Female Ethnicity- 86 % African- American SES- not reported Retrospective No comparison group High school graduation was strongest predictor of resilience. Women who also experienced force Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) economic well being, adherence to community standards) than victim during abuse and those arrested as teenagers were less likely to be resilient. Highly resilient women were less likely to be abused by family, experience severe physical abuse, and live in out-of- home placement, but more likely to come from a stable family, and receive social support Lam & Grossman (1997) Not explicitly defined Participants divided into 4 groups based on high and low scores on the Child Abuse and Trauma Scale and the composite index of protective factors Child Sexual Abuse Endorsed having a traumatic sexual experience as a child or revealed that their relationship with parents involved a sexual experience N= 264 undergraduates (48 reported CSA) from a Northeastern, urban University Age- mean aee 18.8(17- 25, and one 46 year old) Gender- 100% female Ethnicitv- maioritv white SES- 2/3 estimated familv income over $40,000 Additional demographics- mostly single diverse in faith, 163 not employed Retrospective Comparison group were the 216 participants with no self- reported history o f CSA. There were no differences between the two groups except more Composite index of protective factors significantly correlated with depression, symptoms, and social adjustment (those scoring higher, better adjusted, scored lower on all three outcomes) Protective factors were more beneficial for participants with CSA. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) 17 o f 44 abused women had higher levels of protective factors (above median) abused women were employed Liem, James, O’Toole, & Boudewyn (1997) Positive self- worth and absence of depression (40 resilient, 105 nonresilient) Sexual abuse = unwanted sexual contact before 14, involving a 5 year age difference overt coercion 16% of men, 24% women were abused N= 687 undergraduates (145 reported CSA) from urban commuter college campuses in the northeast Age- 16-65 G ender- 253 males, 434 females Ethnicity- 66% Caucasian, rest ethnic minority SES- 77% working or middle class Additional demogranhics- 52% Catholic; 72% single Retrospective Nonresilient abused group reported more overt physical force Males with history o f CSA more likely be considered resilient than females (50% vs. 20%) [note: this may be due to the fact that men, overall, did not report depression] Resilient group had more internal vs. external attributional styles, were less likely to blame themselves for the CSA, were less likely to be chronically self destructive, had less stressful family environments during childhood, and were more likely to have Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) sought therapy Lynskey & Fergusson (1997) Did not specifically discuss resilience but divided abused group into three groups based on adjustment: no adjustment problems/no diagnoses (n=26), 1 adjustment difficulty (n=26) 2 or more (n=55) Sexual Abuse = unwanted sexual activities before age 16 divided into 3 groups: noncontact (n=24), contact (n=47), intercourse (n=36) N= 107 sexually abused participants, drawn from a longitudinal study of 1265 children bom in New Zealand in 1977 Ase- 18 Gender- approximately 80% female * no other demographic information given Retrospective accounts of sexual abuse, within a longitudinal study Severity o f abuse was related to increase in rates of adjustment difficulties 26 participants exposed to CSA did not meet criteria for any adjustment difficulties (resilience?) Age at onset (younger), decreasing paternal care, and increasing affiliations with delinquent or substance using peers were associated with increased adjustment difficulties in sexually abused subjects after controlling for affiliations with delinquent or substance using peers and paternal care, CSA does not predict adjustment difficulties McGloin, Successful Abuse & Neglect = N= 1,196 (676 abused and Longitudinal Abused and neglected Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) & Widom (2001) functioning across multiple domains (employment, homelessness, education, social activity, psychiatric disorder, substance abuse, criminal behavior). Individuals illustrating resilience in 6 out of 8 domains were labeled resilient. substantiated cases of child abuse and neglect drawn from juvenile and adult criminal court records. Physical Ahuse= “cruelty to children” with allegations that an individual “knowingly and willfully inflicted unnecessarily severe corporal punishment” or “unnecessary physical suffering” upon the child Neglect - cases in which the court found no proper parent care or guardianship, children were found to be destitute, homeless or living in a physically dangerous environment neglected; 520 controls) 96 cases were sexual abuse, 110 physical, 543 neglect (note: in some cases multiple types of abuse were experienced) Age- 18-41 (mean= 28.7) at follow-up Gender- 48.7% female; 51.3% male. Ethnicity- 2/3 White Comparison group matched on age, sex, race or ethnicity and approximate family background (e.g. SES) females differed from controls on 7 o f the 8 domains assessed (psychiatric disorder, employment, education, homelessness, substance abuse, arrest records, and self- reported violence). Abused and neglected males differed from controls on 6 o f the 8 domains (psychiatric disorder, employment, education, homelessness, arrest records and self- reported violence) Sexual abuse and neglect were significant predictors o f resilience; physical abuse was not. More controls met the criteria for success than abused and Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) Sexual Abuse= ranged from nonspecific charges of assault and battery with intent to gratify sexual desires to more specific acts such as fondling, sodomy, rape and incest. neglected participants. Moran & Eckenrode (1992) Did not specifically examine resilience. Examined protective personality characteristics, such as locus of control orientation and self-esteem. Child Maltreatment 11 neglected, 13 sexually abused, 4 physically abused, 5 physical + sexual abuse (identified via four social service agencies) Nff 33 Age- 12-18 G ender- females Additional dem osraphics- lower middle class Cross sectional Used comparison group o f 112 non-maltreated females ages 12-18. Comparison group from higher SES Self-esteem and locus o f control for good events interacted with maltreatment status to affect depression- Maltreated adolescents with low self-esteem who did not take responsibility for good events reported the highest levels of depression Long-term maltreatment associated with significantly lower levels o f self-esteem, lower levels of internal locus of to •t*. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) control for good events and greater depression. Spaccarelli & Kim (1995) Absence of clinical levels of symptomatology and 2) maintenance o f age-normative levels of social competence Sexual Abuse 70% had at least one type o f “invasive” abuse (anal or vaginal copulation or oral copulation), 21% experienced fondling, 9% experienced non- contact abuse IN N 43 Ase- 10-17 vear old (mean = 14) G ender- 100% female Ethm citv- 74% Caucasian, 19% Hispanic, 5% Black; SES- mean annual familv income = $20,707 Additional dem osraohics- marital status of participating parent- 81% single; 7% separated/divorced; 9% married Cross sectional No comparison group Variability in outcome based on measures (depression, social competence, or combination o f both) and rater agreement (self-report vs. parent report) Higher levels o f abuse stress, negative appraisals, lower levels o f parental support, higher levels o f aggressive control predicted clinical symptomatology. Parent support associated with resilience, as assessed by social competence.. Tremblay, Hebert, & Piche, (1999) Did not specifically examine resilience Sexual Abuse “an action committed by a person giving or seeking an inappropriate sexual 50 Age- Ages 7-17 (mean 9.2) G ender- 39 girls, 11 bovs Ethnicity- 96% Caucasian Cross sectional No comparison group Sexual abuse characteristics not predictive of internalization, social support. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 1 (Continued) stimulation considering the age of the developmental level of the child or adolescent, thus affecting the physical and/or emotional integrity o f the victim, whereas the aggressor has a blood relation with the victim, or excert [sic] a responsibility, authority, or domination position over him or her” All subjects reported a sexually abusive experience w/in the previous 6 months 72% considered very serious abuse, 22% serious, 6% least serious (using severity criterion proposed by Russell). 52% of perpetrators were immediate family, 14% extended (Canadian sample) SES- not reported Additional demographics- 48% from single parent families; average parental education level- 10-11 years Reliance on avoidant coping strategies related to lower global self worth Children who perceived more support by friends and family positively evaluated their global self worth and exhibited less externalizing problems. 27 Tremblay, et al. 1999). No longitudinal studies examining resilient outcomes for victims of childhood sexual abuse were identified. Previous research has acknowledged particular difficulties with the use of cross-sectional designs to study child maltreatment. Specifically, aspects of maltreatment can easily be confounded with the age or developmental stage of the research participants (Trickett & McBride-Chang, 1995). In addition, cross- sectionai, retrospective designs require that independent and dependent variables be assessed at the same time. For example, a retrospective study of CSA asked participants at age 18 if they had any unwanted sexual experiences prior to age 16. Concurrently, participants were questioned about psychiatric symptoms from the period of 16-18 years (Lynskey Sc Fergusson, 1997). Another limitation is that cross-sectional retrospective studies require participants to rely on memories of past events and feelings. The current study examines data from a prospective, longitudinal study of sexually abused females. Outcome measures used to assess resilience in past research It is clear that resilience to maltreatment has been conceptualized in various ways in previous studies. Competent functioning in abused children has been operationalized as: lack of clinical symptoms/pathology (Liem. et al, 1997; Feinauer & Stuart, 1996; Kaufman, Cook, Amy, Jones, & Pittinsky, 1994; Moran & Eckenrode, 1992); positive social adjustment (Lam & Grossman, 1997; Spaccarelli Sc Kim, 1995); positive self-worth/self-esteem (Tremblay, et al., 1999; Liem, et al, 1997; Moran & Eckenrode, 1992); and graduation from high school (Hyman & Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 Williams, 2001; Herrenkohl, et al, 1994). However, criteria for defining resilience are often determined arbitrarily and there is some disagreement about how to accurately identify resilient individuals. For example, Kaufman et al (1994) consider absence of pathology to be a less strict criterion than the existence of above average functioning in other domains, and as such employed absence of pathology as their outcome criterion in order to maximize the likelihood of identifying resilient children. Conversely, some researchers have suggested that social competence may be easier to maintain for sexually abused girls and that symptomatology may be a more sensitive measure of resilience (Luthar, 1991; Spaccarelli & Kim, 1995). Some have suggested that resilient children react to stressful experiences in an internalizing way, which presents itself as depression or anxiety. In addition, some researchers (Radke-Yarrow & Browne, 1993; Luthar, 1991) suggest that even a child who shows remarkable degrees of success in dealing with stress shows setbacks and successes at different times in different domains. In fact, based on her findings that the "resilient" children in her study were significantly more depressed and anxious than were equally competent children from less stressful backgrounds, Luthar (1991) suggests that some type of difficulties may inevitably be related to severe life stresses. In contrast, findings from a study of resilient outcomes for adolescents experiencing high interparental conflict indicated that the resilient adolescents consistently displayed significantly lower levels of internalizing problems (Neighbors, et al., 1993). Given these discrepancies, it is important to investigate whether resilience is consistent across domains. In other words, will an Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 individual who Is categorized as resilient In one domain likewise be classified as resilient on other domains? Domain Specificity In past research resilience has often been studied as if it was a unidimensional construct, rather than allowing for the possibility that individuals vary along multiple dimensions. Resilience research has therefore examined positive outcomes in several domains independently. The most frequently investigated domains include social competence, achievement, and [lack of] psychopathology (Luthar, Zigler, & Goldstein, 1992). Recent research, however, suggests that an individual may be competent in one area yet incompetent in another (McGloin & Widom, 2001; Masten & Coatsworth, 1998). For example, results from a study of maltreated children (mean age, 9 years, 7 months) revealed that almost two-thirds of the participants were academically resilient, whereas only 21% exhibited resilience in the domain of social competence (Kaufman et al 1994). Similarly, results of a study of sexually abused girls indicated that they were much more likely to be classified as resilient based on social competence, rather than lack of symptomatology, based on scores on the CBCL (Spacarelli & Kim, 1995). One explanation for the variation in resilience findings can be attributed to the variability in outcome measures used across studies. While many researchers study the same domains in order to examine resilient outcomes, the same measures are rarely used. Table 2 provides a summary of the measures that have been used to study resilient outcomes in abused and neglected individuals. The table is based on Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 the three most frequently studied domains in resilience research, as cited above. It is worth mentioning that although these three domains are said to be the most frequently investigated domains in resilience research as a whole, clearly they have not been examined as often in terms of resilient outcomes to abuse and neglect. Whereas symptomatology has been used frequently to assess resilient outcomes in victims of abuse and neglect, social and academic competence have been examined less frequently. A review of findings relevant to these three domains follows. Psychological competence. As previously noted, psychopathology is one of the most commonly analyzed outcomes in abuse and neglect victims. As such, lack of symptomatology is frequently studied as an indicator of resilience. Although numerous studies of the impact of CSA on development have shown a bivariate association between CSA & depressive symptomatology, even when this association is present, a large number of subjects do not exhibit symptoms (Zuravin & Fontanella, 1999). Only one of the fourteen studies cited in Table 2 failed to assess symptomatology, and that study was specifically focused on the successful resolution of developmentally salient tasks such as attachment (Farber & Egeland, 1987). Although depression has been one of the most commonly assessed constructs in resilience research, several different measures have been used to assess depression. Some studies assessed depression using either the Child Depression Inventory (Cicchetti & Rogosch, 1997; Spaccarelli & Kim, 1995) or the Beck Depression Inventory (Moran & Eckenrode, 1992; Lam & Grossman, 1997; Liem et al, 1997), whereas others used subscales of measures such as the Trauma Symptom Checklist Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 Table 2. Outcomes used to assess resilience in prior studies of maltreatment Study Symptomatology Social Competence Academic Achievement Cicchetti & Rogosch (1997) Farber & Egeland (1987) CDF Created a prosocial composite score based on peer nomination scores and counselor behavior ratings School risk index Feinauer & Stuart (1996) TSC-33b Herrenkohl, Herrenkohl, & Egolf (1994) Himelein & McElrath CBCLC teacher report form SCL-90-Rd , High school graduation, or still in school (1996) Affectometer 2e Kaufman, Cook, Amy, Jones, & Pittinsky, (1994) Hyman & Williams (2001) CBCLc parent report form & teacher report form TSC-40b Harter1 - child and teacher reports K-ABCk High school graduation Lam & Grossman (1997) BDIf; SCL-90-Rd SAS-SR' Liem, James, O’Toole, & BDIf, b s i8 Boudewyn (1997) Lynskey & Fergusson TSC-33b (1997) McGloin & Widom (2001) Moran & Eckenrode DSM-III-R BDIf Asked questions about extent of participation in a variety of social activities High school graduation (1992) Spaccarelli & Kim (1995) CDF, RCMASh , Tremblay, Hebert, & CBCLC CBCLC Piche, (1999) “ Child Depression Inventory (Kovacs, 1981)) b Trauma Symptom Checklist-33, and -40 - five scales: anxiety, depression, dissociation, post-sexual abuse trauma, & sleep disturbance; measures the psychological sequelae o f childhood trauma (Briere & Runtz, 1985,1989) “ Child Behavior Checklist (Achenbach & Bdelbrock, 1981) d Symptom Checklist-90 Revised- depression & anxiety subscales (Derogatis, 1983) “ Affectometer 2 assesses extent to which positive feelings exceed negative feelings (Kammann & Flett, 1983) f Beck Depression Inventory (Beck & Beck, 1972) “ Tlrief Symptom Inventory (Derogatis & Melisaratos, 1983) ' ‘ Revised children's manifest anxiety scale assesses acute anxiety (Reynolds & Richmond, 1978) 'Perceived Competence Scale for Children (Harter, 1982) 'Social Adjustment Scale Self-report (Weisman & Bothwell, 1976) k Kaufinan Assessment Battery for Children assess children’s factual knowledge and skills using 6 subscales (Kauftnan & Kaufman, 1983) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 (TSC; Feinauer & Stuart, 1996; Hyman & Williams, 2001; Lynskey & Fergusson, 1997), the Symptom Checklist-90 Revised (SCL-90-R; Himelein & McElrath, 1996), and the Child Behavior Checklist (CBCL; Tremblay et al, 1999; Herrenkohl, et al., 1994; Kaufman et al, 1994). The TSC consists of five subscales and was designed to measure traumatic impact, particularly regarding long-term effects of child sexual abuse, whereas the SCL-90-R is a global severity symptom index that measures both frequency and intensity of distress. Each has a sub scale that measures depressive symptomatology. Another important point is that depression is not always used to define resilient functioning; rather sometimes it is looked at as a correlate of resilience. That is, some researchers examine depression as an outcome variable that signifies resilient functioning (Spaccarelli & Kim, 1995; Liem et al, 1997) whereas others classify participants as resilient or nonresilient first and then compare the groups in terms of depressive and other internalizing symptoms (Luthar, 1991; Luthar et al, 1993). This discrepancy can also be seen in studies that use social competence as an indicator of resilience. Social competence. As with studies of the negative impact of CSA, social competence has not been studied as frequently as depression in studies of resilient outcomes for victims of abuse. Furthermore, studies that have examined social competence to signify resilience have used a number of different outcome measures to represent the construct. For example, Cicchetti & Rogosch (1997) created a composite score of prosocial behavior based on peer nomination scores and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 counselor behavior ratings. Lam & Grossman (1997) assessed social competence using the Social Adjustment Scale Self-Report, an instrument designed to measure instrumental role performance at work, leisure activities, and in relations with extended and immediate family members, Kaufman et al. (1994) used the pictorial version of the social competence subscale of Harter’s Self-Perception Profile for Children. In general, findings on social competence in victims of abuse and neglect have indicated lower levels in abuse victims. Cicchetti & Rogosch (1997) found that the maltreated participants in their study exhibited less prosocial behavior than children in the comparison group. What remains open to investigation is whether resilience can be accurately assessed using social competence as an indicator. Only a few studies have examined this question. Lam & Grossman (1997) investigated social competence as it related to a composite of protective factors (generated by summing standard scores of 16 protective factors) in a sample of college students who reported having a traumatic sexual experience during childhood. Their results indicated that individuals scoring higher on the composite index of protective factors were better adjusted (i.e. more socially competent) than those with lower composite scores. In their sample of maltreated children, Kaufman et al. (1994) noted that categorizing an individual as socially competent, or not, varied based on the identity of the informant. The researchers assessed social competence via self- and teacher- reports and found that self-reports of social competence yielded higher identification of resilience than teacher reports (61% and 21%, respectively). More research is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 needed in this area in order to determine whether social competence is a good gauge of resilient functioning. Academic competence. Academic competence is the least often studied of the three domains under investigation in the present study. Still, the few studies that have examined academic competence have done so using varying criteria. Some have examined high school graduation as a correlate of resilience (Herrenkohl, et al., 1994; Hyman & Williams, 2001; McGloin & Widom, 2001) whereas others have developed a school risk index (Cicchetti & Rogosch, 1997). Still others have used a measure developed specifically to assess children’s factual knowledge and skills by assessing expressive vocabulary, arithmetic, verbal concept formation, reading/decoding, reading comprehension, and general knowledge capabilities as a means of defining resilience (K-ABC; Kaufman, et al., 1994). A longitudinal study investigating the effects of child abuse and neglect used CBCL scores of cognitive/academic, social, and emotional functioning to classify participants as resilient (Herrenkohl, et af, 1994). Twenty-five of 105 abused and/or neglected participants were classified as high-functioning, in that they scored in the top 40% of the participants on all three categories. At a subsequent assessment, information about 23 of the 25 resilient participants was obtained. Whereas 14 of the 23 resilient children in the sample had graduated from, or were on course to graduate from high school, only six had B averages or better. Two of the remaining 8 had C averages while 6 had D averages. This study illustrates the arbitrary nature of many resilience definitions. If resilience had been defined as graduating or being on track Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 to graduate from high school with a C average or above, six of the "resilient" participants would have been excluded. Some researchers have examined high school graduation as a characteristic of the individual, rather than using it as an indicator of resilience. Hyman & Williams (2001) compared resilient vs. nonresilient individuals in terms of graduation rates. Resilience was determined using a 13-point resilience scale devised by the authors. The scale incorporated variables related to psychological well-being, health status, interpersonal relationships, social deviance, and economic well-being. Individuals with scores ranging from 10-13 were classified as resilient, whereas those with a score of 2-9 were considered nonresilient. Results from this study of sexually abused women indicated that resilient participants were significantly more likely than nonresilient participants to have graduated from high school (56% vs. 32%; respectively). It is clear at this point that estimates of resilience rates have been influenced by the outcome measure that is chosen to assess competent functioning. The current review illustrates mixed findings which result, at least in part, from the wide array of outcome measures that have been used to define resilience. Another component of resilience research that has resulted in mixed findings involves factors that protect individuals from negative outcomes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 Potential moderators of the relationship between sexual abuse and competent functioning As previously noted, existing resilience research has assessed several factors that contribute to positive outcomes despite exposure to risk. These factors can be seen as protective factors, or moderators of the abuse experience. Some of the most frequently studied factors include social support (O’Grady & Metz, 1987; Spaccarelli & Kim, 1995; Herman-Stahl & Peterson, 1996; Lynskey & Fergusson, 1997; Tremblay, et al., 1999, Masten et al., 1999) and cognitive ability and/or intellectual capacity (Luthar, 1991; Gribble, et al., 1993; Herrenkohl, et al, 1994; Masten et al., 1999). The findings regarding the protective function of these two factors have been mixed. Social support. The literature indicates that social support can positively influence the way in which a child deals with abusive experiences (see Elliot & Carnes, 2001 for a review). Specifically, social support is believed to influence psychological health and adjustment in children and adolescents (Spaccarelli, 1994; Feiring, Taska & Lewis, 1998). Social support has been assessed in a number of ways in the literature, including the number of friends and family members one reports having, as well as by the quality of those relationships. Poor relationships with parents have been found to be related to deficits in academic functioning, cognitive competence, and increased anxiety (Forehand, Middleton & Long, 1987). Neighbors, et al (1993) found that highly competent individuals who had experienced high parental conflict reported significantly better relationships with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 their mothers than participants with lower competence ratings. Competence ratings were based on scores on Harter’s Teacher’s Rating Scale of cognitive competence, with individuals scoring in the top third classified as highly competent and those with scores in the lowest third classified as low competence. The same findings held for both mother- and self-reports of parent-child relationships. While this study did not assess resilient outcomes in abused individuals, the findings illustrate the impact of the mother-child relationship in stressful environments. Other researchers have found conflicting results. Cicchetti & Rogosch (1997) found a significant interaction for abuse status and the child’s reported desire for maternal closeness. More specifically, for maltreated children, a greater need for closeness to mother was related to lower levels of functioning. Of note is the fact that neither of these studies examines the mother-child relationship in sexual abuse victims. In a study of sexually abused girls aged 8 to 15, Feiring, Taska & Lewis (1998) found that parental support was predictive of lower rates of depression. No differentiation was made between the importance of mother-child vs. father-child relationships. The current study specifically examines mother-daughter relationships in sexually abused females. Cognitive ability. In her study of a group of inner-city adolescents ranging in age from 14-17, Luthar (1991) found that intelligence served a protective function for some individuals while serving as a vulnerability factor for others. Luthar (1991) suggests that her findings may indicate that whereas intelligence serves a protective function for younger children, adolescents might invest less in school and put their Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 talents to use in other areas. Other research indicates that high intellectual functioning is predictive of later competence. In their 3-year study of adaptation in school-age (maltreated and non maltreated) children Cicchetti & Rogosch (1997) found that individuals having higher levels of intellectual fonctioning were more likely to be in the high fonctioning group in terms of overall adaptation. Similarly, Masten et al (1999) found that IQ served a moderating role for academic achievement, conduct and peer social competence in their urban community sample. In addition to social support and cognitive ability, severity of the abuse experience has been found to be related to competent functioning in abuse victims. Variability in the severity o f abuse. Whereas much of the extant research on resilience has focused on characteristics of the individual or characteristics of the environment, few studies have focused on variability in adversity. A major shortcoming in the extant research is that almost no one has considered whether risk factors vary systematically. That is, whether variation in the intensity, or severity, of the risk factor plays a role in the outcome. For example, although poverty is often examined as a risk factor, almost no studies examine different levels of poverty. For the purposes of the current research, variability in the characteristics of the sexual abuse experience will be assessed. Of particular interest is how variation in the sexual abuse experience impacts one’s social, academic and psychological fonctioning. As previously noted, a good amount of extant research on child abuse and maltreatment has examined the impact specific characteristics of the abuse Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 experience have on outcomes. However, only a few studies specifically assessing resilience in victims of sexual abuse have addressed this issue. Lynskey & Fergusson (1997) suggest that aspects of the abuse experience may mitigate or modify effects of child sexual abuse on later adjustment. Their study examined factors that ameliorate or protect against adverse consequences of childhood sexual abuse using retrospective accounts of females who reported unwanted sexual experiences that occurred before age 16. Their findings indicated that the mean number of adjustment difficulties an individual experienced increased with exposure to increasingly severe forms of CSA. In addition, increased rates of adjustment difficulties were noted for participants who were younger at the time the abuse began (i.e. less than 10 years old). Other research offers a contrasting view of the impact of abuse related characteristics. Findings from the Liem et al. (1997) study suggest little or no differences between resilient and nonresilient adults with histories of CSA. For the 145 participants in their study, the nature of the abuse experience was similar for both groups, and occurred at around the same age, for similar periods of time. In addition, for both groups the perpetrator was a family member in approximately 30% of the cases whereas the abuse was extrafamilial 70% of the time. The only significant difference between the groups was reported use of physical force, with nonresilient participants reporting more overt physical force. These findings seem to negate the popular assumption that more resilient outcomes can be attributed to experiencing less severe trauma. However, it should be noted that the participants in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 this retrospective study were undergraduate college students. These participants might be considered resilient simply based on the fact that they are functioning well enough to secure entry into and attend college. An alternate way of examining this issue was proposed by Tremblay, et al. (1999). In their model, the influence of abuse related characteristics on internalization, extemalization, and global self-worth was viewed as indirect in that it was proposed to be modulated by coping strategies and social support. Results from their study of 50 sexually abused children (39 girls; 11 boys) indicated no indirect effect of abuse-related characteristics on children’s adjustment based on coping strategies and social support. However, perpetrator identity was significantly associated with children’s internalizing behavior, whereas duration and severity were not. It should be noted that none of the three abuse related characteristics (perpetrator identity, duration, severity) were associated with externalizing behavior problems or global self worth. It seems, therefore, that characteristics of the abuse experience may be related to some outcomes but not others. This highlights the importance of determining criteria by which an individual will be categorized as resilient and identifying factors that are correlated with each domain. In sum, there are several limitations to the extant research on resilient outcomes for victims of sexual abuse. Although eight studies were identified that examined resilience in sexually abused individuals, the following problems were consistently observed: (1) All of the studies used a cross-sectional design, resulting in the simultaneous measurement of both dependent and independent variables. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. importance of this shortcoming is exemplified in the fact that certain constructs (e.g. depression) were considered evidence of resilience in some studies and examined as correlates of resilience in other studies; (2)The majority used retrospective accounts of childhood sexual abuse, relying on self-identification as a means of confirming abuse; and (3) Little attention was paid to definitions of sexual abuse or to the differential impact that characteristics of the abuse can have on outcome. Each of these limitations requires further attention in order to accurately assess resilient outcomes in victims of sexual abuse. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 Chapter 2: The Present Study This study addresses some of the shortcomings of prior research on resilience, specifically in sexually abused individuals. Most studies have assessed competent functioning in one domain in order to determine resilience, while ignoring other domains. McGloin & Widom (2001) assert that “looking at one domain when attempting to identify successful individuals ignores the pervasiveness of risk and undermines the validity or accuracy of the term ‘resilience’”. They therefore support sampling across domains of functioning in an attempt to obtain a more comprehensive picture. Only three of fourteen studies aimed at examining resilience in maltreated individuals assessed symptomatology, social competence, and academic achievement (McGloin & Widom, 2001; Cicchetti & Rogosch, 1997; Kaufman et al, 1994). In each of these studies the three domains were examined using different outcome measures. Many studies have examined resilient outcomes without a clear definition, or understanding of, what it really means to be resilient. It appears that many researchers have hastened to test hypotheses regarding resilient outcomes without first ascertaining what it means to be resilient. It therefore seems likely that the field needs to disregard the common practice of summarily identifying individuals as resilient and instead refer to people as functioning competently in specific domains. For the purposes of this investigation resilience refers to successful adaptation, reflected by psychological competence (e.g. low depressive symptomatology), social competence, and academic achievement, despite exposure Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 to childhood sexual abuse. This study adds to existing research in this area by including a comparison group of nonabused females with demographic backgrounds similar to those of the abused participants. Both abused and nonabused participants were evaluated in terms of competent functioning in each of the aforementioned areas. A higher proportion of the comparison group was expected to meet the criteria for competence however they would not be classified as resilient, given that theoretically they have not experienced the same risk factors as participants in the sexual abuse group. Similar parameters have been used in previous studies of resilience that included comparison groups (McGloin & Widom, 2001). To reiterate, whereas a few studies have investigated whether individuals classified as resilient in one domain are also considered resilient in other domains, even fewer have examined the issue as it pertains to sexual abuse victims. The present study investigates whether discrete categories of competence exist in a sample of sexually abused and nonabused women. Data from a longitudinal study was used in order to assess 1) which of the three domains of functioning results in identification of the largest number of competent participants; 2) the extent to which participants in both groups meet criteria for competent functioning in each of the three domains; 3) the extent to which individuals who meet the criteria for competent functioning in one domain also meet the criteria in the other domains; 4) whether cognitive functioning, social support and other social network variables serve a protective function and are therefore related to competent outcomes; and 5) whether Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 certain characteristics of abuse can be used to predict which individuals are ultimately categorized as resilient and which are not. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 Chapter 3: Method Participants Participants were taken from an ongoing longitudinal study of the psychobiological impact of child sexual abuse on female development (Putnam & Trickett, 1987). The sample for the current study consisted of those individuals who participated in the initial wave of data collection, and again in the fifth wave. The original sample (at Time 1) consisted of 84 abused and 82 comparison participants, for a total sample of 166. Abused participants were referred by protective service agencies in the greater Washington, DC, metropolitan area. Eligibility criteria included the following: 1) the victim was female, age 6-16; 2) disclosure of referring abuse occurred within 6 months of participation; 3) sexual abuse involved genital contact or penetration; 4) the perpetrator was a family member, including parent, or other relative (e.g. older sibling, uncle, grandparent); and 5) a nonabusing parent or guardian was willing to participate in the project. One unique design element of the present study is the inclusion of a comparison group. As previously noted, most studies of the impact of sexual abuse have neglected to include a comparison group. The majority of the comparison participants were recruited via community advertisements. Whereas it is impossible to have a fully controlled experimental design when studying the impact of sexual abuse, the participants in the comparison group for this study were selected to be similar to the abuse group on the following demographic characteristics: geographic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 location, age, ethnic group, and pre-disclosure socioeconomic status (estimated using the Hollingshead, Four-Factor Scale of Socioeconomic status, 1976). Families ranged from low to middle SES, with mean Hollingshead (1975) scores of approximately 36 (defined as “blue collar” or working class). At the outset of the study, participants from both groups lived in the District of Columbia and six adjacent Maryland and Virginia counties. In 70 percent of the cases the comparison and abused participants resided in the same zip code district, in 20 percent of the cases they resided in adjacent zip code districts and the remaining 10 percent lived in comparable, nearby districts. At the initial assessment, participants ranged in age from 6 to 16 years old (M=l 1.11, SD=3.02). Participants were invited back for four subsequent interviews. Data from 12 of the original comparison participants had to be dropped from analyses due to subsequent disclosure of childhood sexual abuse. Given this, 19 new comparison families were recruited to the study for participation at Times 4 and 5. Data for the current analyses was taken from measures given during the first and fifth waves of data collection. The fifth wave of data collection took place an average of nine years after the initial assessment. In order to increase the total N for the current analyses, outcome data from 12 participants who participated in the Time 4 collection but did not return for the Time 5 assessment were also used, resulting in a sample of 162. At Time 5 participants ranged in age from 13 to 28 years old (M=20.35, SD-3.44). Preliminary analyses revealed that abuse and comparison Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 groups did not differ significantly in terms of age, minority status or SES. See Table 3. Table 3. Means and standard deviations for demographic variables for abused and nonabused participants. Abused («=75) Non-Abused (n=87) T otal (n=162) p-value Age 20.68(3.51) 20.07 (3.39) 20.35 (3.44) .28 SES 35.83 (13.96) 37.22(11.34) 36.57 (12.59) .49 Percent Minority 38.7 50.6 45,1 .13 Measures Psychological competence. The Child Depression Inventory (GDI; Kovacs, 1981, 1992), a 27-item self-report measure of depressive symptoms, was administered at all 5 assessment points in this longitudinal study. Each item contains three statements ranging from mild to severe in terms specific depression symptoms. Participants were instructed to choose the item that described them best over the two weeks preceding the interview. Total scores can range from Oto 54. Past research has indicated acceptable levels of internal consistency (Kovacs, 1985), in addition to criterion and concurrent validity (Saylor, Finch, Spirito, & Bennett, 1984). In this sample, an alpha coefficient of .77 was obtained. The GDI was designed for use with children ages 6-18. By Time 5, the majority of the sample was over 18 years old. As such, the wording of some items was changed to reflect the broader range of experiences possible in a sample that includes people both over and under 18 years of age (Noll, et al., 2001). More specifically, items that refer exclusively to school Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48 experiences were changed to include school and work experiences, so that the questions were relevant to all participants, regardless of their level of experience. For example: M y schoolwork (or work at my job) is not as good as before. Social competence. Social competence was assessed using the perceived social competence subscale of the Self Perception Profile for Adolescents (Harter, 1988). The social competence subscale consists of five items that address satisfaction with one's social skills and ability to make Mends easily. Items are scored from 1 to 4. A score of 1 indicates low perceived competence whereas a score of 4 reflects higher perceived competence. Scores are subsequently summed and averaged, with a resulting score ranging from 1 to 4. Alpha reliabilities from .78 to .84 have been reported for this measure (Harter, 1985). The alpha coefficient on the social competence subscale for the current sample was .79. Academic Competence. More than half of the sample was 18 or older by Time 5. For those individuals, high school graduation was used as an indicator of academic competence. Respondents were asked to indicate their highest level of education at the time of the interview. These responses were then converted into a dichotomous variable, with graduating from college indicating academic competence. Similar methods have been used in a recent study of resilience among abused and neglected children (McGloin & Widom, 2001). Given that there were still some participants younger than 18 at Time 5 an alternate approach to examining academic ability/achievement was employed. The Applied Problems subtest from the Woodcock Johnson Psychoeducational Battery Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 (WJ-R; Woodcock & Johnson, 1989) was administered to all participants at Time 5. The WJ-R is an individually administered battery of subtests of cognitive abilities and achievement that was standardized on more than 6300 subjects ranging in age from 2-95. The Applied Problems subtest measures skill in analyzing and solving • practical numerical problems. Items on this subtest require a participant to discriminate between relevant and irrelevant information in a problem, identify the procedures that need to be used, and then perform the necessary mathematical calculations (McGrew, 1994). Concurrent validity studies have shown correlations in the .60 to .70 range when the WJ-R is compared to other cognitive abilities or achievement tests. Predictors of resilience Social Support. Social support was assessed using items from a modified version of the Children’s Social Support Questionnaire (CSSQ; Bogat, Chin, Sabbath & Schwartz, 1985). This instrument assesses the qualitative and quantitative characteristics of the adolescent’s social support network and yields data on the number, intensity and satisfaction of relationships with peers, relatives and other adults. The variables used for the current analyses include the number of people in the child’s social network and the child’s average happiness rating for people in their social network. Number of people in the child’s social network was further separated into number of males and number of females for the current analyses based on previous research which suggests that the impact of ones social network might differ based on the gender of the peers and/or family members being investigated. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 Happiness ratings range from 1 (very unhappy) to 5 (very happy). A happiness rating was obtained for each individual the child listed as part of their network and those ratings were then summed and averaged. Mother-daughter closeness. The mother portion of the Intimacy Scales (Blyth, Hill & Thiel, 1982) was used to assess the degree of closeness participants perceive themselves to have with their mothers. Items on this scale examine closeness vis-a-vis perceived maternal support (e.g. how much can you count on your mother no matter what) and identification (how much do you want to be like your mother). Participants responded to eight items on a 3-point scale (1= not at all; 2= a little; 3= a lot). Total scores can range from 8 to 24. Higher scores signify more closeness. An alpha coefficient of .72 was obtained for this sample. Cognitive ability. Form L of the Peabody Picture Vocabulary Test Revised (PPVT-R; Dunn & Dunn, 1981) was used to assess receptive vocabulary at Time 1. Although it is not a direct measure of intelligence, it has been shown to correlate .70 with foil IQ scores and has well established reliability and norms. Therefore, for the purposes of this study the PPVT-R was used as an indicator of general intellectual functioning or cognitive maturity (Aber & Allen, 1987). Previous studies have suggested that higher intellectual functioning might serve as a resource for disadvantaged children (Cicchetti & Rogosch, 1997). Characteristics o f the abuse. Data regarding characteristics of the abuse were obtained at the outset of this longitudinal study using the Caseworker Abuse History Questionnaire (CAHQ). Information about six characteristics of abuse was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 examined: 1) Age at onset o f abuse, coded in years. 2) Duration o f abuse, coded in months. 3) Identity o f the prim ary perpetrator, classified as either biological father, other father figure including mothers’ live-in boyfriend, or other relative. 4) Severity o f abuse, using a count of whether or not the abuse included vaginal penetration, anal intercourse, digital penetration, oral sex, masturbation, fondling/sexual kissing. Thus, possible scores range from 1 to 6. 5) M ultiple perpetrators, indicating whether a child experienced sexual abuse with a person in addition to the one identified as the primary perpetrator. A score of 0 was assigned if there were no other perpetrators or a score of 1 if there was. 6) Physical violence, representing one of three categories: (l)th e abuse was accompanied by physical force or violence; (2) the abuse was accompanied by threats of violence and/or whether threats were made concerning disclosure of abuse; and (3) the child experienced physical abuse separate from that accompanying the sexual abuse. Procedures At Time 1, information about sexual abuse was reported by caseworkers, using the Caseworker Abuse History Questionnaire. This instrument was designed to record information that led to referral of a case to a protective service agency and then resulted in the classification of abuse. The PPVT-R and CSSQ were administered to participants during a 2-3 hour testing session, during which other measures of social, emotional, and physical development were administered. Time 5 data was collected during a 3-4 hour assessment, which included paper and pencil measures, computerized measures, and interview style questions designed to assess Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 various physical, psychological and social development variables. At both assessments a non-abusing caretaker accompanied participants who were not yet 18 years old. However caretakers waited in a location separate from the testing area so that they were not present when the participant was completing the measures. Monetary compensation was distributed at the end of each assessment. Missing data imputation In order to avoid losing valuable information, data was imputed at the first order. First-order missing data refer to those in which a participant did not respond to a specific item of a scale (e.g., in the GDI, or Harter). The Estimation- Maximization (EM) algorithm (Dempster, Laird & Rubin, 1977; Orchard & Woodbury, 1972) was used to replace the missing information with a “best guess.” These procedures use information (that is related to the missing information) in multiple imputations that converge over iterations to a minimum-error estimation of the missing information. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 Chapter 4: Results Using cutoff scores to examine competence For the initial set of analyses cutoff scores were used to determine competent functioning. A description of how the cutoff scores for each domain were determined follows. Psychological competence. In general, scores greater than 12 on the CDI have been used to indicate mild depression, whereas scores of 19 or above have been found to indicate clinically significant levels of depression (Smucker, Craighead & Craighead, 1986). However cutoff scores are typically determined in consideration of the population being studied (Sitarenios & Kovacs, 1999). For example, when administered in a clinical setting, lower cutoff scores would be used. Kovacs (1992) has recommended the raw score of 20 as an acceptable cutoff score for use with the general population. As such, the current study used a raw score of 20 as a cutoff for identifying persons who are at-risk for a depressive disorder, and therefore noncompetent/nonresilient. Conversely, participants with CDI scores below 20 were classified as competent/resilient. Social competence. As yet there has been no precedent for determining cutoff scores on Harter’s subscale of social competence. However, one study was identified that used a score of 3.0 as the cutoff (Kaufman, et al. 1994). In general, scores below the midpoint of 2.5 are considered relatively low and those below 2.0 are even more problematic (S. Harter, personal communication, December 5, 2001). For the purposes of this research a score of 3.0 was used to indicate competent Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 54 functioning, based on the assumption that a person who is resilient should not be functioning at a “relatively low” level. Thus, participants scoring below 3.0 were classified as noncompetent/nonresilient. Academic competence. For the purposes of this study, graduation from high school was used as one indicator of academic success [for participants over the age of 18], Participants who were over the age of 18 and had graduated from high school were considered to be functioning competently, whereas participants over 18 who had not graduated were classified as noncompetent/nonresilient. Because nearly 34% of the sample was not yet 18, scores on the WJ-R Applied Problems were used in subsequent analyses as an indicator of academic competence for the entire sample. The User’s Manual for the WJ-R classifies individuals scoring more than one standard deviation below the mean as below average (Woodcock & Johnson, 1989). As such, scores were converted into a dichotomous variable with participants scoring more than one standard deviation below the mean classified as not resilient or not functioning at a competent level. Table 4 depicts the cutoff points used to classify participants as competent/resilient, in addition to the means, standard deviations and number of participants scoring in the competent range for each of the measures. As can be seen, scores on the depression measure (CDI) resulted in identification of the largest number of competent participants (88.3%) followed by scores on the social competence subscale (78.4%). The fewest number of participants were classified as competent based on scores on the WJ-R Applied Problems (69.1%). Chi-square Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 analyses indicated that there were no significant group differences in the number of participants scoring in the competent/resilient range on each domain of functioning2 . Table 5 depicts the means, standard deviations, and percent of participants scoring in the competent range of each domain by group. Table 4. Means, standard deviations, cutoff scores, and percent of participants i in the competent range for entire sample. Measure Mean (SB) Cutoff Score Number Competent Harter CDI Applied Problems 3.33 (.61) 13.19 (5.36) 522.46 (17.45) 3.0 20 515 127 (78.4%) 143 (88.3%) 112(69.1%) Table 5. Means, standard deviations, and percent of participants scoring in the competent range _______ Abused Non-Abused Measure Number (percent) in competent range Number (percent) in competent range Harter CDI Applied problems 3.26 (.67) 13.93 (5.95) 520.07 (18.14) 60 (80%) 63 (84%) 50 (66.7%) 3.30 (.56) 12.55 (4.6) 524.52 (16.67) 67 (77%) 80 (92%) 62 (71.3%) .212 2.46 .399 2 Additional analyses of covariance were performed for each domain using continuous scores on all three outcome measures and controlling for demographic variables. These analyses revealed that abuse status was significantly related to academic competence (i.e. scores on the WJ-R Applied Problems) when controlling for age, SES, and minority status. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 Overall competence/resilience. Based on whether or not they scored in the competent/resilient range, participants were assigned a score of 0 (not competent) or 1 (competent) for each of the three domains of functioning. Scores were then summed across all three domains in order to determine the number of domains in which each participant was classified as competent/resilient. Participants that scored in the competent/resilient range on all three domains were considered to have obtained overall competence/resilience. Table 6 illustrates the number of participants who received scores within the resilient range on each of the domains. Only fifty percent of the total sample scored in the competent/resilient range in all three domains. Another 38% of the sample was competent in two out of three domains whereas a little over 10% were competent in only one domain. Three participants did not receive scores in the competent/resilient range in any o f the domains. Although there were no significant group differences, the three participants that failed to receive scores in the competent/resilient range on any of the domains were from the abuse group. Table 6. Total number of domains in which participants received scores in the competent/resilient range Number of Domains Abused Comparison Total 0 3 (4%) 0 3 (1.9%) 1 9 (12%) 8 (9.2%) 17 (10.5%) 2 25 (33.3%) 36(41.4%) 61(37.7%) 3 38 (50.7%) 43 (49.4%) 81 (50%) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 Next the domains of competence were paired and chi-square analyses were conducted on each pair in order to determine the extent to which participants were more likely to be classified as resilient in one domain than another. Tables 7-9 present cross tabulations representing participant classifications on each pair of domains. As can be seen in Table 7, participants were significantly more likely to be classified as competent/resilient on the measure of psychological competence than on the social competence measure (chi-square = 12.23, d f - 1, p < .001). When academic and social competence scores were cross tabulated comparable numbers of competent/resilient participants were identified. Similarly, equivalent numbers of competent/resilient participants were identified by cross tabulating scores on measures of academic and psychological competence. Table 7. Cross tabulation of competent/resilient classifications for social Social Competence P sK ltltS ic tl competence Non-Competent Competent TOTAL Non-competent 10 (6.2%) 9 (5.6%) 19(11.7%) Competent 25 (15.4%) 118(72.8%) 143 (88.3%) TOTAL “T3“”Tr— 35 (21.6%) 127 (78.4%) 162 (100%) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Table 8. Cross tabulation o f competent/resilient classifications for academic competence and psychological competence. Acafejk C cjU fflstaiffiE Psy^hqlpgtcai cpmt s&nce Non-competent Competent TOTAL p = .324 Non-Competent 4 (2.5%) 46 (28.4%) 50 (30.9%) Competent 15 (9.3%) 97 (59.9%) 112(69.1%) TOTAL 19(11.7%) 143 (88.3%) 162 (100%) Table 9. Cross tabulation of competent/resilient classifications for academic competence and social competence. Academic, Social P mp s f o m Non-competent Competent TOTAL "W ^~245, p = .621 Non-Competent 12 (7.4%) 38 (23.5%) 50 (30.9%) Competent 23 (14.2%) 89 (54.9%) 112(69.1%) TOTAL 35 (21.6%) 127 (78.4%) 162 (100%) A step-wise discriminant analysis was performed in order to determine which of the variables assessed at Time 1; abuse status, number of males and females in social network, overall happiness with members of social network, mother-daughter closeness and cognitive ability, would best discriminate between the two groups (competent/resilient vs. noncompetent/nonresilient) on overall resilience. Cognitive ability, as defined by Time 1 PPVT scores, was the primary predictor for the discriminant function, which accounted for 18% of the variance. On the basis o f all Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 5 variables, 68.9% of participants were correctly classified into the 2 original groups (see Table 10). As can be seen, correct classification was nearly the same for both groups: 70% for the non-competent group and 67.7% for the competent group. Table 10. Classification results from the Discriminant Function Analysis fedjcM ,^ ffllRMembershij ActMaLGreup # o f tor-Com petest Competent cases Non- 60 42 18 competent (70%) (30%) Competent 62 20 42 (32.3%) (67.7%) Note: 68.9% of original grouped cases correctly classified Using continuous outcome variables to examine competence Some researchers have questioned the use of cutoff scores in determining competent/resilient outcomes. In order to examine the impact that using cutoff scores might have on the findings o f the current study, additional analyses were performed in which the data was examined in its original form. Scores on the CDI, Harter, and WJ-R Applied Problems were examined as continuous variables. The relationship between social competence, psychological competence, and academic competence and the variables assessed at Time 1 (abuse status, number of males and females in social network, overall happiness with members of social network, mother-daughter closeness, and cognitive ability) were examined via multiple regression. Intercorrelations of demographic and predictor variables for the entire sample are presented in Table A! in the Appendix. Separate regression Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 analyses were conducted for each of the three domains of competence. In order to conduct these analyses each of the social network variables, the mother-daughter closeness variable, and the cognitive ability variable were centered, or converted into deviation score form, as suggested by Aiken & West (1991). The abuse status variable was recomputed such that abused participants were coded 1 and comparison participants were coded -1. Given that there are relatively equivalent group sizes this recode allows the group variable to have a mean of zero and a standard deviation of approximately one, much like the standardized social network and cognitive ability variables. Product terms were then computed for each of the social network variables and the cognitive ability variable with abuse status in order to test the possibility that competent outcomes are moderated by the interactions between abuse status and the social network and cognitive ability variables. For each of the three regression analyses, the domain of competence (i.e. social, psychological, academic) was regressed onto main effects (abuse status, number of males and females in social network, mother-daughter closeness, and cognitive ability) and all 2-way interaction terms simultaneously. All analyses controlled for age, SES, and minority status. Within these analyses, significant interaction effects would indicate moderation. Table 11 summarizes the results of the multiple regressions for each of the three domains. Significant adjusted R2s were obtained for two of the three equations (social competence & academic competence). There was a significant interaction effect for number of males in social network x abuse status for social competence. The analysis also revealed a significant main effect of overall happiness with social Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 11. Results of multiple regression analyses using social network variables, mother-daughter closeness, and cognitive ability as predictors o f domains of competence Social Competence Psychological Competence Academic Competence Age -.047 -.043 .217** SES .092 -.215* .203** Minority Status -.053 -.099 -.073 Abuse status -.165* .109 -.047 # of males in social network (#MSN) # of females in social network (#FSN) Overall happiness w/ social network (HWSN) Mother-daughter closeness (MDC) -.046 .080 .262** -.166 .050 .041 -.147 -.066 -.083 .148** -.002 -.026 PPVT-R scores .011 -.069 .594*** #MSN x abuse status -.294** .040 -.092 #FSN x abuse status .109 .068 .078 HWSN x abuse status .069 .055 .085 MDC x abuse status -.028 .068 -.080 PPVT x abuse status -.012 -.039 -.178** Adjusted R2 .092 -.013 .568 F 1.79* .904 11.22*** * p< 05, ** p< 01, ***p< 001 *p<-10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 62 network for social competence. Finally, there was a marginally significant main effect of abuse status for social competence. In the domain of academic competence, a significant cognitive ability x abuse status interaction was detected. In addition, the analysis revealed significant main effects of age, SES and number of females in the social network for academic competence. Significant two-way interactions were probed further using procedures outlined by Aiken & West (1991). Aiken & West (1991) suggest that when significant interactions are found the slopes of the simple regression lines should be analyzed in order to assess whether they differ significantly from zero. Simple slopes of the line predicting social competence from number of males in social network were calculated at conditional values of abuse status (abuse = 1, comparison = -1). T-tests of the simple slopes predicting social competence reveal that at the abuse value (abuse status = 1), the relation between abuse status and number of males in social network was significant and negative, whereas at the comparison value the relationship was marginally significant and positive. Figure 1 presents the regression lines for social competence regressed on number of males in social network at conditional values of abuse status. Simple slopes of the line predicting academic competence from Time 1 PPVT scores were also calculated using conditional values of abuse status. The slope predicting academic competence was significantly positive at both abuse values, however the relation was steeper at the comparison value. Figure 2 presents the regression lines for academic competence regressed on Time 1 PPVT scores at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1 . Regression lines predicting social competence from males in social network at conditional values of abuse status 3.5 8 I 3 .3 £ 0 u a 1 3.1 -0.8 - 0.6 -0.4 -0.2 0.2 0.4 0.6 as Mates in social network in SD from Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 Figure 2. Regression lines predicting academic competence from Time 1 PPVT scores at conditional values of abuse status SB- 500 485 I u u s < 0.6 1 -0.2 0.2 0.4 0.6 - 1 -0.8 -0. 6 -0.4 0 Cognitive a bil!*y [at Tim s 1} is SDfrom mean Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 conditional values of abuse status. Table 12 presents significance tests for simple slopes. Table 12. Significance tests of simple slopes o f regression lines predicting social competence from number of males in social network and academic competence from Time 1 PPVT scores at conditional values of abuse status Domain ________________ §___________________ t Social competence Abuse value -.200 -2.33* Comparison value .146 1.91^ Academic competence Abuse value 7.07 4.12** Comparison value 13.18 7.55** Effects of characteristics of abuse in sexually abused group In order to investigate the impact of characteristics of abuse on resilient outcomes in participants from the abuse group, separate regression analyses were performed for each of the three domains of competence. In each analysis, the domain o f competence was regressed onto main effects (perpetrator identity, multiple perpetrators, age at onset, duration, physical violence, and severity of abuse). A separate regression analysis was done to examine the impact of the abuse characteristics on overall resilience (across all three domains). Intercorrelations of the demographic variables, predictor variables and characteristics of abuse are presented in Table A2 in the Appendix. As can be seen in Table 13, only the equation for academic competence produced a significant adjusted R2, accounting for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 13. Results of multiple regression analyses using characteristics of abuse as predictors of domains of competence Social Competence Psychological Competence Academic Competence Overall resilience Age .257 -.337 -050 .246 SES .038 -.364** .236* .131 Minority Status -.054 -.139 -.346** -.212 Perpetrator (bio father vs. other -.273* -.152 -.194* .148 relative) Multiple .079 -.032 -.012 .108 perpetrators Age at onset -.169 .336 .241 .014 Duration -.045 .442* .083 -.014 Physical violence -.022 -.133 -.303* -.291* Severity o f abuse -.071 -.059 .088 .024 Adjusted R2 -.057 .017 .303 .115 F .619 1.12 4.10*** 1.93* *p<.05, ** p<.01, ***p<.001 *p<.10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 30% of the variance. Results indicated that physical violence was significantly related to academic competence when controlling for age, SES and minority status. Specifically, more physical violence was associated with lower scores on the WJ-R Applied Problems. In addition, there was a significant main effect for duration of abuse on psychological competence; longer durations were associated with higher depression scores. Finally, the identity of the perpetrator (biological father vs. other) was a marginally significant predictor of academic and social competence (p=.09 and p=.06, respectively). This finding indicates that girls who were sexually abused by their biological fathers were more likely to score lower on WJ-R Applied Problems, and have lower scores of self-reported social competence, than girls who were abused by another relative. Finally, additional analyses were conducted using cutoff scores to examine the impact of abuse characteristics on each of the three domains of competence. The purpose of these analyses was to determine whether the results would be comparable to those obtained when the impact of the characteristics of abuse was examined using continuous scores. The results of these analyses are presented in Table 14. There were two significant differences between the groups. First, there was a significant difference in the identity of the perpetrator for the social competence measure (% = 5.96, p = .015). Second, sexual abuse accompanied by physical violence was significantly related to lower academic competence (% 2 = 8.24, p = .001). These findings are similar to those obtained using continuous scores. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 14. Characteristics of abuse for resilient and nonresilient sexually abused participants in each domain o f competence Social Competence Pyschological Competence Academic Competence R NR Difference R NR Difference R NR Difference ______________ (n=59) (n=15)_________________ (n=62) (n=12)_________________ (n=61) (n=13)_____________ Age at onset 7.89 7.60 Duration (in months) 25.13 32.94 Perpetrator identity Biological 10 7 father (58.8%) (41.2%) Other 49 8 relative (86.0%) (14.0%) Multiple perpetrators Yes 26 6 (81.3%) (18.8%) No 33 9 (78.6%) (21.4%) Violence Present 28 8 (77.8%) (22.2%) Absent 31 7 (81.6%) (18.4%) .084 .847 X = 5.96* 7.95 24.94 14 (82.4%) 48 (84.2%) 27 (84.4%) X2 = .081 (83.3°/ 7.21 35.89 (17.6%) 9 (15.1 (80.6%) 33 (86.8%) (15.6%) 7 (16.7%) 7 (19.4%) 5 (13.2%) F = .467 F = 1.41 7.59 8.96 28.09 20.29 9 8 (52.9%) (47.1%) 40 17 20 12 18 (50%) 31 (50%) 7 = .538 (81.6%) (18.4%) F = .169 F = 1.89 % = .033 (70.2%) (29.8%) X 2 = 1.74 (62.5%) (37.5%) 29 13 X 2 - .015 (69%) (31%) x2 = -348 8.24** Note: R=resilient; NR=nonresilient; *p<05, **p<001 Os 69 Chapter 5: Discussion In recent years there has been an Increase In research on resilient outcomes for abused and/or neglected children. Little has been done, however, to specifically examine resilient outcomes in sexually abused individuals. Furthermore, the few studies that have been done in this area have been plagued with many methodological problems. The current study sought to address some of the problems in the existing research by I), examining resilient outcomes using data from a prospective longitudinal study; 2). including a comparison group; 3). examining competent functioning across multiple domains; 4). attempting to replicate previous findings that indicate that certain variables serve a protective function and are thus predictive of competent outcomes; and 5). examining the impact that specific characteristics of the sexual abuse experience may have on competent functioning later in life. The largest number of competent participants was identified in the psychological competence domain. This finding supports the work of previous researchers who consider absence of pathology to be a less strict criterion for resilience than the existence of above average functioning in other domains. Based on this assumption, Kaufman et al (1994) used absence of pathology as their outcome criterion in order to maximize the likelihood of identifying resilient children. Findings from their study indicated that children were more likely to be classified as resilient based on ratings of clinical symptomatology than measures of social competence. This finding contrasts the findings of Spaccarelli & Kim (1995) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 70 in which sexually abused girls were more likely to be classified as resilient based on social competence, opposed to lower symptomatology. Similarly, Luthar (1991) reported that "resilient" (i.e. socially competent) children in her study were significantly more depressed and anxious than were equally competent children from less stressful backgrounds. There are several feasible explanations for the difference in outcomes in these studies. First, each study employed very different samples. Luthar examined resilience in inner city youth whereas the participants in the Kaufman et al (1994) study were maltreated school-age children and the participants in the Spaccarelli & Kim (1995) study were sexually abused girls. Secondly, differential findings may be due in part to variations in the outcome measures and definitions of resilience used in each study. Finally, the two studies that examined resilient outcomes in abused samples (Kaufman et al, 1994; Spaccarelli & Kim, 1995) were cross sectional and therefore do not address psychological functioning across time. Additional research is needed that specifically examines lower psychological symptomatology as an indicator of resilient functioning using sexually abused samples. In the past researchers have suggested that an individual may be competent in one area and incompetent in another (Masten & Coatsworth, 1998). As previously noted, many researchers believe that individuals are likely to show setbacks and successes at different times in different domains (Radke-Yarrow & Browne, 1993; Luthar, 1991). Results of the current study offer some support for this theory. Although use of the term “resilience” has been criticized in that it somehow implies Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 a unidimensional construct rather than allowing for the possibility that individuals vary along multiple dimensions, the findings of the current study suggest that individuals do in fact vary In their levels of competence from one domain to another. Participants in the current research were significantly more likely to be categorized as competent/resilient in terms of low depressive symptomatology than on the social competence measure. However, measures of academic achievement and clinical symptomatology resulted in identification of comparable numbers of resilient children. These findings are similar to those reported by Kaufman et al. (1994). The majority of the studies that have examined competence across multiple domains have done so using only at-risk samples. The current study adds to existing research in that it utilizes a comparison group drawn from a non-clinical setting. Findings from the current study conflict with those ofMcGloin & Widom (2001), who found that the controls in their study were less likely to meet the criteria for success than abused and neglected participants. Cicchetti et al (1993) reported findings similar to those of the current study. The maltreated children in their study had fewer areas of adaptive functioning than the nonmaltreated children. The two groups (abused and non-abused) in the current study were equally likely to score in the competent range on all three domains. One explanation for these conflicting findings is that the criteria for resilience in the McGloin & Widom (2001) study were more stringent in that there were more domains under investigation. Participants were required to exhibit competence in 6 of 8 domains in order to be considered Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 successful. Perhaps if the current study had included additional domains of competence a distinction between the two groups might have been detected. The analyses reported here demonstrate the difficulty in operationalizing the concept of resilience. It is still unclear at this point how best to determine resilience using multiple domains of competence. Although using multidimensional assessment strategies to define resiliency reduces the likelihood of incorrectly labeling children who are competent in one domain but not in others, there is still no clear resolution as to which are the most important domains of functioning, and how many domains one must demonstrate competency in so as to be classified as resilient. There remains, then, a question as to whether it would be “safer” to label people as competent in specific domains, rather than attempting to come up with a composite of overall resilience. For example, perhaps labeling participants as psychologically competent, or socially competent, would be more accurate than saying that they are “resilient”. In this sense the term resilient holds little meaning unless it is specifically defined so that the reader understands exactly which domains of functioning are under examination. Saying that a person is resilient seems to imply, at least in lay terms, that they individual is functioning satisfactorily in all facets of life. In fact, findings from the current study do not support this assumption. It seems much more plausible to specifically describe people as functioning competently in certain domains, even if they are not functioning well in others. One other consideration is that there are some factors that may influence competent functioning in certain domains independent of abusive or traumatic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 experiences. Ethnicity, or more specifically, minority status, is one such variable. In the current study minority status was a significant predictor of academic competence. Specifically minority status was negatively related to scores on the WJ-R Applied Problems. This finding emphasizes the need for statistical control of such variables in future research in this area. It also highlights the fact that most of the previous research on competent functioning in abused individuals has not included ethnically diverse samples. Of the studies of sexual abuse discussed here, many had an overwhelming majority of Caucasian participants (Lam & Grossman, 1997; Liem et al, 1997; Spacarrelli & Kim, 1995; Tremblay, Hebert & Piche, 1999); another sample was 86% African American (Hyman & Williams, 2001); and others did not report ethnic breakdowns (Feinauer & Stuart, 1996; Lynskey & Fergusson, 1997) so it is impossible to know whether ethnicity played a role in their findings. It is imperative that future research address ethnicity as a potential mediating factor in the impact of sexual abuse on development. Conversely, it is possible that the presence or absence of certain protective factors might impact competent functioning. Thus, the presence or absence of those factors could influence the number of domains in which an individual is competent. Protective factors The findings of the current study did not indicate a significant relationship between an individual’s perceived closeness with their mother and positive outcomes later in life. This finding contrasts earlier work in this area that has consistently suggested that parental support is associated with adjustment in sexually abused Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 74 children (see Elliot & Carnes, 2001 for a review). One explanation for this may be the outcome measure used to assess mother-daughter closeness. Given that the measure used in the current study was only an 8-item instrument, one suggestion is that future studies could benefit from a more detailed instrument to assess perceived maternal support. For the sexually abused participants, additional items tapping into how much the girl felt supported by her mother immediately following disclosure should definitely be included. For example, questions that specifically ask how the girl’s mother responded to the disclosure, whether or not she believed her daughter, and how supportive she was subsequent to the disclosure are just a few items that should be included in future research on the impact of mother-daughter closeness in sexually abused girls. It would be negligent to think that, based on the findings presented here, mother-daughter closeness is not related to positive outcomes later in life for sexually abused girls. Research in this area could also benefit from having data from multiple informants, for example both mother- and daughter- reports on the quality of their relationship. In addition, when possible future studies should also examine the father-daughter relationship. Lynskey & Fergusson (1999) found that reports of increasing paternal care, defined as the extent to which young people describe their father as providing support, affection and nurturing, was associated with decreased adjustment difficulties. Similarly, Herman-Stahl & Peterson (1996) found that the resilient adolescents in their study reported more intimacy with their fathers. Others have reported contrasting results. Neighbors et al (1993) found that only the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 « adolescent’s relationship with their mother was associated with different levels of competency. Given that the participants in this study were victims of intrafamilial abuse, and that 23% of the sexually abused girls were abused by their biological father, it was not feasible to assess the extent to which paternal care impacted competent outcomes in the current study. When other social network variables were examined, the results indicated that the impact of number of males in ones social network was moderated by abuse status. The number of males in ones social network appears to have different implications for abused vs. nonabused girls in this study. Whereas the relationship between abuse status and number of males in social network was negative at the abuse value, at the comparison value the relationship between abuse status and number of males in social network was positive. This finding has not been reported in previous research however it appears to be a revealing discovery that should be examined further in future studies of resilient outcomes, especially in sexually abused females. It seems possible that having more males in ones social network is a detriment particularly for sexually abused girls. One explanation is that for the girls in the sexually abused group being exposed to more males meant an increase in the likelihood that they would experience abuse. For example, the girls in the sexually abused group may have had more contact with males who were not their father, father figure, or other immediate family member (e.g. mother’s live-in boyfriend, step-brothers, etc.) than girls in the nonabused group. Again, this theory is difficult to assess in the current study given that participants were victims of intrafamilial Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 abuse. Future research in this area should closely examine the number of and identity of males that the participant has lived with throughout her life. It seems possible that as the number of non-related males that a girl has contact with increases so too will the likelihood that she will experience some type of abuse. The results of the current study indicate that the relationship between cognitive ability assessed at Time 1 and subsequent academic competence was moderated by abuse status. Although slope analyses indicated that the relationship was significantly positive at both abuse values, the relation was steeper at the comparison value. It seems that for the abused participants in this study some other variable might be interfering with academic outcomes. The findings of this study also corroborate earlier findings that maltreated children score significantly lower than nonmaltreated children on the PPVT (Bamett, Vondra & Shonk, 1996). Sexually abused girls also tended to score lower on the WJ-R Applied Problems when age, SES and minority status were controlled for. Future studies are needed that investigate other factors that might influence academic competencies in sexually abused girls. In terms of social and psychological competence, the current study did not provide support for the hypothesis that individuals having higher levels of intellectual functioning are more likely to be resilient (Cicchetti & Rogosch, 1997). Cognitive ability at Time 1 was not significantly related to scores on the social competence and psychological competence measures. One possible explanation is that whereas intelligence serves a protective function for younger children, it does not serve the same function in adulthood (Luthar, 1991). Perhaps being Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 Intellectually adept Is more Important for younger children who’s peers are more apt to tease or make fun of them than for adolescents, who’s peers might not place as make Importance on one’s intellectual capabilities. In adolescence a child who is extremely affable and charismatic might be socially competent despite not having the same intellectual capabilities as Ms or her peers. It is also plausible that the domains of psychological and social competence could have been more accurately assessed using alternate measures. TMs possibility will be discussed in more detail later. It Is clear though, that additional longitudinal research is needed to test these assumptions. Characteristics of abuse As previously noted, findings regarding the impact that certain characteristics of the abuse experience have on competent functioning have been mixed in prior research. Whereas some researchers suggest that abuse characteristics impact later adjustment (Lynskey & Fergusson, 1997) others have found few differences (Liem et al., 1997). Findings for the current study varied depending on the abuse characteristic being examined, as well as the domain of competence that was under investigation. The current results are somewhat similar to those of Liem et al. (1997), in that when resilience was defined as academic competence, nonresilient girls were more likely to have experienced sexual abuse accompanied by physical abuse. Similarly, Liem et al. (1997) found that nonresilient sexually abused girls were more likely than resilient girls to have reported the use of overt physical force. However, there were no significant differences between reports of physical violence Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 by resilient and nonresilient participants in the psychological and social competence domains of the current study. Many researchers have suggested that abuse by someone close to the victim is more traumatic than abuse by a stranger or an acquaintance (Browne & Finkeihor, 1986; Sirles, Smith & Kusama, 1989; Adams-Tucker, 1982; McLeer, Deblinger, Atkins, Foa & Ralphae, 1988; Russell, Schurman, & Trocki, 1988). The results of the current study provide limited support for this hypothesis. Sexually abused girls who were abused by their biological father scored lower in academic and social competence than girls abused by other relatives. Perpetrator identity (biological father vs. other relative) was not associated with psychological competence in this study. In previous work using data from earlier assessments in this study (Trickett, Reiffman, Horowitz & Putnam, 1997) abuse by a biological father was found to be associated with problem behaviors such as demanding suspicious behavior, sexual acting out, withdrawn secretive behavior, dissociative hyperactivity and aggressive delinquency. Taken together, these findings support the assertion that the identity of the perpetrator should be considered when examining the impact of sexual abuse on resilient outcomes. Abuse by ones biological father seems to be more detrimental than abuse by other relatives but this may be true only for certain domains of functioning. One additional consideration is that girls who were abused by their biological fathers usually experienced the abuse for longer durations, which likely impacted subsequent functioning. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 Limitations of the study and implications for future research This research contributes to the literature on resilient outcomes for sexually abused individuals and presents new data regarding moderating effects of social network variables and cognitive ability on competent functioning. There are however, some limitations that impact the conclusions that can be drawn from the current research. First, this study attempted to examine the implications of using cutoff scores to define resilient functioning, as compared with using continuous outcome variables. It has become somewhat common practice in the resilience literature to use cutoff scores to differentiate between resilient and nonresilient participants. One problem with this method is that although the cutoff scores used in this study were arrived at based on those used in existing research, as well previously established clinical norms, they are still somewhat arbitrary. Evidence of this is seen in the variation in the results for analyses using cutoff vs. continuous scores. Given their arbitrary nature, it is difficult to ascertain how well the cutoffs used in the current study represent the domains being investigated. For instance, post hoc exploration of the data revealed that the number of participants who are categorized as competent varies depending on the cutoff score used on the GDI. Based on existing research on depression, the analyses for the current study utilized 20 as a cutoff score for identifying people who may be at-risk for a depressive disorder. This resulted in 88% of the sample being classified as competent. However, the literature also suggests that a score of 12 on the GDI indicates mild depression. With this is mind, the data were examined using 12 as a cutoff score. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 The results indicated that only 58% of the sample would have been categorized as competent if 12 had been implemented as the cutoff (compared with 88% when 20 was used as the cutoff). These contrasting findings highlight the arbitrariness of using cutoffs to determine whether or not an individual is competent. One could easily make the argument that a person who is even mildly depressed is not functioning at a competent level and therefore should not be classified as resilient. It seems to make more sense, then, to examine each domain on a continuum. Another consideration is that each of the domains examined in the current study are derived from single indicator measures. Additional research is warranted that attempts to replicate the findings reported here using multiple indicators. Along the same vein, it is possible that the findings of the current research are reflective of the fact that all of the outcome measures are based on self-reports. In the past researchers have noted that relying on an individual’s own self-perception does not account for over- or underrating (Kinard, 1998). Previous studies have frequently shown that children’s accounts of their own well-being and competence may differ drastically from parent or teacher reports (see Achenbach, McConaughy, & Howell, 1987 for a review). In fact, the results of a study by Spaccarelli & Kim (1995) indicate variation in outcome depending on informant identity. Similarly, in their sample of maltreated children Kaufman et al. (1994) noted that categorizing an individual as socially competent, or not, varied based on the identity of the informant. The researchers assessed social competence via self- and teacher-reports and found that self-reports of social competence yielded higher identification of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 resilience than teacher reports. Likewise, Barnett, Vondra & Shook (1996) examined discrepancies in teacher and child ratings of social competence and found that specifically for younger children (6-7 year olds), maltreated participants tended to overrate their social acceptance by peers, as compared to their teacher’s perception of their acceptance by peers. Some researchers have dealt with this problem by requiring scores in the resilient range from both parent- and child-reports in order for an individual to be classified as resilient (Kaufman et al, 1994). In the current study reports by multiple informants could have been quite enlightening, particularly with respect to academic achievement. In future research additional information could be obtained from school officials, including verification of graduation as well as transcripts indicating the individual’s grade point average during high school. For the younger portion of the sample (i.e. those who had not yet graduated from High School) it would be interesting to also assess current grade point average, school attendance, and extracurricular activities. While the majority of the participants in this study were adults, it would still be helpful to have data from other informants to verify that participants are functioning at the level that their scores on the self-report measures suggest. Future research in this area would be greatly enhanced by having data from multiple informants. In addition to gathering information from multiple informants, future longitudinal research on competent outcomes for sexually abused children could benefit from examining additional domains in order to assess competent functioning across a wider array of developmental outcomes. For example, future research on Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82 resilience should examine occupational attainment and satisfaction in adults. Whereas the current study examined graduation from high school, no data regarding whether or not the participant had attended higher education, or whether they were happy in their current position was obtained. Information about how many jobs the individual has had and how often they change jobs would also be quite telling as a means of measuring stability. It would also be interesting to examine the personal relationships that the participants are (and have been) involved in. The current study only examined social functioning in terms of ones overall perceived social competence. Examining the social network variables that were assessed at Time 1 again at Time 5 would likely provide much insight into the personal relationships of the participants, including the number of peers and family members one has in their social network, as well as whether s/he is actively involved in a romantic relationship. Assessing happiness with peers, family members, and significant others in adulthood and how that relates to other developmental outcomes in sexually abused women would be a great addition to the literature. One final consideration is that the abused participants in this study may be self-selected. That is, the girls in the abused group were required to have a nonabusing caretaker who was willing to participate in the study. Thus it is likely that the abuse group is somewhat skewed and does not represent the lower levels of functioning that might be possible in a sexually abused population. It is highly possible that girls who do not meet the criteria for resilience would not choose to participate in a study of this nature. It also seems likely that mothers who are not Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 supportive of their daughters would not have agreed to participate, which may have impacted the findings regarding social support. To summarize, research on resilient outcomes in sexual abuse victims is still in its infancy. Whereas a fair amount of research has been done on victims of child abuse and maltreatment, less has been done specifically pertaining to resilient outcomes in sexual abuse victims. There is fertile ground for future research in this area. Specifically, one area that future research could benefit from is examining the stability of competence/resilience over time. One of the basic tenets of developmental psychology is that issues in one developmental period lay the groundwork for subsequent issues. This is the reason that patterns of adaptation and maladaptation are often categorized in terms of the individual’s ability to successfully manage developmentally salient tasks (Sroufe & Rutter, 1984). In his discussion of psychosocial resilience and protective mechanisms Rutter (1990) asserts that people who cope successfully at one stage in their life might react adversely to other stressors in different situations. In addition, assessing risk and resilience across the life span is important in illustrating the role that biological, psychological and environmental factors play in fostering or hampering positive adaptation (Cicchetti & Toth, 1997). Some research on the stability of resilience over time has indicated that individuals who are classified as resilient early in life may have adjustment difficulties later in life (Farber & Egeland, 1987; Rad.ke-Yarrow & Brown, 1993; Luthar, Doemberger & Zigler, 1993; HerrenkoM, et al., 1994) whereas others have Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 reported that resilience is stable across time (Werner & Smith, 1992; Radke-Yarrow & Brown, 1993). One explanation for conflicting findings regarding the stability of resilience is that being classified as resilient might vary depending on the developmental stage an individual is at for any given assessment. It seems plausible that an individual’s designation as resilient, or not, could vary from one assessment to the next. A person may not be classified as resilient based on one assessment but may later “qualify” as resilient, indicating some sort of recovery of function. These discrepancies highlight the need for longitudinal studies in which data from multiple assessments can be evaluated in order to determine whether individuals who are classified as resilient at one stage of development are still functioning at resilient levels at subsequent assessments. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 References Aber, J.L. & Allen, IP . (1987). The effects of maltreatment on young children’s socio-emotional development: An attachment theory perspective. Developmental Psychology, 23, 406-414. Adbulrehman, R.Y., & De Luca, R.V. (2001). The implications of childhood sexual abuse on adult social behavior. Journal of Family Violence, 16, 193-203. Achenbach, T.M., McConaughy, S.H., & Howell, C.T. (1987). Child/adolescent behavioral and emotional problems: Implications o f cross-informant correlations for situation specificity. Psychological Bulletin, 101(2), 213. Adams-Tucker, C, (1982). Proximate effects of sexual abuse in childhood: A report on 28 children. American. Journal o f Psychiatry, 139, 1252-1256. Aiken, L.S. & West, S.G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Bagley, & Ramsey, R. (1986). Sexual abuse in childhood: Psychosocial outcomes and implications for social work practice. Journal of Social Work and Human Sexuality, 4, 33-47. Barnett, D,, Vondra, XL, Shook, S.M. (1996). Self-perceptions, motivation, and school functioning of low-income maltreated and comparison children. Child Abuse & Neglect, 20(5), 397-410. Beitchman, J. H., Zucker, K. X, Hood, I. E., daCosta, G. A., Akman, D., & Cassavia, E. (1992) A review of the long-term effects of child sexual abuse. Child Abuse cmdNeglect, 16, 101-118. Bernard, H.R. (1999). Social Research Methods: Qualitative and Quantitative Approaches. Sage Publications. Bogat, G.A., Chin, R , Sabbath, W., & Schwartz, C. (1985). The children’s social support questionnaire (Technical Report #3). East Lansing, MI: Michigan State University. Blyth, D.A., Hill, J.P., & Thiel, K.S. (1982). Early adolescents' significant others: Grade and gender differences in perceived relationships with familial and non-familial adults and young people. Journal o f Youth and Adolescence, 11, 425-450. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 Browne, A., & Finkelhor, D, (1986). Initial and long-term effects: A review of the research. In D. Finkelhor & Associates, A sourcebook on child sexual abuse (pp. 143-179). Beverly Hills, CA: Sage. Cicchetti, D. & Rogosch, F. (1997). The role o f self-organization in the promotion of resilience in maltreated children. Development & Psychopathology, 9, 797-815. Cicchetti, D., Rogosch, F.A., Lynch, M., & Holt, K.D. (1993). Resilience in maltreated children: Processes leading to adaptive outcome. Development & Psychopathology, 2, 629-647. Conte, J.R., & Schuerman, J R. (1987). Factors associated with an increased impact of child sexual abuse. Child Abuse & Neglect, 1 1,201-211. Dempster, A. P., Laird, N.M., & Rubin, D.B. (1977). Maximum likelihood estimation from incomplete data via the EM algorithm (with discussion). Journal o f Royal Statistical Society, B39, 1-38. Dumont, M. & Provost, M.A. (1999). Resilience in adolescents: Protective role of social support, coping strategies, self-esteem, and social activities on experience of stress and depression. Journal o f Youth and Adolescence, 28(3), 343-363. Dunn, L. & Dunn, L. (1981). M anual fo r the Peabody Picture Vocabulary Test Revised. Circle Pines, Minnesota: American Guidance Service, Inc. Eckenrode, J., Laird, M,, & Doris, J. (1993). The school performance of abused and neglected children. Developmental Psychology, 29, 53-62. Egeland, B. (1997). Mediators o f the effects o f child maltreatment on developmental adaptation in adolescence. In Cicchetti, D. & Toth, S.L. (Eds.) Developmental perspectives on trauma: Theory, research, and intervention. Rochester symposium on developmental psychology, Vol. 8. (pp. 404-434). Rochester, NY, USA: University of Rochester Press, xvii, 613 pp. Egeland, B., Carlson, E., & Sroufe, L.A. (1993). Resilience as process. Development & Psychopathology, 5, 517-528. Einbender, A.J., & Friedrich, W.M. (1989). Psychological functioning and behavior of sexually abused girls. Journal o f Consulting and Clinical Psychology, 57, 155-157. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 Elliot, A.N. & Carnes, C.N. (2001). Reactions of nonoffending parents to the sexual abuse of their child: A review of the literature. Child Maltreatment, 6(4), 314-331. Erickson, M.F., Egeland, B. & Plant, R. (1989). The effects of maltreatment on the development of young children. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and research on the causes and consequences o f child abuse and neglect. Cambridge: Cambridge University Press. Faxber & Egeland (1987). Invulnerability among abuse and neglected children. In E.J. Anthony & B J. Cohler (Eds.) The invulnerable child. New York: Guilford Press. Feinauer, L.L. & Stuart, D.A. (1996). Blame and resilience in women sexually abused as children. The American Journal o f Family Therapy, 24, 31-40. Feiring, C., Taska, L.S., & Lewis, M. (1998). Social support and children’s and adolescents’ adaptation to sexual abuse. Journal o f Interpersonal Violence, 13(2), 240-260. Forehand, R., Middleton, K. & Long, N. (1987). Adolescent functioning as a consequence of recent parental divorce and the parent-adolescent relationship. Journal o f Applied Developmental Psychology, 8, 305-315. Friedrich, W. N., Beilke, R. L., & Urquiza, A. I. (1987). Children from sexually abusive families: A behavioral comparison. Journal o f Interpersonal Violence, 2, 391-402. Gobble, P.A., Cowan, E.L., Wyman, P.A., Work, W.C., Wannon, M., & Raoof, A. (1993). Parent and child views of parent-child relationship qualities and resilient outcomes among urban children. Journal o f Child Psychology and Psychiatry, 34, 507-520. Harter, S. (1982). The perceived competence scale for children. Child Development, 53-87-97. Harter, S. (1985). M anualfor the Self-Perception Profile fo r Children. University of Denver, Denver, CO. Harter, S. (1988). M anualfor the self-perception profile fo r adolescents. University of Denver. Harter, S., Alexander, P C., Neimeyer, R.A. (1988). Long-term effects of incestuous child abuse in college women: Social adjustment, social cognition, and family characteristics. Journal o f Consulting and Clinical Psychology, 5-8. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 Heller, S.S., Larrieu, J.A., D'Imperio, R., Boris, N.W. (1999). Research on resilience to child maltreatment: Empirical considerations. Child Abuse & Neglect, 23(4), 321-338. Herman-Stahl, M. & Petersen, A.C. (1996). The protective role of coping and social resources for depressive symptoms among young adolescents. Journal of Youth and Adolescence, 25(6), 733-753. Herrenkohl, E.C., Herrenkohl, R.C., Egolf, B. (1994). Resilient early school-age children from maltreating homes: Outcomes in late adolescence. American Journal o f Orthopsychiatry, 64(2), 301-309. Himelein, M.J. (1995). Risk factors for sexual victimization in dating. Psychology o f Women Quarterly, 19(1), 30 Himelein, M.J. & McElrath, J.V. (1996). Resilient child sexual abuse survivors: Cognitive coping and illusion. Child Abuse & Neglect, 20, 747-758. Hollingshead, A.F. (1975). Four Factor Index of Social Status. New Have, CT: Yale University, Department of Sociology. Hunter, J.A. (1991). A comparison o f the psychosocial maladjustment of adult males and females sexually molested as children. Journal o f Interpersonal Violence, 6, 205-217. Hyman, B. & Williams, L. (2001). Resilience among women survivors of child sexual abuse. Affilia, 16(2), 198-219. Kamsner, S. & McCabe, M.P. (2000). The relationship between adult psychological adjustment and childhood sexual abuse, childhood physical abuse, and family-of-origin characteristics. Journal of Interpersonal Violence, 15, 1243- 1261. Kaufman, J., Cook, A., Amy, L., Jones, B., & Pittinsky, T. (1994). Problems defining resiliency: Illustrations from the study o f maltreated children. Development & Psychopathology, 6, 215-229. Kendall-T ackett, K.A., Williams, L.M., & Finkelhor, D. (1993). Impact of sexual abuse on children: A review and synthesis of recent empirical studies. Psychological Bulletin, 113, 164-180. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 Kinard, E.M. (1995). Perceived social support and competence in abused children: A longitudinal perspective. Journal o f Family Violence, 10, 73-99. Kovacs, M. (1981). Rating scales to assess depression in school-age children. Acta Paedopsychiatrica, 46, 305-315. Kovacs, M. (1985). The Children’s Depression Inventory. Psychopharmacology Bulletin, 21, 995-998. Kovacs, M. (1992). Child Depression Inventory M anual Toronto: Multi-Health System. Lam, I , & Grossman, F.K. (1997). Resiliency and adult adaptation in women with and without self-reported histories of childhood sexual abuse. Journal o f Traumatic Stress, 10(2), 175-196. Liem, J.H., James, J.B., O’Toole, J.G., Boudewyn, A.C. (1997). Assessing resilience in adults with histories of childhood sexual abuse. American Journal o f Orthopsychiatry, 67(4), 594-606. Lipovsky, J. A., Saunders, B.E. & Murphy, S.M. (1989). Depression, anxiety, and behavior problems among victims of father-child sexual assault and nonabused siblings. Journal o f Interpersonal Violence, 4(4). Luthar, S. (1991). Vulnerability and resilience: A study of high-risk adolescents. Child Development, 62, 600-616. Luthar, S.S., Cicchetti, D., & Becker, B. (2000). The construct of resilience: A critical evaluation and guidelines for future work. Child Development, 71(3), 543-562. Luthar, S.S., Doemberger, C.H., & Zigler, E. (1993). Resilience is not a unidimensional construct: Insights from a prospective study of inner-city adolescents. Development & Psychopathology, 5(4), 703-717. Lynskey, M.T., & Fergusson, D.M. (1997). Factors protecting against the development of adjustment difficulties in young adults exposed to childhood sexual abuse. Child Abuse & Neglect, 21(2), 1177-1190. Mannarino, A. P., Cohen, J. A , Smith, J. A., & Moore-Motily (1991). Six- and twelve-month follow-up of sexually abused girls. Journal o f Interpersonal Violence, 6, 494-511. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 Masten, A.S., Hubbard, 11, Gest, S.D., Tellegen, A., Garmezy, N., & Ramirez, M. ( 1 9 9 9 ) Competence in the context of adversity: Pathways to resilience and maladaptation from childhood to late adolescence. Development & Psychopathology, 11, 143-169. Masten, A., & Coatsworth, J.D. (1998). The development of competence in favorable and unfavorable environments: Lessons from research on successful children. American Psychologist, 53(2), 205-220. McClellan, 1, McCurry, C., Ronnei, M., Adams, 1, Eisner, A., & Storck, M. (1996). Age of onset of sexual abuse: Relationship to sexually inappropriate behaviors. Journal o f the American Academy o f Child and Adolescent Psychiatry, 34(10), 1375-1383. McGloin, J.M. & Widom, C.S. (2001). Resilience among abused and neglected children grown up. Development & Psychopathology, 13, 1021-1038. McGrew, K.S. (1994). Clinical interpretations o f the Woodcock-Johnson tests o f cognitive ability-revised. Boston: Aliya and Bacon. McLeer, S.V., Deblinger, E.B., Henry, D., Orvaschel, H. (1992). Sexually abused children at high risk for post-traumatic stress disorder. Journal o f the American Academy o f Child and Adolescent Psychiatry, 131, 875-879. Mennen F, Meadow D. (1995). The relationship of abuse characteristics and symptoms in sexually abused girls. Journal o f Interpersonal Violence, 10, 250 Moran, P.B. & Eckenrode, J. (1992). Protective personality characteristics among adolescent victims of maltreatment. Child Abuse & Neglect, 16, 743-754. Mullen, P., Martin, 1, Anderson, J., Romans, S., & Herbison, G. (1996). The long - term impact of the physical, emotional, and sexual abuse of children: A community study. Child Abuse & Neglect, 20, 7-20. Nash, M.R., Zivney, O.A., & Hulsey, T. (1993). Characteristics of sexual abuse associated with greater psychological impairment among children. Child Abuse & Neglect, 17(3), 401-408. National Center of Child Abuse and Neglect (NCCAN) (1988). Study of national incidence and prevalence of child abuse and neglect, 1988. Washington, D. C.: U. S. Department of Health and Human Services. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 National Center o f Child Abuse and Neglect (NCCAN) (1996). Third study of national incidence and prevalence of child abuse and neglect, (preliminary findings). Washington, DC: US Department of Health and Human Services. Neighbors, B., Forehand, R., & McVicar, D. (1993). Resilient adolescents and interparenta! conflict. American Journal o f Orthopsychiatry, 63(3), 462-471. Noli, J.G., Trickett, P.K., & Putnam, F.W. (2002, in press). A prospective investigation of the impact of childhood sexual abuse on sexuality in adolescence and early childhood. Journal o f Consulting and Clinical Psychology. O’Grady, D. & Metz, J.R. (1987). Resilience in children at high risk for psychological disorder. Journal o f Pediatric Psychology, 12(1), 3-23. Orchard, T., & Woodbury, M.A. (1972). A missing information principle: Theory and applications. In Proceedings o f the 6th Berkeley Symposium on Mathematical Statistics and Probability, 1 (pp. 697-629). Washington, DC: University of California Press. Perez, C. & Widom, C. (1994). Childhood victimization and long-term intellectual and academic outcomes. Child Abuse and Neglect, 18, 617-633. Perkins, D. F., & Luster, T. (1999). The relationship between sexual abuse and purging: Findings from community-wide surveys of female adolescents. Child Abuse and Neglect, 23, 371-382. Putnam, F.W. & Trickett, P.K. (1987). The psychobiological effects o f child sexual abuse (protocol #87-M-14). Washington, DC: National Institutes of Mental Health. Radke-Yarrow, M., & Brown, E. (1993). Resilience and vulnerability in children of multiple-risk families. Development & Psychopathology, 5, 581-592. Rowe, E. & Eckenrode, J. (1999). The timing of academic difficulties among maltreated and nonmaltreated children. Child Abuse & Neglect, 23(8), 813- 832. Russell, D.E.H. (1986). The secret trauma: Incest in the lives o f girls and women. New York: Basic Books. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 Saylor, C.F., Finch, A.I., Jr., Spirito, A., & Bennett, B. (1984). The children’s depression inventory; A systematic evaluation of psychometric properties. Journal o f Consulting and Clinical Psychology, 52, 955-967. Sedney. M. A., & Brooks, B. (1984). Factors associated with a history of childhood sexual experience in a nonclinical female population. Journal o f the American Academy o f Child Psychiatry, 2 3 ,215-218. Sirles, E.A., Smith, J.A., & Kusama, H. (1989). Psychiatric status of intrafamilial child sexual abuse victims. Journal o f the American Academy o f Child and Adolescent Psychiatry, 28(2), 225-229. Sitarenios, G. & Kovacs, M. (1999). Use o f the children ’ s depression inventory. In M.E. Maruish (Ed.) The use of psychological testing for treatment planning and outcomes assessment. Mahwah, NJ: Lawrence Erlbaum Associates, Publishers. Sloan, V.J., Jason, L.A., Bogat, G.A. (1984). A comparison of orientation methods for elementary school transfer students. Child Study Journal, 14(1), 47-60. Smucker, M.R., Craighead, W.E., & Craighead, L.W. (1986). Normative and reliability data for the Children’s Depression Inventory. Journal o f Abnormal Child Psychology, 14, 25-39. Spaccarelli, S. & Kim, S. (1995). Resilience criteria and factors associated with resilience in sexually abused girls. Child Abuse & Neglect, 19(9), 1171- 1182. Spaccarelli, S. (1994). Stress, Appraisal, and coping in child sexual abuse: A theoretical and empirical review. Psychological Bulletin, 116(2), 340-362. Sroufe, L.A. & Rutter, M. (1984). The domain of developmental psychopathology. Child Development, 55, 17-29. Tong, L., Oates, K., & McDowell, M. (1987). Personality development following sexual abuse. Child Abuse and Neglect, 11, 371-383. Tremblay, €., Hebert, M., & Piche, C. (1999). Coping strategies and social support as mediators of consequences in child sexual abuse victims. Child Abuse & Neglect, 23(9), 929 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 Trickett, P.K. & McBride-Chang, C. (1995). The developmental impact of different forms of child abuse and neglect. Developmental Review, 15, 311-337. Trickett, P.K., McBride-Chang, C. & Putnam, F.W. (1994). The classroom performance and behavior o f sexually abused females. Development and Psychopathology, 6, 183-194. Trickett, P.K., Noll, J.G., Reiffman, A., & Putnam, F.W. (2001). Variants in intrafamilial sexual abuse experience: Implications for short- and long-term development. Development and Psychopathology, 1 3 ,1001-1019. Werner, E.E. & Smith, R.S. (1992). Overcoming the odds: High-risk children from birth to adulthood. Cornell University Press, Ithaca, NY. Whiffen, V.E., Thompson, I.M. & Aube, I. A. (2000). Mediators of the link between childhood sexual abuse and adult depressive symptoms. Journal o f Interpersonal Violence, 15, 1110-1120. Widom, C.S. (1989). Does violence beget violence? A critical examination of the literature. Psychological Bulletin, 106(1), 3-28. Woodcock, R. W., & Johnson, M. B. (1989). Woodcock-Johnson Psychoeducational Battery-Revised. Allen, TX: DLM Teaching Resources. Wyatt, G.E. & Powell, G.J., Eds. (1998). Lasting effects o f child sexual abuse. Beverly Hills, CA: Sage. Zingraff, M.T., Leiter, J., Myers, K.A., Johnsen, M.C. (1993). Child maltreatment and youthful problem behavior. Criminology, 31(2), 173-202. Zuravin, S.J. & Fontanella, C. (1999). The relationship child sexual abuse and major depression among low-income women: A function of growing up experiences? Child Maltreatment, 4, 3-12. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix Intercorrelation tables Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table A1. Intercorrelations of demographic and predictor variables for entire sample 1 2 3 4 5 6 7 8 1. Age 2.SES -.004 3. Minority status ,144t -.200* 4. Abuse status -.090 .055 .119 5. Males in network .004 ,157f ,151T .027 6. Females in network .095 .218* .106 .096 .360** 7. Happiness w/network -.137 -.048 .021 .077 .108 .077 8. Closeness w/mother -,172t 4* -.170' -.071 -.079 .023 -.104 .174* 9.PPVT -.011 .259** .329** .071 -.096 .076 .042 -.129 *p< 05; **p< 01; fp<10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table A2. Intercorrelations of demographic variables, predictor variables and characteristics of abuse 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Age 2. SES .033 3. Minority status .271* -.237* 4. Males in network .051 .178 .126 5. Females in network .217* .225* .216* .352** 6. Happiness w/network -.139 -.097 -.127 -.003 -.039 7. Closeness w/mother -.032 -.130 -.100 -.046 -.054 .276* 8. PPVT -.095 .340** -.381** -.029 .140 .135 -.096 9. Severity o f abuse .017 .149 -.109 .144 -.132 .059 -.115 .137 10. A ge at onset .702** -.134 .514** .082 .329** -.147 -.135 -.152 -.231* 11. Duration .179 .284* -.346** .039 -.175 .081 .171 .130 .266* ..444** 12. Perpetrator (bio father vs. .066 -.125 -.175 -.080 -.161 -.131 -.026 -.091 .093 -.204* .371** other relative) 13. Physical violence -.061 -.327** .144 -.042 -.060 .107 .432** -.340** .099 -.055 .001 .068 14. Multiple perpetrators -.126 -.195* -.086 -.089 -.174 .225* .024 -.194 .164 -.113 -.128 -.023 .299** *p<.05; **p<.01; ^pc.lO
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Anorexia nervosa: A developmental perspective
PDF
Dyslexic's development of word recognition and phonological skill over time
PDF
Hostile family environments and children's perceptions of social support
PDF
Effects of parenting style and ethnic identity on European American and Asian Indian adolescents' academic competence and self esteem
PDF
Alcohol treatment entry and refusal in a sample of older veterans
PDF
Distributional cues and subject identification in the production of subject -verb agreement
PDF
Family functioning, intergenerational conflict, and psychological symptomology of Asian American college students
PDF
Does first-year treatment intensity predict outcome in young autistic children receiving Lovaas' ABA intervention?
PDF
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
PDF
Gender differences in motivation for sexual intercourse: Implications for risky sexual behavior and substance use in a university and community sample
PDF
Ethnic identity in Parsee teenagers
PDF
"Insane passions": Psychosis and female same -sex desire in psychoanalysis and literary modernism
PDF
Gender and sexuality in a grave new world: Feminist ethics, the body and gothic sensibilities
PDF
Comparative analysis of the stress process in AIDS caregivers: Gay male partners, Anglo women, and Latinas
PDF
An integrative analysis of racism and quality of life: A comparison of multidimensional moderators in an ethnically diverse sample
PDF
Dementia caregiving and ethnicity: African American caregivers and the sociocultural stress and coping model
PDF
An integrative model of subjective well-being: Culture, personality, and demographics
PDF
Emotion regulation as a mediator between family-of-origin aggression and marital aggression
PDF
Atypicality: Benefit and bane. Provocation, trigger, typicality, and the expression of aggression and generalization
PDF
A phenomenological inquiry into the essential meanings of the most intense experience of religiosity and spirituality
Asset Metadata
Creator
Kurtz, Dawn A.
(author)
Core Title
Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,Psychology, clinical,psychology, developmental,women's studies
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Manis, Franklin (
committee chair
), [illegible] (
committee member
), Trickett, Penelope K. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-370389
Unique identifier
UC11334973
Identifier
3103923.pdf (filename),usctheses-c16-370389 (legacy record id)
Legacy Identifier
3103923.pdf
Dmrecord
370389
Document Type
Dissertation
Rights
Kurtz, Dawn A.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
psychology, developmental
women's studies