|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001919537
007 cr mnu|||uuuuu
008 071218s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002076
Psychological consequences of child sexual abuse and the risk and protective factors influencing these consequences
h [electronic resource] /
by Kelli-Lee Harford.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Although a number of negative consequences of childhood sexual abuse (CSA) have been identified, research has shown that some survivors of CSA are fairly resilient and do not demonstrate these negative outcomes. The current study examined differences between sexually abused and non-abused children on a number of emotional and behavioral dimensions and on achievement. In addition, the role of factors such as intelligence, abuse severity, gender, history of previous psychological interventions and number of foster homes on outcomes in a group of 117 children between the ages of 7 and 16 with sexual abuse histories and 80 controls who did not have a reported history of sexual abuse was examined. Results suggested that children with CSA histories were rated by their caregivers as exhibiting significantly more overall behavior problems than children without CSA histories. CSA history was not found to be significantly associated with self reported depressive symptoms and there was not a significant relationship between gender and caregiver reported behavior problems. However, consistent with expectations, CSA history was significantly associated with intelligence and higher levels of intelligence being predictive of better functioning in a number of areas. Similarly, CSA history was significantly associated with achievement and as expected, higher levels of intelligence were significantly associated with higher overall achievement. Number of foster care placements, abuse severity, history of previous psychological treatment and age at time of testing were generally not found to be significantly associated with resilience.
Dissertation (Ph.D.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 71 pages.
Advisor: Bill Kinder, Ph.D.
t USF Electronic Theses and Dissertations.
Psychological Consequences of Child Sexual Abuse and the Risk and Protective Factors Influencing These Consequences by Kelli-Lee Harford A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Bill Kinder, Ph.D. Katurah Jenkins-Hall, Ph.D. Jon Rottenberg, Ph.D. Marcia Finkelstein, Ph.D. Toru Shimizu, Ph.D. Date of Approval: May 31, 2007 Keywords: resilience, intelligence, severity, foster homes, achieveme nt Copyright 2007, Kelli-Lee Harford
Acknowledgements I would like to thank my advisor, Dr. Bill Kinder, for his advice and suppor t during my time at USF, and for reviewing countless versions of various manuscript s over the years. I would also like to convey my gratitude to my committee members, D rs. Marcia Finkelstein, Katurah-Jenkins Hall, Jonathan Rottenberg, and Toru Shimi zu as well as Dr. David Drobes for their invaluable suggestions and encouragement whil e completing this dissertation. Finally, I would like to express my appreciation to Todd and my family for helping provide balance and support throughout my graduate training.
i Table of Contents List of Tables ii List of Figures iii Abstract v Chapter 1 Introduction 1 Prevalence 2 Definition of Sexual Abuse 4 Psychological Consequences and Behaviors 5 Cognitive Dysfunction 9 Severity 11 Foster Care Placements as a Risk Factor 14 Resilience 16 Intelligence as a Protective Factor 17 Chapter 2 The Present Study 22 Chapter 3 Hypotheses 24 Chapter 4 Method 25 Measures 25 Procedure 29 Participants 30 Chapter 5 Results 33 Chapter 6 Discussion 52 References 61 About the Author End Page
ii List of Tables Table 1 Locations from which participants were drawn 31 Table 2 Independent Samples T-Test between Sexually Abused and Non-Sexually Children on Demographics, IQ, CDI and Achievement 34 Table 3 Independent Samples T-Test between Sexually Abused and Non-Sexually Children on CBCL Variables 35 Table 4 Independent Samples T-Test between Sexually Abused Boys and Girls on CBCL variables 37 Table 5 Correlations between Full Scale IQ, Number of Homes, Severity of Abuse, History of Psychological Treatment and Total CDI Score 39 Table 6 Correlations between Full Scale IQ, Number of Homes, Severity of Abuse, History of Psychological Treatment and CBCL Scores 40 Table 7 Correlations between Full Scale IQ, Number of Homes, Severity of Abuse, History of Psychological Treatment and Achievement Scores 43 Table 8 Correlations between Full Scale IQ, Number of Homes, CDI, Achievement and CBCL Scores for the Sexually Abused Group 44
iii List of Figures Figure 1 Moderating Effect of IQ on Abuse and Total Achievement 49 Figure 2 Moderating Effect of IQ on Abuse and CBCL Internalizing 50
Psychological Consequences of Child Sexual Abuse and the Risk and Resilience Factor s Influencing These Consequences Kelli-Lee Harford ABSTRACT Although a number of negative consequences of childhood sexual abuse (CSA) have been identified, research has shown that some survivors of CSA are fairly resi lient and do not demonstrate these negative outcomes. The current study examined differenc es between sexually abused and non-abused children on a number of emotional and behavioral dimensions and on achievement. In addition, the role of factors such as intelligence, abuse severity, gender, history of previous psychological inter ventions and number of foster homes on outcomes in a group of 117 children between the ages of 7 and 16 with sexual abuse histories and 80 controls who did not have a reported history of sexual abuse was examined. Results suggested that children with CSA histories were rated by their caregivers as exhibiting significantly more overall behavior problems than children without CSA histories. CSA history was not found to be significantly associ ated with self reported depressive symptoms and there was not a significant relat ionship between gender and caregiver reported behavior problems. However, consistent wit h expectations, CSA history was significantly associated with intellig ence and higher levels of intelligence being predictive of better functioning in a number of areas. Simi larly, CSA history was significantly associated with achievement and as expe cted, higher levels
of intelligence were significantly associated with higher overall ac hievement. Number of foster care placements, abuse severity, history of previous psychological tre atment and age at time of testing were generally not found to be significantly associat ed with resilience.
Risk and Protective Factors in Child Sexual Abuse 1 Psychological Consequences of Child Sexual Abuse and the Risk and Protective Factors Influencing these Consequences Introduction Researchers have identified a number of psychological and behavioral consequences in survivors of child sexual abuse such as posttraumatic stress disorder a nd eating disorders (Browne & Finkelhor, 1986). Factors such as the severity of the abuse experience (Kendall-Tackett, Williams & Finkelhor, 1993) and cognitive functioning (Trickett, McBride-Chang & Putnam, 1994) have been identified as putting an individual at greater risk of developing these negative consequences. In addition, since ma ny children who have been sexually abused are removed from their home and placed into foster care, the impact of the number of foster care placements on sexual abuse s urvivors, therefore, is an important factor to consider when discussing risk factors in survi vors of CSA. Stanley, Riordan and Alaszewski (2005) suggested that a high number of foster placements were related to mental health problems in the children in their sample Research has also indicated, however, that not all survivors of CSA experience these negative consequences (Spaccarelli & Kim, 1995). Factors such as intelligence ha ve been purported to act as protective factors in high risk children (Luthar, Zigler, & Gol dstein, 1992) but have only been examined to a limited degree in individuals who were sexually abused as children.
Risk and Protective Factors in Child Sexual Abuse 2 Prevalence In 1984, the Second National Incidence and Prevalence Study of Child Abuse and Neglect was mandated by Congress. This study demonstrated that girls were significantly more likely to be victims of sexual abuse than boys. The incidence rate of sexual abuse among girls was 3.28 per 1000, while for boys it was 1.00 per 1000. The study also reported that low income was a significant risk factor for abuse in g eneral, both physical and sexual, but there was a greater gender difference for lower income families, with girls in low income families at greater risk for sexual abuse rather than physical abuse (Cappelleri, Eckenrode, & Powers, 1993). Putnam (2003) suggested that the majority of the research reviewed has been retrospectively collected from adult samples, and that prevalence rates vary widely as a function of the selection and response rate, the definition used, and the method by which the history is obtained (p. 270). Putnam (2003) reported that based on statistics from the U.S. Department of Health and Human Services (1998), before the 1970s the incidence of CSA was rare. These statistics should be interpreted with care, however, bas ed on low rates of reporting any type of abuse, especially sexual abuse during this ti me period. There was then an increase to about 149,800 substantiated or indicated cases in 1992 and as of the most recent statistics in 2000, this number has decreased to about 88,000. Putnam (2003) reported that studies in community samples indicate that prevalence ra tes are usually between 12% and 35% in female samples and between 4% to 9% in male samples. Gorey and Leslie (1997) found the prevalence rates of CSA in women to be
Risk and Protective Factors in Child Sexual Abuse 3 16.8% and for men to be 7.9% after adjusting for factors such as sample related variat ion, response rates and differences in definition. Loeb, Williams, Carmona, Rivkin, Wyatt, Chin and Asuan-O'Brien (2002) suggested that prevalence rates of CSA do not appear to differ among ethnic groups. In addition, research from other countries indicates that prevalence rates arou nd the world are comparable to those in the United States. Finkelhor (1994) conducted a review of studies from 19 countries and found that the rates of CSA in women were between 7% and 36% and in men were between 3% and 29%. Some researchers have suggested that the prevalence rates may be higher because many individuals who have been abused do not report this abuse (Dimock, 1988; Finkelhor, 1994). This may be especially significant in the case of male victims of CSA. Watkins and Bentovim (1992) reviewed research on male children and adolescents and found that males were less likely than female s to report the abuse to authorities. Males may be less likely to report the abuse bec ause they may accept greater blame for the abuse (Myers, 1989); there is greate r stigma attached to the abuse (Dimock, 1988); because of fear of being labeled as homosexual (Black & DeBlassie, 1993); they may be less likely to view the sexual activity as abuse, especially if the perpetrator was female (Dhaliwal, Guazas, Antonowicz, & Ross, 1996); and the y may feel that their reports will be taken less seriously by the authoritie s (King & Woolett, 1997). Lab, Feigenbaum and De Silva (2000) found that mental health professionals are less likely to ask males about histories of CSA.
Risk and Protective Factors in Child Sexual Abuse 4 Definition of sexual abuse Sexual abuse has been defined in a variety of ways by different researchers The definitions of sexual abuse vary within the sexual abuse literature from no contact events such as exhibitionism to fondling and sexual intercourse (Rumstein-McKean & Hunsl ey, 2001). One of the most widely used definitions is that of Finkelhor (1979) whose definition includes sexual activity between a child and an older person, including simulated, attempted or actual intercourse, kissing, hugging or fondling in a sexua l manner, sexual overtures and exhibitionism. This contact was described as sex ual abuse if it occurred between a child 12 or under and an adult over 18; or more than 5 years older than the child; or between an adolescent and an adult at least 10 or more years ol der than the adolescent. Russell (1986) defines sexual abuse as any sexualized behavior between a minor child and anyone who is 5 years older than the child. Bartoi and Kinder (1998) defined sexual abuse as oral, vaginal or anal intercourse, or genital mani pulation with someone at least 5 years older, being touched in a way that made the individual feel violated or being coerced into unwanted sexual activity, or had ever been touched in a way that made them feel violated. Bartoi and Kinder (1998) developed a brief measur e using this definition that has been used in a number of studies. The definition that will be used in the current study is based on the definition used by the Department of Children and Family Services (DCF) which is found in Chapte r 39 of the Florida Statutes. Florida state statute 39.01 defines sexual abuse as involvi ng one or more of the following by an adult to a minor child : penetration; any sexual conta ct not intended for medical purposes or as normal caregiver responsibility; masturbati on in the
Risk and Protective Factors in Child Sexual Abuse 5 presence of a child; exposure of genitals for the purpose of sexual arousal or gra tification, aggression, degradation, or for a similar purpose; sexual exploitation including encouraging, or forcing a child to solicit for or engage in prostitution, or engage in a sexual performance. Florida statues define a minor child as a child below 18 year s old. Sexual abuse can also be committed by a juvenile whereby the sexual behavior occ urs without consent, without equality, or as a result of coercion. DCF determines t hat a child has been sexually abused if there is substantive evidence that the abuse occur red. The results of the DCF investigations are classified as no indicators of abuse, some indicators of abuse, or verified abuse. After the investigation has been classi fied into one of these categories, the child may be adjudicated dependent and placed in the care of DCF. Research has demonstrated many negative consequences of CSA. The outcomes found by each study are often dependent on the particular definition of sexual abus e used by the researchers. Negative symptomatology can be viewed as internaliz ing and/or as externalizing and many sexual abuse survivors may demonstrate characteri stics of a variety of these problems. Psychological Consequences and Behaviors Psychological effects such as depression, anxiety, fear, distress, guilt a nd shame have been associated with women who have been sexually abused (Browne & Finkelhor, 1986; Chaffin, Silovsky, & Vaughn, 2005; Johnson & Kenkel, 1991; Kendall-Tackett, Williams, Finkelhor, 1993; Paolucci, Genuis, & Violato, 2001; Saywitz, Mannarino, Berliner & Cohen, 2000; Spaccarelli & Fuchs, 1997). Some behaviors that have been
Risk and Protective Factors in Child Sexual Abuse 6 identified by researchers as possibly stemming from abuse are aggre ssion, oversexualized behavior, eating disorders, substance abuse, self injurious behaviors, soma tic complaints, dissociation, sexual perpetration, academic difficulties, interpe rsonal difficulties, and suicidality (Browne & Finkelhor 1986; Inderbitzen-Pisar uk, Shawchuck & Hoier 1992; Kendall-Tackett, Williams, Finkelhor 1993; Martin, Bergen, Richardson, Roeger & Allison, 2004; Monahan & Forgash, 2000; Newman, Clayton, Zuellig, Cashman, Arnow, Dea, & Taylor, 2000; Paolucci, Genuis, & Violato, 2001; Saywitz, Mannarino, Berliner & Cohen, 2000; Smith M.S. & Smith M.T., 1999; Spaccarelli & Fuchs, 1997). Martin et al (2004) found that boys with histories of CSA were more likely to exhibit suicidality which included suicidal ideation, plans, threats, delibera te self injury and suicide attempts than girls with histories of CSA when controlling for hopele ssness, depression and family functioning. They found that girls suicidality was bette r accounted for by depression, hopelessness, and family functioning while boys were more likely t o demonstrate suicidality even after controlling for these factors. Garnefski and Arends (1998) examined emotional and behavioral problems and suicidality in boys and girls with CSA histories and compared them with matche d controls that had not experienced CSA. Emotional problems assessed were loneliness, anxiety, depressed mood and self esteem and were measured by subscales on an emotional problems scale. Behavioral problems assessed included use of alcohol, aggressive behavior, criminal behavior, use of drugs and truancy and were assessed by subscales on a behavioral problems scale. They found that boys and girls with histor ies of
Risk and Protective Factors in Child Sexual Abuse 7 CSA reported significantly more emotional and behavioral problems and suicidal problems than matched controls. In comparing boys and girls with histories of CS A, there were no significant differences with regard to overall levels of emotional proble ms being reported, although boys reported significantly more loneliness than girls and gi rls reported significantly more feelings of anxiety than boys. Boys with hist ories of CSA had significantly more overall behavior problems than girls with histories of CSA including significantly more use of alcohol and drugs, suicidality, aggressive behavior a nd criminal behavior, and truancy. Some researchers have found that the circumstances surrounding CSA, the way that the abuse is processed and the effects of the abuse may differ by ethnic group. For example, Wyatt (1990) reported that while there were few differences betw een African American and Caucasian women with regard to their initial response to CSA, the s hort term impact of the abuse, as well as many of the long term consequences, Afri can American women were more likely than Caucasian women to avoid men resembling the perpetrator. This research is still in its infancy and more research needs t o be conducted to examine whether the effects of sexual abuse is different for males and fe males in different cultures, as the majority of research examining cultural fac tors to date has looked these factors with regard to women. Weinstein, Staffelbach, and Biaggio (2000) in their review of the literature on PTSD and Attention Deficit Hyperactivity Disorder (ADHD) in individuals wi th histories of CSA reported that while PTSD is the diagnosis most often given, ADHD is the second most common. The authors suggested that it may be difficult to make a differential
Risk and Protective Factors in Child Sexual Abuse 8 diagnosis because PTSD and ADHD may share similar symptomatology such as problems concentrating and impulsivity. In addition, PTSD and ADHD may truly co-occur in some individuals. Kisiel and Lyons (2001) reported that dissociation, described as a natural, protective response to stress can, with prolonged stress become an automatic res ponse to stress. They also suggested that aggressive, risk taking behaviors can occur i n the context of dissociative experiences when individuals feel out of control and forced to do something against their will. They examined the hypothesis that dissociati on mediated the relationship between abuse history and psychopathology among children and adolescents in a state run residential facility and found that sexual abuse was significa ntly associated with dissociation. Dissociation was measured using the Adolescent Dissociati ve Experiences Scale (Armstrong, Putnam, Carlson, Libero & Smith; 1997) and the Child Dissociative Checklist (Putnam, Helmers, Horowitz & Trickett; 1993). Furthe r, they suggested that their results indicated that dissociation may mediate the rel ationship between sexual abuse and psychiatric symptoms and risk-taking behavior. Kaplow, Hall, Koenen, Dodge, and Amaya-Jackson, (2005) examined attention problems in sexually abused children. They found that dissociation immediately following disclosure of abuse was the strongest predictor for attention problem s. Dissociation was assessed by the Trauma Symptom Checklist for Children (TSC C; Briere, 1996). They suggested that this may be because dissociation interfered wi th the childs perception of their environment and the integration of incoming information, which may in turn lead to concentration problems. In addition, they hypothesized that
Risk and Protective Factors in Child Sexual Abuse 9 dissociation has been associated with less connectivity in the corpus callosum whic h may lead to the attention and information processing problems seen in children who have been sexually abused. The authors found that PTSD was only related to attention problems by way of dissociation. Kaplow et al (2005) also found that being abused by a family member was also predictive of attention problems. This study did not include a control group of children who had not been sexually abused. Cognitive Dysfunction Trickett, McBride-Chang and Putnam (1994) suggested that the psychological consequences that have long been identified as stemming from sexual abuse likely a ffect cognitive functioning in those who have been abused. They suggested that because abused children are distracted by their abuse experience, this takes away cognitive energy from other pursuits. In addition, they suggest that the abused childs internal motivation to please significant others which can provide the motivation needed to succeed in academic pursuits may be diminished because it is likely that these or other s ignificant individuals in the childs life are perpetrating the sexual abuse. Additionally, because of low self esteem, abused children may be less inclined to participate in and learn f rom challenging experiences. Other researchers such as Nakano, et al (2002) hav e found that prolonged stress can result in structural brain changes which can then impact m emory functioning. Nakano et al (2002) examined hippocampal volume in cancer survivors. They found that in cancer survivors with a history of distressing cancer relate d memories, the volume of the left hippocampus was 5% smaller than in cancer survivors without a history of distressing memories. It may be that prolonged stress experienc ed by sexual
Risk and Protective Factors in Child Sexual Abuse 10 abuse survivors may result in changes in the structure of the brain, which in turn affect s cognitive functioning. Bremner, Randall, Scott, Capelli, Delaney, McCarthy and Charney (1995) examined memory and intellectual functioning in adult survivors of severe childhood physical and sexual abuse. They found that while there were no significant differ ences between adults with abuse histories and controls with regard to IQ, the survivors of abus e had deficits in verbal short term recall, as well as immediate and delayed re call. There is some evidence to suggest, however, that there is a relationship specifically between sexual abuse and cognitive dysfunction in those who have been abused. Trickett et al (1994) found that, in comparison to controls, sexually abused girls did not do as well in classroom social competence and overall academic performance. In addition, participants in the sexually abused sample had higher levels of school avoidant behavior than controls. They also found that sexual abuse was negatively related to cognitive ability as measured by the Peabody Picture Vocabulary Test Revis ed (PPVTR), but that sexual abuse history was not related to grades. It should be noted, however, that the PPVT-R has been criticized on the grounds that it is heavily dependent on environmental exposure to the test stimuli. In their review of the literature, Veltman and Browne (2001) suggested that the research on the effect of sexual abuse on intellectual ability and academi c functioning is mixed, with some research suggesting a link between sexual abuse and lower intel lectual and/or verbal ability while other research does not find such a link. This link seems to be more tentative when sexual abuse is distinguished from physical abuse. For exampl e,
Risk and Protective Factors in Child Sexual Abuse 11 Perez and Widom (1994) found that while abused and neglected individuals had significantly lower IQ scores than controls, when the sample of those who had been sexually abused only was separated from their larger sample, the results indic ated that young adults who were sexually abused as children did not differ from controls with regard to IQ and reading ability. Harford and Kinder (2004) also found that there we re no significant differences between women with (M = 102.74, SD = 8.14) and without (M = 101.85, SD = 7.57) histories of CSA with regard to intelligence. This study was, however, based on a college student sample and it may be that women who are able to attend college are more resilient to the abuse than others with CSA histories. Similarly, Porter, Lawson and Bigler (2005) compared children who had been sexually abused and in therapy with matched controls and found that there were not significant differences between abused and non abused children with regard to memor y and intellectual abilities when controlling for socioeconomic status and overall intelligence. Severity Many researchers have examined the impact of abuse characteristics on t he subsequent functioning of sexual abuse survivors. One characteristic that has often be en studied is the severity of the abuse. Feinauer and Stuart (1996) examined outcomes in women who had been sexually abused as children. They defined severity in terms of age at onset, identificati on of perpetrator, frequency, duration and type of abuse. They found that severity of abuse w as significantly related to current level of trauma symptoms.
Risk and Protective Factors in Child Sexual Abuse 12 Morrow and Sorell (1989) similarly found that greater severity of abuse was associated with lower self esteem, greater depression and higher frequenc y of negative behaviors. They also investigated the effects of other abuse characteristi cs such as duration, frequency and self blame. They hypothesized that severity was the most important predictor in predicting mental health outcomes in their study because vic tims of abuse viewed activities not involving intercourse as less damaging and less taboo t han activities involving intercourse. In their review of the existing literature, Kendall-Tackett, Williams and Finkelhor (1993) examined six studies about the impact of severity on survivors of CSA. Five of these studies indicated that severity of sexual abuse did significantly impac t survivors of CSA. The majority of these studies appeared to involve penetration in their definition of CSA. Kendall-Tackett et al reported, however, that one of the problems of these studie s was that the actual definition of severity varied among the studies. Merrill, Guimond, Thomsen and Milner (2003) examined different pathways by which CSA may result in both greater and fewer sexual partners using a path anal ytic model. They found that women with a history of CSA reported greater use of both avoidant and self-destructive coping strategies than women without a history of CS A. They also found that women who used self destructive coping in response to CSA were more likely to engage in dysfunctional sexual behavior and this in turn was associa ted with a greater number of sex partners. Dysfunctional sexual behavior was me asured by the Dysfunctional Sexual Behavior (DSB) Scale of the Trauma Symptom Inventor y (Briere, 1995). The DSB scale assesses sexual behaviors that are self defe ating or
Risk and Protective Factors in Child Sexual Abuse 13 maladaptive because of an indiscriminant quality, potential for self-harm, or use for nonsexual purposes (p.990). They also found that women who engaged in avoidant coping to deal with CSA were more likely to have fewer sexual partners than women who did not use avoidant coping. Since Kendall-Tackett et als (1993) review, there have been some studies which call into question the impact of abuse severity on sexual abuse survivors. Paradise, Rose Sleeper and Nathanson (1994) found that there was no relationship between abuse characteristics such as severity, frequency, duration or relationship to the pe rpetrator and problems in sexual abuse survivors. The participants in this study may have experie nced particularly severe abuse as all the participants had experienced abuse i nvolving contact, that is, children were only included in the study if the child, their parents or their c linician reported that they were physically touched in a way considered sexual by som eone 5 or more years older than them, had been touched sexually by a peer and objected to the contact, or had a newly diagnosed sexually transmitted disease. Similarly, Feiring, Taska, and Lewis (2002) examined the effect of abuse s everity in a sample of 147 children and adolescents. They found that abuse severity did not account for additional variation in adjustment in the sexual abuse survivors one year a fter abuse discovery when controlling for adjustment at the time of discovery. Instead t hey found that shame and attributional style accounted for adjustment following abuse discovery even after previous adjustment at the time of discovery is controlled. U nlike several previous studies, Feiring, et al (2002) used a complex definition of sexual a buse
Risk and Protective Factors in Child Sexual Abuse 14 severity involving level of sexual contact, relationship to perpetrator, frequenc y, duration, and use of force. Foster Care placements as a risk factor Existing research indicates that children in foster care are at a higher risk for a variety of psychological problems than children that have not been placed in foster car e. Since many children who have been sexually abused end up in foster care for varying periods of time, studying the effects of foster care placements may have im portant implications for individuals with CSA histories. Stanley, Riordan and Alaszews ki (2005) investigated the mental health of children in foster care and residential plac ements. They divided their sample into children who had a high, medium and low level of need for mental health services. They found that 47.5% of their sample had been in more than three placements during their time in the foster care system. In addition, they found that a large percentage (47%) of the children who had experienced more than three placeme nts had a high need for mental health services. Thirty seven percent of children who had experienced three or more placements had a medium level of need for mental health services, while only 16% had a low level of need for mental health services. The author s concluded that based on their sample, and the supplemental information gathered on their sample, it was likely that the high number of placements both resulted from and exacerbated mental health problems on the children in their sample. Pecora, Kessler, Williams, OBrien, Downs, English, Hiripi, White, Wiggins and Holmes (2005) found that 54.4% of foster care alumni in their study had clinical levels of at least one mental health disorder, while 19.9% had three or more mental health
Risk and Protective Factors in Child Sexual Abuse 15 problems. They found that PTSD (25.2%) and major depression (20.1%) were the most common mental health concerns of foster care alumni over the past twelve months be fore the study, and that rates of PTSD were twice as high as for U.S. war veterans Rutter (2000) suggested that this may be due to a number of factors acting together such as genetic factors, exposure to physical trauma, psychosocial experiences b efore entering the foster care system, their experiences while in the foster care syst em as well as their experiences after leaving the foster care system. Flynn, Ghazal, Legault, Vandermeulen and Petrick (2004) examined mental health in children ages 5-15. They separated the group into 5-9 year olds and 10-15 year olds. They found that children in foster care did not differ from a normative sample i n the areas of health outcome and self esteem. However, only 20% of the 10-15 year olds and 22% of the 5-9 year olds were found to be resilient in comparison to a normative sample. They also found that both groups of children in foster care had more negative outcomes in the areas of anxiety and emotional distress. Ackerman and Lindhiem (2005) examined inhibitory control and oppositional behavior in five and six year old children who were in foster care. They examined inhibitory control using the Day/Night Stroop task. They found that children who had a history of greater placement instability performed significantly w orse on the Stroop task than those with a history of greater placement stability, indicating that chil dren with greater instability had inhibitory control difficulties. They also found that ther e was a significant relationship between placement instability and oppositional behavior consistent with inhibitory control difficulties. These differences were found eve n after
Risk and Protective Factors in Child Sexual Abuse 16 controlling for IQ, and performance on a working memory task. The researchers a lso found that number of past placements was more predictive of inhibitory control and behavioral self regulation than prenatal substance exposure, prematurity, or doc umented maltreatment history. This study did not utilize a control group and did not involve sexually abused children, but rather a sample of physically abused and neglecte d children. Pecora et al (2005) also found that fewer placement changes while in fost er care was predictive of less mental health concerns in alumni in their study. Resilience While the negative effects of CSA are well documented, a number of researcher s have found that many individuals who have experienced CSA have been fairly resilient to the abuse (Bartoi & Kinder 1998; Himelein & McElrath, 1996; Liem, James, OToole, & Boudewyn, 1997; Masten & Wright, 1998; Monaghan-Blout, 1996; Spaccarelli & Kim, 1995). Resilience has been defined as a process by which individuals demonstrate positive adaptation in the face of adversity or trauma. Adversity refers to n egative life situations that are known to be associated with difficulties in adjustment. Exampl es of these include abuse or neglect, and low socio-economic status. Positive adaptation can be seen in terms of high social competence or the absence of psychological distress Three main factors have been associated with resilience: (1) personal character istics of the individual such as intelligence; (2) aspects of the individuals families suc h as cohesion or discord; (3) characteristics of the individuals environment such as their soci al support systems. Due to the dynamic nature of resilience, however, even when personali ty
Risk and Protective Factors in Child Sexual Abuse 17 characteristics of the individual are serving as protective factors, these characteristics are always being shaped by interactions between the individual and their environment. Individuals who are able to successfully overcome adversities under certain c onditions may not be able to do so under different conditions. Research seems to suggest, however, that while individuals may show changes over time, overall, individuals who do well i n certain areas continue to show positive adaptation over time. Protective factors m ay act in two ways: (1) by changing the meaning of the risk factors for the individual a nd (2) changing the individuals exposure to the risk factor (Luthar, Cicchetti & Bec ker, 2000; Luthar, Zigler & Goldstein, 1992; Rutter, 1987; Luthar & Cicchetti, 2000). Intelligence as a protective factor In studying intelligence as a protective factor, it is important to note that not all experts agree with the use of conventional assessment scales to measure inte lligence. In addition, while intelligence is often viewed as a trait, it may be influenced b y a number of environmental factors such as the context of testing, social class, parental e ducation, prejudice, and English as a second language (Vaillant & Davis, 2000). Luthar, Zigler, and Goldstein (1992) found that high achieving, gifted adolescents showed more positive psychological adjustment than their peers who were not identifie d as gifted. They concluded that this may be due to the gifted adolescents being more cognitively mature, as well as from experiential factors like those associ ated with frequent past successes. Luthar, Woolston, Sparrow, Zimmerman, and Riddle (1995) also found that achievement was strongly associated with social competence, and appeared to mediate
Risk and Protective Factors in Child Sexual Abuse 18 associations between intelligence and aspects of competence. Academic achi evement was also associated with adaptive behaviors in the contexts of personal care, domesti c skills, and skills used in the community. They concluded that success in one domain of competence is often linked with striving for success in other aspects as well. The authors suggested that these findings are useful for intervention and prognosis, as relati vely high achieving children seem to be those most likely to engage in adaptive behavior acros s different domains. Cederblad, Dahlin, Hagnell, and Hansson (1995) found that intelligence and other beneficial temperamental traits such as high activity and energy leve l, high sociability and good impulse control and persistence were associated with lower frequencies of some psychiatric diagnoses. They also found that different traits seemed to be rela ted to different diagnoses. For example, high intelligence was associated with a low er risk of depression, psychopathy, neurosis and alcoholism. Werner (1994) reported on a longitudinal study of high risk children on the Hawaiian island of Kauai and suggested that the individual dispositions of the resilie nt individuals in the study led to them seeking out environments that rewarded their competencies. While parental competence and social support were important for adult competence, this impact was less direct than the individuals disposition. Masten and Coatsworth (1998) suggested that there are three main predictors of competence in favorable and unfavorable environments: the parent-child relationship; good cognitive development or intellectual functioning; and the childs self-reg ulation of attention, emotion, and behavior. They suggested that children with good cognitive skills
Risk and Protective Factors in Child Sexual Abuse 19 may be better able to cope with unfavorable situations, because they can manage the cognitive load inherent in adverse situations. Masten and Coatsworth (1998) also suggested that IQ may act as a moderator of risk by acting as a protecti ve or risk factor in the processes linking adversity to social conduct. The authors suggested that doi ng well on IQ tests requires a variety of information-processing skills that may al so help the child to cope with adversity. For example, children with higher IQs may be able to solve problems or protect themselves better and/or have better self-regulation skil ls. On the other hand, children with below average IQs may be less able to cope with adverse situations or learn from their experiences to the same degree as children with hi gher IQs. While intelligence has been shown to be correlated with competence among high risk children, at high levels of stress children with high intelligence seem to l ose their advantage and demonstrate school based competence levels more similar to their l ess intelligent peers. There are a variety of explanations offered for these inte ractions between intelligence and stressors as predictors of competence. Children wit h a high IQ may be better at problem solving and coping, be better able to evaluate the conseque nces of their behaviors, to delay gratification, and to contain impulses. Intelligence ma y, however, act as a vulnerability factor because children with higher IQ may be more sensitive to their environments, which makes them more susceptible to life stres sors than individuals with lower IQs. Intelligent inner-city youth were found to show conside rably more variation in school based performance depending on levels of ego development than their less intelligent peers. Ego development was measured by an abbreviate d version of the Loevingers (1985) Sentence Completion Test, Form 81. Increasing levels of eg o
Risk and Protective Factors in Child Sexual Abuse 20 development have been associated with increasingly mature functioning across t he domains of impulse control, cognitive style, moral development, and interpersonal relations. Intelligent inner-city youth were also found to show more variation in sc hool based performance depending on the degree to which they experienced an internal locus of control than their less intelligent peers. However, their levels of compete nce never went below those of their less intelligent peers (Luthar & Zigler, 1992). Tiet, Bird, Davies, Hoven, Cohen, Jensen, and Goodman (1998) found that while IQ had no impact in children at low risk for psychopathology, children at high risk for psychopathology and with higher IQs may have coped better and therefore avoided the harmful effects of adverse life events. In their study, the children who showed pos itive adjustment also tended to live in higher functioning families, and receive more guida nce and supervision from their parents and other adults in the family. These authors hypothesized that higher educational aspirations may also provide high-risk yo uth with a sense of direction and hope. There have been a number of reasons suggested for the superior functioning of intellectually gifted children. They may have greater cognitive mat urity, which leads to improvements in their ability to actively structure their experiences and therefore be better able to control them. Also, because their intellectual skills are devel opmentally advanced, they may have a relatively wide variety of modes for the adaptive handli ng of their experiences. Therefore, children who are intellectually gifted ma y show better psychological adjustment than their non-gifted peers because of the greater fl exibility of their coping strategies. The psychological adjustment of gifted children may also be due
Risk and Protective Factors in Child Sexual Abuse 21 to experiential variables. For example, intellectual achievement often le ads to experiences of high prestige and success in the peer group, school, and family. This history of frequent successes could, therefore, in conjunction with these superior coping strat egies contribute to the better adjustment levels shown by academically and intellec tually gifted children (Luthar, Zigler & Goldstein, 1992).
Risk and Protective Factors in Child Sexual Abuse 22 The present study Previous research has indicated that a history of CSA can have a number of negative consequences such as attentional problems, cognitive dysfunction, internali zing problems such as depression and externalizing problems such as aggression. Gender differences have been observed in the nature of the psychological consequences observed. Research has also indicated that severity of abuse may mediate the re lationship between CSA and psychological sequelae. These studies, have however, not been conclusive as a number of studies have also found different results. There have been a number of risk and protective factors that have been studied in relation to at risk chil dren. One protective factor that has been identified in at risk children and which was exa mined in this study is intellectual functioning. As previous researchers have suggested, i t may be that children with higher IQs are better able to cope effectively with be ing abused, have more positive academic and interpersonal experiences from which to draw during t imes of stress, or may in fact be less prone to abuse. Furthermore, a history of CSA and the sequelae that may follow from being abused during a childs formative developm ental years may prove to be an obstacle to achieving ones cognitive potential, possibl y by changing pathways in the brain, and thus result in lower IQ than otherwise may ha ve been achieved. This lower IQ may then make future coping with stressful events m ore difficult and the likelihood of emotional and/or behavioral problems higher as well as having great difficulty with achievement when compared to children with lower IQ.
Risk and Protective Factors in Child Sexual Abuse 23 There may also be an association between IQ and severity as children with low er IQs may be less prone to reporting abuse and thus may suffer from prolonged abuse than children with lower IQs (Mansell, Moskal, & Sobsey, 1998). The number of foster care placements was also examined as a possible risk factor. Since many childre n who have been sexually abused are removed from the home, it may be that some of the negative consequences observed in sexually abused children may be moderated by the number of foster care placements that the children have experienced. As previously sta ted, many studies have identified a link between number of home placements and negative outcomes. While these factors have been previously studied extensively, little r esearch has examined the relationship among these particular variables as the curre nt study has attempted to do. While there were efforts made in the current study to derive a comparison group of non-sexually abused children, it is recognized that while the children in the non-abused sample do not have any reported sexual abuse history, they may have been sexually abused in the past without the abuse being reported to authorities.
Risk and Protective Factors in Child Sexual Abuse 24 Hypotheses Based on the previous research, the following hypotheses are made to further existing research and address the limitations of some of the previous studies: ( 1) Children in the sexually abused sample will have higher scores on the clinical scales of the Achenbach Child Behavior Checklist (CBCL; Achenbach, 1991, 2001), specifically with regard to attention problems, aggression, anxiety, depression and somatic complai nts; (2) Boys with CSA histories will demonstrate higher clinical elevations than girls on the aggression and rule breaking scales of the CBCL, while girls will demonstrate hi gher clinical elevations on the anxious and depressed scales of the CBCL (3) Children wi th histories of CSA will also have higher scores on the Childrens Depression Inve ntory (CDI; Kovaks, 1992) than children without histories of CSA; (4) Children with histories of CSA will have significantly lower Full Scale IQ scores; (5) Children w ith histories of CSA will have lower total achievement scores; (6) Children with higher score s on intelligence tests will have lower Achenbach and CDI scores and higher achie vement scores than children with lower intelligence scores; (7) Children who have been i n fewer foster homes will have lower Achenbach and CDI scores and higher achievement sc ores than children who have been in more foster homes; (8) Children who have undergone more severe sexual abuse will have higher Achenbach and CDI scores and lower achievement scores than those with less sexual severe abuse; (9) History of pr evious psychological treatment will be associated with lower Achenbach and CDI s cores and
Risk and Protective Factors in Child Sexual Abuse 25 higher achievement scores; (10) Intelligence will moderate the relat ionship between sexual abuse, CBCL, CDI and Achievement scores; (11) Number of foster homes wi ll moderate the relationship between abuse, CBCL, CDI and Achievement scores; (12) age of child at testing will also moderate the relationship between abuse, CBCL, CD I and Achievement scores; (13) Severity of abuse will also moderate the relations hip between abuse, CBCL, CDI and Achievement scores (please see below for definition of s everity to be used).
Risk and Protective Factors in Child Sexual Abuse 26 Method Measures The Wechsler Intelligence Scales for Children (WISC; Wechsler, 1991, 2003) a re instruments used to assess intellectual functioning for children aged 6-16. The two versions of the test that were used in this study were the WISC-III (1991) and the WISCIV (2003). The WISC is an individually administered test. A full scale IQ is deri ved from both scales and this full scale IQ was used in the analysis. Both versions of thi s test have good validity and reliability and are widely used in research and practice (Kaufm an, Flanagan, Alfonso, & Mascolo, 2006; Needelman, Schnoes, & Ellis, 2006; Sattler, 2001). The CBCL (Achenbach, 1991, 2001) is a measure which is used to assess the competencies and problems of children and adolescents. The revised scale differs f rom the original scale with regard to some items and norming sample. Additionally, w hile the names of the scales differ between the versions, they are roughly equivalent. T he CBCL/6-18 (Achenbach, 2001) uses parents to rate their childrens problems and competencies. It is a revision of the CBCL/4-18 and consists of 113 items. It is desig ned for children aged 6-18 years and was normed on a national sample which was representative of the population. The CBCL scales have good reliabilities wit h alphas for the competence scales (Activities, Social, School and Total Competenc e) ranging from .63 to .69. For the empirically based scales (Anxious/Depressed Withdrawn/Depressed, Somatic Complaints, Social Problems, Thought
Risk and Protective Factors in Child Sexual Abuse 27 Problems, Attention Problems, Rule-Breaking Behavior, Aggressive Beha vior, Internalizing, Externalizing and Total Problems scales) alphas ranged from .78 to .97. Alphas for the DSM-Oriented scales ranged from .72 to .91. For the Competence scales, test-retest reliability ranged from .82 to .93 over a one week period. For the empirically based scales, test-retest reliabilities ranged f rom .82 to .94. Test retest reliabilities ranged from .80 to .93 for the DSM-oriented scales (Achenbach & Re scorla, 2001). The authors of the manual state that the content validity of these scales is well supported by years of research and consultation. Additionally, multiple regressions odds rations and discriminant analyses all showed significant ( p < .01) discrimination between referred and non-referred children indicating good criterion-related vali dity. Good construct validity was also reported with the CBCL being significantly ass ociated with other scales with similar dimensions and long term predictions of outcomes (Ache nbach & Rescorla, 2001). The CDI is a 27 item self rating scale used to assess depressive symptomat ology in children and adolescents aged 7-17. Items on the scale range from 0-2 with hig her scores indicating higher levels of the particular symptom endorsed. A total score is derived by adding all the individuals responses and this total score will be used in t he analysis. The CDI also produces five subscale scores Negative Mood, Interpers onal Problems, Ineffectiveness, Anhedonia and Negative Self Esteem (Kovacs, 1992). T he CDI is a widely used instrument with acceptable validity and reliability ( Saylor, Finch, Spirito & Bennett, 1984). Saylor et al (1984) examined the psychometric properties of t he
Risk and Protective Factors in Child Sexual Abuse 28 CDI in a group of psychiatric inpatients and normal controls. The CDI demonstrated acceptable inter-rater ( r = .38 to .87) and split-half ( r = .57 to .75) reliabilities and good internal reliability ( r = .80 to .94). The CDI also demonstrated acceptable criterion validity with significant differences being found between psychiatric inpa tients and controls (t (46) = 2.48, p <.05) and good construct validity, with significant correlations found between the CDI and the Piers-Harris Self Concept Scale ( r (26) = .46, p <.05). There were some concerns about the specificity of the CDI with regard to depress ive symptomatology, as it appeared to be better suited to identifying general psychopathology. The Woodcock-Johnson Tests of Achievement-Third Edition (2001), its precursor, the Woodcock-Johnson Psycho-Educational Battery-Revised (WJ-R; 1989) and the Wechsler Individual Achievement Tests (WIAT, 1992; WIAT II, 2002) are widely used tests designed to assess academic functioning. They each provi de Reading, Math and Writing Composites and have good reliability and validity for all scales (Sattler, 2001). Martell and Smith (1994) examined the relationship between the WJ-R and the WIAT in a sample of children. They found that the Reading Composites for the two measures were highly and significantly correlated ( r = .70, p < .001). Similarly, the Math Composites were significantly correlated ( r = .54, p < .001). While the Writing Composites were also moderately correlated ( r = .59), the relationship was not significant because of the small numbers of children that completed the writing composites f or each test (n = 11).
Risk and Protective Factors in Child Sexual Abuse 29 Severity was assessed in a similar manner to Merrill et al (2003). They used a global index of CSA severity by assigning one point for (1) penetration (2) force or threats (3) father or stepfather as perpetrator (4) more than one perpetrator (5) more than five incidents. In this study, age at onset will also be included as a severity var iable with abuse occurring before the age of 13 being assigned one point. Higher scores indicat e more severe CSA. A global index of CSA severity was then determined based on the number derived from the above procedure after review of case files of the partici pants which were conducted by the examiner and a research assistant. A yes, no measure of previous psychological treatment was also compil ed. In addition, information about the number of home placements each child had was gathered. Procedure Available information on the children with and without histories of sexual abuse at the collection sites was gathered by the researcher and/or the rese arch assistant. This included scores on the WISCs, CBCLs, CDIs and WJs and WIATs. For the analyses conducted, Cohen (1992) suggested that to detect a medium effect size, approximatel y 84 participants would be required. A sample of 20 files was examined by both raters to assess inter-rater reliability ( r = .90, p < .01) of the severity measure. Each record was identified by a number and there were no names associated with the records in the database to ensure confidentiality. The records of each child were scre ened for sexual abuse history. Records of children with CSA histories were examined for abuse characteristics and effects of CSA. The records were then compared with t hose of children who did not have a documented history of CSA in their files to test the
Risk and Protective Factors in Child Sexual Abuse 30 hypotheses. Attempts were made to match participants in the sexually abuse d and nonsexually abused group by age. Participants were stratified by age for anal ysis with children 7-12 being in one group and children 13-16 in the other group. Case files were examined in order from most recent to least recent. Once the data on sexually abus ed children was obtained, an equal number of children, matched by age and without documented histories of sexual abuse in order of most to least recent was gathe red as a control group. Descriptive statistics for the sample demographics were computed. A ser ies of Ttests were carried out to test for sample differences, and to test hypotheses 1-5. Correlations were then carried out to test hypotheses 6-9. Hypotheses 10-13 were tes ted using the guidelines provided by Baron and Kenny (1986), Aiken and West (1991) and Holmbeck (2002) for testing moderator effects. The main analyses were conducte d on the entire sample using the Total score on the CDI, the Total, Internalizing and E xternalizing Scores of the CBCL, and the Total Achievement score of the WJ or WIAT. Within gr oup differences for the sexually abused group were examined for hypothesis 6 and 7. G ender was statistically controlled for within SPSS, where relevant. Participants Data comprised existing records from The Childrens Home, a group home for children who have been removed from the home due to abuse or neglect, The University of South Florida Psychological Services Center, and the University of Florida Divi sion of Child and Adolescent Psychiatry both of which offer outpatient assessment and therapeutic services to the community, and the University of South Florida, Divi sion of
Risk and Protective Factors in Child Sexual Abuse 31 Child Development, an outpatient clinic providing multidisciplinary evaluations to children (see Table 1). Participants were between the ages of 7 and 16. Informa tion regarding ethnicity and gender was collected from each childs file, wher e available. Table 1 Locations from which participants were drawn Location Group N Childrens Home Sexually Abused Non-Abused 95 0 University of South Florida Division of Child Devel opment Sexually Abused Non-Abused 1 2 University of South Florida Psychology Clinic Sexua lly Abused Non-Abused 3 79 University of Florida Division of Child and Adolesc ent Psychiatry Sexually Abused Non-Abused 18 0 There were a total of 198 participants and of these 59.1% (n = 117) had a history of sexual abuse, according to the above criteria, while 40.4% (n = 80) did not have a history of reported sexual abuse. Participants ranged in age from 7-16 (M = 10.89, SD = 2.65). When the sample was stratified by age, there were a total of 139 (70.2%) under the age of 13 at the time of testing and 59 (29.8%) who were 13 or older. Seventy percent of the sample was identified as Caucasian, 11.1% as Hispanic, 9.6% as African Amer ican, 0.5% as Asian, and 8.1% were identified as from another group, such as being of mixed
Risk and Protective Factors in Child Sexual Abuse 32 descent, and 0.5% of the participants did not have information available about their ethnicity. Ethnicity was not significantly associated with abuse history or the moderators of interest, that is, Full Scale IQ, age at time of testing, severity of abuse or history of previous treatment. The sample was made up of both male and female participants, w ith 53.5% being male and 46.5% being female. Fifty two percent of the sample had a received psychological treatment in the past while 16.2% had not. Psychological treatment history was not available for 31.3% of the sample.
Risk and Protective Factors in Child Sexual Abuse 33 Results Given that multiple outcome variables were tested instead of a single outcome variable for each set of analyses, a Bonferroni correction is indicated to co ntrol for experiment-wise error. However, the relatively small sample size invol ved in these analyses, along with a Bonferroni correction, would reduce the likelihood of finding significant results and in turn inflate the possibility of committing a Type II error. Therefore, after examining the factors for and against using a Bonferr oni correction, the present study analyzed the data according to plan by testing multiple crit erion variables separately in each set of analyses. Given the aforementioned problems with usi ng a strict, Bonferroni-adjusted alpha level, the results are discussed using the original, conventional alpha level of .05. T-tests indicated that there were no significant differences between the se xually abused and non-sexually abused groups in terms of age (t (151) = -.42, p >.05) and ethnicity (t (195) = -.47, p >.05). However, the groups did differ significantly in terms of gender (t (179) = -4.29, p <.05), number of placements (t (193) = -10.47, p <.05) and history of previous psychological treatment (t (134) = 12.15, p <.05), with the sexually abused group containing more girls and children with larger number of home placement s and children in the non-sexually abused group having a higher likelihood of having sought previous psychological treatment.
Risk and Protective Factors in Child Sexual Abuse 34 The results of the t-tests conducted to test hypothesis 1 indicated that having a history of sexual abuse was associated with higher levels of Total Behavior P roblems (t (145) = 6.18, p <.05), Internalizing problems (t (145) = 4.63, p <.05), and Externalizing Behavior Problems (t (145) = 6.70, p <.05) on the CBCL (see Table 2). When the individual subscales of the CBCL were examined, having a history of sexual abuse w as associated with greater problems with Anxious/Depressed Behavior (t (14 1) = 6.12, p <.05), Withdrawn/Depressed Behavior (t (133.56) = 4.15, p <.05), Social Problems (t (141) = 6.09, p <.05), Thought Problems (t (141) = 3.75, p <.05), Rule Breaking Behavior (t (141) = 7.11, p <.05), and Aggressive Behavior (t (141) = 5.66, p <.05) as measured by the CBCL (see Table 3). Table 2 Independent Samples T-Test between Sexually Abused and Non-Sexually Abused Children on Demographics, IQ, CDI and Achievement Variable Group N Mean SD df T Age Abused Non-Abused 117 80 10.97 10.78 2.48 2.90 151 -.42 Ethnicity Abused Non-Abused 117 80 1.72 1.63 1.33 1.11 195 -.47 Gender Abused Non-Abused 117 80 1.58 1.29 .50 .46 179 -4.29* Number of placements Abused Non-Abused 115 80 7.5 1.1 5.5 .34 193 -10.47*
Risk and Protective Factors in Child Sexual Abuse 35 Variable Group N Mean SD df T History of Treatment Abused Non-Abused 117 80 1.29 2.53 .56 .78 134 12.15* Full Scale IQ Abused Non-Abused 117 80 89.44 96.64 14.40 13.35 195 3.55* Total CDI Abused Non-Abused 81 74 50.89 48.9 11.05 10.05 153 -1.17 Total Achievement Abused Non-Abused 106 75 87.40 95.84 14.43 10.70 179 -4.30* p < .05, two tailed Table 3 Independent Samples T-Test between Sexually Abused and Non-Sexually Abused Children on CBCL Variables Variable Group N Mean SD df T Total Behavior Problems Abused Non-Abused 67 80 69.30 58.89 9.53 10.68 145 6.18* Internalizing Problems Abused Non-Abused 67 80 63.60 55.71 9.46 10.93 145 4.63* Externalizing Problems Abused Non-Abused 67 80 68.42 56.66 9.80 11.21 145 6.70* Anxious/Depressed Behavior Abused Non-Abused 63 80 65.29 56.24 10.02 7.66 141 6.12*
Risk and Protective Factors in Child Sexual Abuse 36 Variable Group N Mean SD df T Withdrawn/Depressed Behavior Abused Non-Abused 63 80 64.32 57.76 9.37 9.41 133.56 4.15* Somatic Complaints Abused Non-Abused 63 80 57.71 57.21 8.40 8.38 141 .36 Social Problems Abused Non-Abused 63 80 67.79 58.15 10.29 8.64 141 6.09* Thought Problems Abused Non-Abused 63 80 64.10 58.00 11.57 7.83 141 3.75* Attention Problems Abused Non-Abused 63 80 66.98 63.71 12.48 9.00 141 1.82 Rule Breaking Behavior Abused Non-Abused 63 80 68.08 57.53 9.39 8.34 141 7.11* Aggressive Behavior Abused Non-Abused 63 80 68.63 58.75 10.93 9.92 141 5.66* p < .05, two tailed The results of the t-tests for the second hypothesis suggested that while, as previously discussed, there was a significant relationship between gender and a buse history, there was not a significant relationship between gender and the Total (t (65) = .63, p >.05), Internalizing (t (65) = -.80, p >.05) and Externalizing (t (65) = -.60, p >.05) scales of the CBCL (see Table 4).
Risk and Protective Factors in Child Sexual Abuse 37 Table 4 Independent Samples T-Test between Sexually Abused Boys and Girls on CBCL Variables Variable Group N Mean SD df T Total Behavior Problems Boys Girls 28 39 68.43 69.92 9.75 9.45 65 -.63 Internalizing Problems Boys Girls 28 39 62.50 64.38 9.66 9.37 65 -.80 Externalizing Problems Boys Girls 28 39 67.57 69.03 11.32 8065 65 -.60 Anxious/Depressed Behavior Boys Girls 25 38 63.48 66.47 9.79 10.11 61 -1.16 Withdrawn/Depressed Behavior Boys Girls 25 38 63.72 64.71 9.48 9.40 61 -.41 Somatic Complaints Boys Girls 25 38 56.92 58.24 7.85 8.81 61 -.61 Social Problems Boys Girls 25 38 64.88 69.71 9.63 10.37 61 -1.86 Thought Problems Boys Girls 25 38 63.56 64.45 9.40 12.90 61 -.30 Attention Problems Boys Girls 25 38 65.12 68.21 12.12 12.72 61 -.96
Risk and Protective Factors in Child Sexual Abuse 38 Variable Group N Mean SD df T Rule Breaking Behavior Boys Girls 25 38 67.84 68.24 9.71 9.31 61 -.16 Aggressive Behavior Boys Girls 25 38 67.64 69.29 12.26 10.09 61 -.58 p < .05, two tailed Results of the analysis for hypothesis 3 indicated that sexual abuse history wa s not significantly associated with the total depression score (t (153) = -1.17, p >.05) as measured by the CDI (see Table 2). Analyses conducted for hypothesis 4 that as predicted, children with histories of sexual abuse had significantly lower Full Scale IQ scores than children without sexual abuse histories (t (195) = 3.55 p <.05) (see Table 2). Because the WJ and WIAT are highly correlated, analyses for hypothesis 5 we re based on whichever test was administered to each participant. A Total Achieve ment Score for each child was computed by averaging their Broad Reading, Bro ad Math and Broad Written Language Composite Scores from either the WJ or the WIAT. Us ing this Total Achievement Score, children with sexual abuse histories scored signific antly lower than children without sexual abuse histories (t (179) = -4.30, p <.05) (see Table 2). Hypotheses 6-9 were tested using zero-order correlation analyses. Analyses to t est hypothesis 6 suggested that for the overall sample, IQ was significantly ass ociated with the Total CDI score with higher IQ scores being associated with signif icantly lower Total
Risk and Protective Factors in Child Sexual Abuse 39 scores (see Table 5). Additionally, higher IQ was significantly assoc iated with lower Total Behavior Problems and Externalizing Problems of the CBCL (see Table 6 ). Higher IQ was also significantly associated with higher Total Achievement scor es (see Table 7). Table 5 Correlations between Full Scale IQ, Number of homes, Severity of Abus e, History of Psychological Treatment and Total CDI Score Variable Abuse History FSIQ # of homes Severity Hx of tx Total CDI Score Abuse History 1 -.25** .60** -.68** .09 FSIQ 1 -.07 -.11 .07 -.21** # of homes 1 .19* -.24** .00 Severity 1 -.10 .05 Hx of tx 1 -.04 Total CDI 1 Note. Dashes mean either already reported or not c omputed. FSIQ = Full Scale IQ; # of homes = Number of home placements; Severity = severity of abuse; H x of tx = History of previous treatment p < .05, **p < .01
Risk and Protective Factors in Child Sexual Abuse 40 Table 6 Correlations between Full Scale IQ, Number of homes, Severity of Abuse, Hist ory of Psychological Treatment and CBCL Scores Variable FSIQ # of homes Severity Hx of tx Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thgh t. Attn. R. Break. Agg. FSIQ 1 -.07 -.11 .07 -25** -.14 -.25** -.12 -.09 -.10 -.29** -.14 -.24** -.16 -.21* # of homes 1 .19* -.24** .27** .22** .34** .32** .19* -.12 .30** .14 -.02 .34** .30** Severity 1 -.10 .21 .14 .23 .11 .15 .13 .12 .21 .01 .06 .20 Hx of tx 1 -.07 -.07 -.10 -.04 -.10 -.04 -.12 .00 -.01 -.07 -.12 Total 1 .77** .88** .73** .61** .45** .73** .62** .69** .75** .82**
Risk and Protective Factors in Child Sexual Abuse 41 Variable FSIQ # of homes Severity Hx of tx Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thgh t. Attn. R. Break. Agg. Int. 1 .55** .84** .73** .61** .64** .46** .40** .48** .51** Ext. 1 .59** .41** .27** .55** .49** .52** .85** .93** Anx./ Dep. 1 .62** .42** .62** .46** .38** .54** .60** With./ Dep 1 .31** .45** .43** .38** .46** .39** Som. Comp. 1 .34** .25** .31** .18** .28** Soc. 1 .42** .55** .45** .55**
Risk and Protective Factors in Child Sexual Abuse 42 Variable FSIQ # of homes Severity Hx of tx Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thgh t. Attn. R. Break. Agg. Thght. 1 .55** .52** .49** Attn. 1 .44** .51** R. Break. 1 .78** Agg. 1 Note. Dashes mean either already reported. FSIQ = F ull Scale IQ; # of homes = Number of home placement s; Severity = severity of abuse; Hx of tx = History of previous treatment; Total = Total Behavi or Problems; Int. = Internalizing Behavior Problems Composite; Ext. = Externalizing Behavior Problems Composite; Anx./Dep. = Anxious/Depressed s ubscale; With./Dep. Withdrawn/Depressed subscale; Som. Comp. = Somatic Complaints subscale; Soc. = Social Problems subscale; Thght = Thought Problems subscale; R. Break. = Rule Breakin g subscale; Agg. = Aggressive Behavior subscale p < .05, **p < .01
Risk and Protective Factors in Child Sexual Abuse 43 Table 7 Correlations between Full Scale IQ, Number of homes, Severity of Abuse, Hist ory of Psychological Treatment and Achievement Scores Variable FSIQ # of homes Severity Hx of tx Total Achievement FSIQ 1 -.07 -.11 .07 .58** # of homes 1 .19* -.24** -.13 Severity 1 -.10 -.04 Hx of tx 1 .06 Total Achievement 1 Note. Dashes mean already reported. FSIQ = Full Sca le IQ; # of homes = Number of home placements; Severity = severity of abuse; Hx of tx = History of previous treatment. **p < .01 For the sexually abused group only, IQ was also significantly associated wi th total CDI scores with higher IQ being associated with significantly lower Tot al scores. IQ was not significantly associated with the Total, Internalizing, or Externaliz ing scales of the CBCL. However, as with the overall sample, IQ was significantly associ ated with higher Total Achievement scores (see Table 8).
Risk and Protective Factors in Child Sexual Abuse 44 Table 8 Correlations between Full Scale IQ, Number of Homes, CDI, Achievement and CBC L Scores for the Sexually Abused Group Variable FSIQ # of homes CDI Achievement Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thght. Attn. R. Break. Agg. FSIQ 1 .13 -.32** .62** -.01 .12 -.12 .12 .16 -.12 -.21 -.08 -.23 -.05 -.12 # of homes 1 -.09 .09 -.07 -.03 .03 .03 -.08 -.28* -.01 -.11 -.23 -.02 .01 CDI 1 -.28** .19 .06 .18 .08 .17 .19 .16 .26 .19 .25 .23 Achievement 1 -.07 .12 -.07 .19 .12 -.10 .12 -.13 -.28* -.04 .01 Total 1 .70** .84** .71** .61** .46** .60** .64** .69** .70** .80**
Risk and Protective Factors in Child Sexual Abuse 45 Variable FSIQ # of homes CDI Achievement Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thght. Attn. R. Break. Agg. Int. 1 .46** .87** .65** .58** .50** .50** .33** .41** .41** Ext. 1 .53** .35** .25* .35** .49** .47** .82** .95** Anx./ Dep. 1 .57** .48** .42** .41** .35** .44** .51** With./ Dep 1 .25* .29* .44** .38** .36** .34** Som. Comp. 1 .37** .28* .33** .13 .27* Soc. 1 .30* .45** .23 .33**
Risk and Protective Factors in Child Sexual Abuse 46 Variable FSIQ # of homes CDI Achievement Total Int. Ext. Anx./ Dep. With./ Dep Som. Comp. Soc. Thght. Attn. R. Break. Agg. Thght. 1 .54** .48** .49** Attn. 1 .44** .43** R. Break. 1 .70** Agg. 1 Note. Dashes mean either already reported. FSIQ = F ull Scale IQ; # of homes = Number of home placement s; Total = Total Behavior Problems; Int. = Internalizing Behavior Problems Composite; Ext. = E xternalizing Behavior Problems Composite; Anx./Dep. = Anxious/Depressed subscale; With./Dep. Withdrawn/Depressed subscale; Som. Comp. = Somatic Complaints subscale; Soc. = Social Problems subscal e; Thght = Thought Problems subscale; R. Break. = Rule Breaking subscale; Agg. = Aggressive Behavior subscale p < .05, **p < .01
Risk and Protective Factors in Child Sexual Abuse 47 Results of the correlation analysis conducted to test hypothesis 7 revealed tha t number of home placements was not significantly associated with Total CDI scor es (see Table 5) for the overall sample. However, number of home placements was significa ntly associated with higher Total Behavior Problems, Internalizing problems, and Externalizing Problems on the CBCL (see Table 6). Higher numbers of placeme nts was not associated with Total Achievement (see Table 7). For the sexually abused group, number of home placements was not significantly associated with Total CDI scor es, CBCL scores, or Total Achievement scores (see Table 8). In order to conduct the correlation analysis to test hypothesis 8, severity wa s first assessed by two raters using Merrill et als (2003) measure while revie wing subjects charts. Inter-rater reliability for the severity measure was calc ulated for a subset of the sample and was found to be high ( r = .90). Abuse severity was not significantly associated with Total CDI scores, CBCL subscale scores, nor with achievement scores (see Tables 5, 6, and 7 respectively). Correlation analyses to test hypothesis 9 indicated that history of previous psychological treatment was not significantly associated with Total C DI, any of the scales of the CBCL, or with achievement (see Tables 5, 6, and 7 respectively). A number of individual difference variables (age, number of placements, IQ and severity of abuse) were thought to be potential moderators of the effect of abuse on the outcome variables (CBCL, CDI and Achievement scores) and were presented as hypotheses 10-13. To establish whether these variables moderated the effects of abus e on the outcome variables, a series of hierarchical multiple regressions were c onducted to test
Risk and Protective Factors in Child Sexual Abuse 48 for a significant interaction between each variable and abuse history. The mai n effects of abuse, age at time of testing, IQ and number of home placements on the dependent variables were also examined. Data were screened to ensure that all assum ptions for hierarchical regression analysis were met. As recommended by Aiken and W est (1991), variables of interest were centered prior to computing interaction terms in order to reduce the effect of multicollinearity. This was accomplished by subtracting the means for abuse, number of previous placements, age at time of testing (stratified), and Ful l Scale IQ. Because gender was associated with a number of the variables of interest, it was entered in the first step of regression analyses (after being centered) f or the purposes of statistical control. Abuse was entered in the second step, the potential moderator was entered in the third step, and the interaction of the given variable and abuse was enter ed in the fourth step. A significant interaction occurs when the product term offers addi tional prediction above and beyond that provided by the main effects. Hierarchical multiple regression analyses were carried out for all potential moderators on all outc ome variables. All significant interaction effects were further tested following the post hoc probing techniques recommended by Holmbeck (2002) using conditional moderators. When Full Scale IQ was entered to test hypothesis 10, the overall models were significant ( p < .05) for Total CBCL (F(4, 142) = 10.62) scores, Internalizing (F(4, 142) = 6.77) and Externalizing (F(4, 142) = 11.97) subscale scores of the CBCL, Total CDI scores (F(4, 150) = 2.86), and Total Achievement scores (F(4, 176) = 29.50). These variables accounted for 23%, 16%, 25%, 7%, and 40% of the variance in the models, respectively. Additionally, Full Scale IQ was shown to moderate the rel ationship between
Risk and Protective Factors in Child Sexual Abuse 49 abuse and Total Achievement (see Figure 1). There was also a trend toward sig nificance noted for the Internalizing subscale of the CBCL (see Figure 2). Moderating Effect of IQ on Abuse and Total Achievem ent0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 low med high Abuse Total Achievement IQ high med low Figure1
Risk and Protective Factors in Child Sexual Abuse 50 Moderating Effect of IQ on Abuse and CBCL Internali zing0.00 10.00 20.00 30.00 40.00 50.00 60.00 low med high Abuse CBCL Internalizing IQ high med low Figure 2 The post-hoc test of the interaction of abuse and FSIQ as a predictor of Total Achievement revealed a significant relationship between Abuse and Total Achie vement at medium (t = -2.80, p < .01) and low (t = -4.12, p < .01) levels of IQ, but not at high (t = -.82, p > .05) levels of IQ. Similar post-hoc testing of the interaction of abuse and IQ as a predictor of the Internalizing subscale of the CBCL revealed a significant relationship between abuse and the Internalizing subscale of the CBCL at high (t = 3.98, p < .01) and medium (t = 3.73, p < .01) levels of IQ, but not significant at low (t = 1.46, p > .05) levels of IQ. When number of placements was entered to test hypothesis 11, the overall models were significant ( p <.05) for the Total (F(4, 141) = 9.64), Internalizing (F(4, 141) = 6.12), and Externalizing (F(4, 141) = 11.10) scales of the CBCL as well as the Total
Risk and Protective Factors in Child Sexual Abuse 51 Achievement score (F(4, 174) = 4.68). These variables accounted for 22%, 15%, 24%, 10% of the variance in the models, respectively. The model for the Total CDI score wa s not significant (F(4, 148) = 1.09, p > .05). Similarly, number of placements did not moderate the relationship between abuse and any of the variables. When age at time of testing was entered to test hypothesis 12, the overall models were significant ( p < .05) for the Total (F(4, 142) = 9.84), Internalizing (F(4, 142) = 6.02), and Externalizing (F(4, 142) = 11.64) scales of the CBCL as well the Total Achieve ment score (F(4, 176) = 4.72). These variables accounted for 22%, 15%, 25%, and 10% of the variance in the models, respectively. The model for the Total CDI score was not significant (F(4, 150) = 1.46, p > .05). In addition, age at time of testing did not moderate the relationship between abuse and any of the variables. Hypothesis 13, which aimed to test whether abuse severity would moderate the relationship between abuse and the outcome variables, could not be tested because children who had not been abused did not have a severity score.
Risk and Protective Factors in Child Sexual Abuse 52 Discussion In summary, this study examined differences between children with and without histories of CSA with regard variables such as behavioral and emotional functioning and achievement. The role of risk and protective factors such as gender, IQ, abus e severity, number of foster home placements and history of previous psychological treatment w ere also examined as potentially affecting outcomes in children with CSA histories The results of this study revealed a number of findings which were consistent with expectations based on previous research as well as some unexpected findings. Results of the analyses of the hypotheses indicated that as predicted children w ith CSA histories were rated by their caregivers as exhibiting signifi cantly more total behavior problems, including both internalizing and externalizing behavior problems. In particular, having a history of sexual abuse was associated with greater pr oblems with anxious/depressed behavior, withdrawn/depressed behavior, social problems, thought problems, rule breaking behavior and aggressive behavior, as measured by caregiv er reports. It should again be noted, that the comparison group for this study consisted of a clinical population of children and so, even in comparison to other clinical populations, children with CSA histories were rated as demonstrating significantly m ore problems. The differences between the CSA and non-CSA groups on these dimensions were generally large and those that were statistically significant may al so be considered to be clinically significant as means for the CSA group generally fell c lose to the clinically
Risk and Protective Factors in Child Sexual Abuse 53 significant range for problems on the CBCL, while means for the non-CSA group generally fell further away from the clinically significant ran ge. This finding is consistent with the large body of research which suggests that children with his tories of sexual abuse exhibit significantly more internalizing and externalizing be havioral problems than children without sexual abuse histories (Browne & Finkelhor, 1986; Chaffin, Silovsky, & Vaughn, 2005; Inderbitzen-Pisaruk, Shawchuck & Hoier 1992; Johnson & Kenkel, 1991; Kendall-Tackett, Williams, Finkelhor, 1993; Martin, Bergen, Richardson, Roeger & Allison, 2004; Monahan & Forgash, 2000; Newman, Clayton, Zuellig, Cashman, Arnow, Dea, & Taylor, 2000; Paolucci, Genuis, & Violato, 2001; Saywitz, Mannarino, Berliner & Cohen, 2000; Spaccarelli & Fuchs, 1997). While CSA was not associated with increased attention problems on the CBCL, it should again be noted that the control sample was a clinical sample, which was heavily drawn from ADHD evaluations. It may be that instead of interpreting these results to su ggest that children with CSA do not exhibit attention problems, to infer that children with sexual abuse histories are not significantly different than children presenting with attention problems. The current study also found that there was not a significant relationship betw een gender and the CBCL scales. This is contrary to expectations based on Garnefski a nd Arends (1998) research which suggested that there were differences with rega rd to emotional problems and in which boys demonstrated more behavioral problems, particularly aggressive behavior and rule-breaking than girls. The results of this study are, however, consistent with research by Dube, Anda, Whitfield, Brown, Felitte, Dong a nd
Risk and Protective Factors in Child Sexual Abuse 54 Giles (2005) who found no gender differences in adult survivors of CSA when examining factors such as suicide attempts and interpersonal relationships. Similarl y, Banyard, Williams and Siegel (2004) in examining a number of mental health outcomes such as anxiety, depression, anger, and suicidality found that males and females differe d only with regard to sexual concerns, which was not specifically measured in the curr ent study. Contrary to expectations, CSA was not significantly associated with overall self reported depression when using the CDI, although this result is not altogether surpris ing in the context of the sample from which this data was drawn. These results sugges t that sexually abused children do not appear to be significantly different than another sa mple of clinic referred children on the dimension of self reported depression. In addition, it should again be noted that while children with CSA did report somewhat higher ratings on the CDI, the relatively small data available may not have provided sufficient powe r to detect significant effects. While the results suggest that CSA is not associated with self-reported depre ssion, as previously discussed, results from the CBCL suggested that CSA was associ ated with depression as reported by caregivers. It may be that these seemingly c onflicting results are due to the different sources of the information (i.e. self versus caregiver) and/or that there was more data available for the CBCL than for the CDI to detect signific ant results as there was more power for the analysis of the CBCL data. As expected, children with histories of sexual abuse had significantly lower IQ s than children without sexual abuse histories, an over 7 point difference in means which is both statistically and clinically meaningful. This is consistent with re search by Trickett, et
Risk and Protective Factors in Child Sexual Abuse 55 al (1994) who suggested that the abuse experience can negatively impact cognitive functioning. Another possible explanation may be that children with lower IQs may be more susceptible to being abused than children with higher IQs and may be less like ly to report abuse which may result in the abuse persisting for a longer period of time (Mansell, et al., 1998). While the nature of the difference in IQ cannot be determi ned from the current study, there are important implications of this difference fo r prevention and treatment programs which are discussed below as well as important implic ations for future research to examine possible pathways to the lower IQs of these childre n. Consistent with expectations, IQ did serve as a protective factor with highe r IQ being significantly associated with lower overall self reported depressive symptoms for both the overall and the CSA only groups. IQ was also significantly associated wi th the lower overall behavior problems and in particular lower reports of externalizing pr oblems by caregivers for the overall group, although it was not significantly associa ted with caregiver reports of behavior problems for the CSA group only. There was a trend towa rd significance noted for the caregiver reports of internalizing problems when IQ was entered as a moderator, particularly for high and medium levels of IQ. This f inding is consistent with a number of studies suggesting that IQ can serve as an overall pr otective factor (Luthar, Zigler, & Goldstein, 1992; Masten & Coatsworth, 1998) and in particul ar be associated with lower rates of specific types of psychopathology (Cede rblad, Dahlin, Hagnell, & Hansson, 1995; Tiet, Bird, Davies, Hoven, Cohen, Jensen, & Goodman 1998). The results suggesting that low levels of IQ, however, may not play a moderational role are, however, novel.
Risk and Protective Factors in Child Sexual Abuse 56 Consistent with research by Buckle, Lancaster, Powell and Higgins (2005) and Jones, Trudinger, and Crawford (2004), the current study found that the overall academic achievement of children with sexual abuse histories was significantly lower than children without sexual abuse histories. Additionally, as expected, higher IQ was signifi cantly associated with higher overall achievement in both the overall sample and the CSA only group. IQ was shown to moderate the relationship between abuse and Total Achievement particularly at medium and low levels of IQ. Number of foster care placements was not found to be significantly associated with any of the variables of interest except for higher number of home placements be ing significantly associated with more total behavior problems, including both inte rnalizing and externalizing behavior problems as reported by caregivers for the overall sample. While it could be expected that number of foster care placements would have less of an influence on intelligence and achievement, it is somewhat surprising that number of foster care placements was not significantly associated with self repo rted depressive symptomatology. Again, while these results should be interpreted in the context of the relatively low power of the measure assessing self reported depressive sy mptoms, these findings are consistent with research by Stanley, et al, 2005 and Pecora, et al (2005) who did not find a relationship between self reported depression and number of home placements. Abuse severity, history of previous psychological treatment and age at time of testing were, contrary to expectations not shown to be significantly associat ed with any of the variables of interest. To date, the research on severity appears to be equivoca l. While
Risk and Protective Factors in Child Sexual Abuse 57 Feinauer and Stuart (1996) and Morrow and Sorell (1989) found abuse severity to be associated with poorer functioning, Paradise, et al (1994) and Feiring, Taska, and Lewis (2002) have, like the current study found no relationship between severity and outcome. It may be that, as with Paradise et al (1994), when examining cases with parti cularly severe abuse, such as this study, the role of severity becomes a less importa nt factor when examining outcome. While history of treatment and age at time of abuse would also se em to contribute to outcomes, the current study did not support this hypothesis. While there have been few studies examining these variables, research by Cavanaugh (2005) similarly found there to be no association between age of onset and outcomes as well as treatm ent history and outcomes. As a group, children with CSA histories demonstrated significantly more emotional and behavioral problems than children without CSA histories. However, given the previous discussion of resilience as positive adaptation in the face of adversi ty; in the present study, many individuals with CSA histories did appear to be resilient to the a buse experience with regard to many of the outcomes measured in the current study. In particular children with higher IQs appeared to demonstrate positive adaptat ion with regard to fewer psychological problems per caregiver report, fewer self reported depressive symptoms, and higher achievement than their counterparts with lower IQs. These results indicate, therefore, that IQ does serve an important protecti ve role for children, both with and without CSA histories. This has important implications for prevention and treatment which are discussed below. Although this study did identify the importance of IQ as a protective factor in high risk children, it does not explain the nature
Risk and Protective Factors in Child Sexual Abuse 58 of this relationship, that is, whether the abuse experience played a role in cogniti ve development or whether these children could have been identified as having lower IQ s prior to being abused and future research could shed light on this area. Several limitations should be considered when interpreting the results of the current study. As previously mentioned, because of the relatively large number of analyses conducted on the data, the possibility of committing Type I errors ma y have been inflated. Additionally, causal relationships cannot be ascertained from the c urrent study given its correlational nature. The archival nature of this study may have influenced the consistency with which data was collected. For example, achievement meas ures included two different versions of two different achievement measures, which may have had some differences in reliability and validity which were not be controlled f or in this study. The problem of missing data was also particularly problematic for the current dataset, in part due to its archival nature. The small sample size of some of the m easures used, serve as another limitation of this study. Ascertaining abuse severity from chart reviews may also prove to be a limitation of the current study, as there may have been other factors that influenced severity score that was not available through c hart reviews. While the CSA sample in the current study was drawn from DCF identified case s of sexual abuse and the non-CSA from clinic charts of children presumed to have no sexual abuse history, it is possible that there may have been children in the non-sexually abus ed group that were actually abused. Furthermore, because it was not possible in the c urrent study to examine familial factors such as parental IQ and personality, it is possible that
Risk and Protective Factors in Child Sexual Abuse 59 some of the results seen in the study were due to genetic rather than environmental factors. There are, however, a number of strengths to this study including the relatively ethnically diverse sample that was used for both the CSA and non-CSA groups. Additionally, a number of studies on children with abuse histories have included either males or females and the current study included both sexes. These factors incr ease the generalizability of these findings. Additionally, the current study represent s a real world sample of children with abuse histories who were assessed within a relatively short time frame of their abuse, which can aid our understanding of possible short term implica tions of CSA. The current study can also be considered clinically meaningful by helping practitioners to examine intellectual, academic and psychological factor s that should be considered when working with children with CSA histories. Furthermore, this stud y also adds to the literature on characteristics that affect the abuse experienc e and as such can aid in the development of prevention and treatment programs. For example, given previous research suggesting that children with lower IQs may be at higher risk for sexual abuse and have more negative consequences once abused prevention programs could be targeted at this population. Specifically, teaching this population about sexua lity and sexual abuse and teaching them strategies to prevent abuse could be an important step toward decreasing rates of abuse and the negative sequelae of abuse. Additional ly, close monitoring of agencies serving these populations should be conducted to decrease the likelihood that individuals with lower IQs will be subjected to abuse perpetra ted by
Risk and Protective Factors in Child Sexual Abuse 60 and within these agencies. Furthermore, teaching individuals with lower IQs who may lack assertiveness skills so as to decrease the likelihood that they will be vic timized may also prove useful. With regard to treatment programs, cognitive behavioral treat ments have often been found to be effective for many of the negative consequences of CSA. However, screening for IQ may be an important adjunct to treatment to more effe ctively individualize treatment plans based on cognitive functioning to determine the extent t o which more cognitive aspects of treatment can be effectively utilized in tr eatment. Given the limitations placed on this study by its archival nature, replicating this study with a sample of identified children with abuse histories would likely provi de important insights. As this study was not able to tease apart the influence of environmental versus genetic factors because of the lack of availability of f amily members for participation, future studies will likely benefit from a more t horough exploration of these influences. It would likely be beneficial to examine these va riables with a more matched sample of children, either inpatient or outpatient.
Risk and Protective Factors in Child Sexual Abuse 61 References Ackerman, J.P. & Lindhiem, O. (2005). Placement instability during infancy predicts inhibitory control difficulties at age five. Poster session presented at the biennial meeting of the Society for Research in Child Development, Atlanta, GA. Achenbach, T.M. & Rescorla, L.A. (2001). Manual for the ASEBA School-Age Forms & Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth & Families. Aiken, L. S., & West, S. G. (1991 ). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Baron, R.M. & Kenny, D.A. (1986). The moderator/mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 (6), 11731182. Banyard, V.L., Williams, L.M. & Siegel, J.A. (2004). Childhood sexual abuse: A gender perspective on context and consequences. Child Maltreatment 9 (3), 223-238. Bartoi, M.G. & Kinder, B.N. (1998). Effects of child and adult sexual abuse on adult sexuality. Journal of Sex & Marital Therapy, 24, 75-90. Black, C.A., & DeBlaissie, R.R. (1993). Sexual abuse in male children and adolescents : Indicator, effects, and treatment. Adolescence, 28 123-133.
Risk and Protective Factors in Child Sexual Abuse 62 Bremner, J.D., Randall, P., Scott, T.M., Capelli, S., Delaney, R., McCarthy, G., & Charney, D.S. (1995). Deficits in short-term memory in adult survivors of childhood abuse. Psychiatry Research, 59 97-107. Browne, A., & Finkelhor, D. (1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99 (1), 66-77. Buckle, S.K., Lancaster, S., Powell, M.B. & Higgins, D.J. (2005). The relationship between child sexual abuse and academic achievement in a sample of adolescent psychiatric inpatients. Child Abuse and Neglect 29 (9), 1031-1047. Cappelleri J.C., Eckenrode J., & Powers J.L. (1993). The epidemiology of child abuse: Findings from the Second National Incidence and Prevalence Study of Child Abuse and Neglect. American Journal of Public Health, 83 (11), 1622-1624. Cavanaugh, C. (2005). Psychological resilience of childhood sexual abuse survivors: A group psychotherapy outcome study. Dissertation Abstracts International 66, 4-B. Cederblad, M., Dahlin, L., Hagnell, O., & Hansson, K. (1995). Intelligence and temperament as protective factors for mental health. A cross-sectiona l and prospective epidemiological study. European Archives of Psychiatry and Clinical Neuroscience, 245, 11-19. Chaffin, M., Silovsky, J.F., & Vaughm ,C. (2005). Temporal concordance of anxiety disorders and child sexual abuse: Implications for direct versus artifactua l
Risk and Protective Factors in Child Sexual Abuse 63 effects of sexual abuse. Journal of Clinical Child and Adolescent Psychology, 34 (2), 210-222. Cohen, J. (1992). A power primer. Psychological Bulletin 112, (1), 155-159. Dhaliwal, G. K., Gauzas, L., Antonowicz, D. H., Ross, R. R. (1996) Adult male survivors of childhood sexual abuse: Prevalence, sexual abuse characteristics, and long-term effects. Clinical Psychology Review, 16, (7), 619-639. Dimock, P. T. (1988). Adult males sexually abused as children: Characteristics and implications for treatment. Journal of Interpersonal Violence, 3 (2), 203-221. Dube, S.R., Anda, R.F., Whitfield, C.L., Brown, D.W., Felitti, V.J., Dong, M. & Giles, W. (2005). Long-term consequences of childhood sexual abuse by gender of victim. American Journal of Preventive Medicine 28 (5), 430-438. Feinauer, L., Mitchell, J., Harper, J. & Dane, S. (1996). The impact of hardiness and severity of childhood sexual abuse on adult adjustment. American Journal of Family Therapy, 24 (3), 206-214. Feinauer, L. & Stuart, D.A. (1996). Blame and resilience in women sexually a bused as children. American Journal of Family Therapy, 24 (1), 31-40. Feiring, C., Taska, L, & Lewis, M. (2002). Adjustment following sexual abuse discove ry: The role of shame and attributional style. Developmental Psychology, 38 (1), 79-92. Finkelhor, D. (1979) Sexually victimized children The Free Press. A division of Macmillan Publishing Co., Inc.
Risk and Protective Factors in Child Sexual Abuse 64 Finkelhor D. (1994). The international epidemiology of child sexual abuse. Child Abuse & Neglect, 18, (5), 409-417. Garnefski, N. & Arends, E. (1998). Sexual abuse and adolescent maladjustment: Differences between male and female victims. Journal of Adolescence 21, 99107. Gorey K.M. & Leslie D.R. (1997). The prevalence of child sexual abuse: Integrati ve review adjustment for potential response and measurement biases. Child Abuse & Neglect, 21, (4), 391-398. Hagan, T. & Smail, D. (1997). Power-mapping-II. Practical application: The ex ample of child sexual abuse. Journal of Community & Applied Social Psychology, 7 269284. Holmbeck, G. N. (2002). Post-hoc probing of significant moderational and mediational effects in studies of pediatric populations. Journal of Pediatric Psychology, 27, 87-96. Inderbitzen-Pisaruk, H., Shawchuck, C.R., & Hoier (1992). Behavioral characterist ics of child victims of sexual abuse: A comparison study. Journal of Clinical Child Psychology, 21 (1), 14-19. Johnson, B.K. & Kenkel, M.B. (1991). Stress, coping, and adjustment in female adolescent incest victims. Child Abuse & Neglect 15, 293-305. Jones, D.A., Trudinger, P., & Crawfird, M. (2004). Intelligence and achievement of children referred following sexual abuse. Journal of Pediatrics and Child Health 40 (8), 455-460.
Risk and Protective Factors in Child Sexual Abuse 65 Kaplow, J.B., Hall, E., Koenen, K.C., Dodge, K.A., & Amaya-Jackson, L. (2005). The development of attention problems in sexually abused children: A path analytic model. Poster session presented at the biennial meeting of the Society for Research in Child Development, Atlanta, GA. Kaufman, A., Flanagan, D.P., Alfonso, V.C., Mascolo, J.T. (2006) Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV). Journal of Psychoeducational Assessment, 24 (3), 278-295. Kendall-Tackett, 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, (1), 164-180. Kisiel, C. L., Lyons, J. S. (2001). Dissociation as a mediator of psychopathology among sexually abused children and adolescents. American Journal of Psychiatry, 158 (7) 1034-1039. King, M., & Woollett, E. (1997). Sexually assaulted males: 115 men consulting a counseling service. Archives of Sexual Behavior, 26 (6) 579-588. Kovacs, M. (1992). Childrens Depression Inventory, manual New York: Multi-Health Systems, Inc. Harford, K. & Kinder, B. (2005). The role of intelligence and coping processes on resilience in adult survivors of childhood sexual abuse. Manuscript submitted for publication. Himelein, M.J. & McElrath, J.V. (1996). Resilient child sexual abuse survivors: Cognitive coping and illusion. Child Abuse & Neglect, 20 (8), 747-758.
Risk and Protective Factors in Child Sexual Abuse 66 Lab, D. D., Feigenbaum, J. D., De Silva, P. (2000) Mental health professionals' attitude s and practices towards male childhood sexual abuse. Child Abuse & Neglect, 24 (3), 391-409. Liem, J.H., James, J.B., OToole, J.G., & Boudewyn, A.C. (1997). Assessing resilience in adults with histories of child sexual abuse. American Journal of Orthopsychiatry 67 (4), 594-606. Loeb, T. B., Williams, J. K., Carmona, J. V., Rivkin, I., Wyatt, G. E., Chin, D., AsuanO'Brien, A. (2002). Child sexual abuse: Associations with the sexual functioning of adolescents and adults. Annual Review of Sex Research, 13, 307-345. Luthar, S.S. & Cicchetti. D. (2000). The construct of resilience: Implications f or interventions and social policies. Development and Psychopathology 12 857885. Luthar, S.S., Cicchetti. D., & Becker, B. (2000). The construct of resilience: A cri tical evaluation and guidelines for future research. Child Development, 71 (3), 543562. Luthar, S.S., Woolston, J.L., Sparrow, S.S., & Zimmerman, L.O. (1995). Adaptive behaviors among psychiatrically hospitalized children: The role of intellige nce and related attributes. Journal of Clinical Child Psychology, 24 (1), 98-108. Luthar, S.S. & Zigler, E. (1992). Intelligence and social competence among high ris k adolescents. Development and Psychopathology, 4, 287-299.
Risk and Protective Factors in Child Sexual Abuse 67 Luthar, S.S., Zigler, E., & Goldstein, D. (1992). Psychosocial adjustment among intellectually gifted adolescents: The role of cognitive-developmental and experiential factors. Journal of Child Psychology and Psychiatry, 33 (2), 361373. Mansell, S., Sobsey, D., & Moskal, R. (1998). Clinical findings among sexually abused children with and without developmental disabilities. Mental Retardation, 36 (1), 12-22. Martelle, Y. & Smith, D.K. (1994, March). Relationship of the WIAT and WJ-R Tests of Achievement in a sample of students referred for learning disabilities. Paper presented at the annual meeting of the National Association of School Psychologists, Seattle, Washington. Martin, G., Bergen, H.A., Richardson, A.S., Roeger, L., & Allison, S. (2004). Sexual abuse and suicidality: Gender differences in a large community sample of adolescents. Child Abuse & Neglect, 28 491-503. Masten, A.S. & 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. Merrill, L.L., Guimond, J.M., Thomsen, C.J., & Milner, J.S. (2003). Child sexual abuse and number of sexual partners in young women: The role of abuse severity, coping style and sexual functioning. Journal of Consulting and Clinical Psychology, 71 (6), 987-996.
Risk and Protective Factors in Child Sexual Abuse 68 Monahan, K. & Forgash, C. (2000). Enhancing the health care experiences of adult female survivors of childhood sexual abuse. Women & Health, 30 (4), 27-41. Morrow, K.B. & Sorell G.T. (1989). Factors affecting self-esteem, depression, a nd negative behaviors in sexually abused female adolescents. Journal of Marriage and the Family, 51 (3), 677-686. Myers, M.F. (1989). Men sexually assaulted as adults and sexually abused as boys Archives of Sexual Behavior, 18 (3), 203-215. Nakano, T, Wenner, M., Inagaki, M., Kugaya, A., Akechi, T., Matsuoka, Y., Sugahara, Y., Imoto, S., Murakami, K., Uchitomi, Y. (2002). Relationship between distressing cancer-related recollections and hippocampal volume in cancer survivors. American Journal of Psychiatry 159 (12), 2087-2093. Needelman, H., Schnoes, C.J., Ellis, C.R. (2006). The new WISC-IV. Journal of Developmental and Behavioral Pediatrics, 27 (2), 127-128. Newman, M.G., Clayton, L., Zuellig, A., Cashman, L., Arnow, B., Dea, R., & Taylor, C.B. (2000). The relationship of childhood sexual abuse and depression with somatic symptoms and medical utilization. Psychological Medicine, 30 (5), 1063-1077. Paolucci, E.O., Genuis, M.L., Violato, C. (2001). A meta-analysis of the published research on the effects of child sexual abuse. The Journal of Psychology, 135, (1), 17-36.
Risk and Protective Factors in Child Sexual Abuse 69 Paradise, J.E., Rose, L., Sleeper, L.A., & Nathanson, M. (1994). Behavior, family function, school performance, and predictors of persistent disturbance in sexually abused children. Pediatrics, 93 (3), 452-459. Pecora, P.J., Kessler, R.C., Williams, J., OBrien, K., Downs, A.C., English, D., White, J., Hiripi, E., White, C.R., Wiggins, T., & Holmes, K.E., Improving family foster care: Findings from the Northwest Foster Care Alumni Study. Seattle, WA: Casey Family Programs Available from http://www.casey.org. Perez, C.M. & Widom, C.S. (1994). Childhood victimization and long-term intellectual and academic outcomes. Child Abuse & Neglect, 18, (8), 617-633. Porter, C., Lawson, J.S. & Bilger, E.D. (2005). Neurobehavioral sequelae of child sexua l abuse. Child Neuropsychology 11, 203-220. Putnam F.W. (2003). Ten-year research update review: Child sexual abuse. Journal of the American Academy of Child & Adolescent Psychiatry, 42 (3), 269-278. Rumstein-McKean, O., & Hunsley, J. (2001). Interpersonal and family functioning of female survivors of childhood sexual abuse. Clinical Psychology Review, 21 (3), 471-490. Russell, D.E. (1986). The secret trauma: Incest in the lives of girls and women. New York: Basic Books. Rutter, M. (1987). Psychosocial resilience and protective mechanisms. American Journal of Orthopsychiatry, 57 (3), 316-331.
Risk and Protective Factors in Child Sexual Abuse 70 Sattler, J.M. (2001). Assessment of children: Cognitive applications (4 th edition). California: Jerome M. Sattler Publisher, Inc., Saylor, C.F., Finch, A.J., Spirito, A. & Bennett, B. (1984). The Childrens Depression Inventory: A systematic evaluation of psychometric properties. Journal of Consulting and Clinical Psychology, 52 955-967. Saywitz, K.J., Mannarino, A.P., Berliner, L., & Cohen, J.A. (2000). Treatment for sexually abused children and adolescents. American Psychologist, 55 (9), 10401049. Smith, M.S. & Smith, M.T. (1999). A stimulus control intervention in the gynecological exam with sexual abuse survivors. Women and Health, 30 (2), 39-51. Spaccarelli, S. & Fuchs, C. (1997). Variability in symptom expression among sexual ly abused girls: Developing multivariate models. Journal of Clinical Child Psychology, 26 (1), 24-35. Spaccarelli, S. & Kim, S. (1995). Resilience criteria and factors associat ed with resilience in sexually abused girls. Child Abuse & Neglect 19 (9), 1171-1182. Stanley, N., Riordan, D., Alaszewski, H. (2005). The mental health of looked after children: matching response to need. Health and Social Care in the Community, 13, (3), 239-248. Tiet, Q.Q., Bird, H.R., Davies, M., Hoven, C., Cohen, P., Jensen, P.S., & Goodman, S. (1998). Adverse life events and resilience. Journal of American Academy of Child & Adolescent Psychiatry, 37 (11), 1191-1200.
Risk and Protective Factors in Child Sexual Abuse 71 Trickett, P.K., McBride-Chang, C., & Putnam, F.W. (1994). The classroom performance and behavior of sexually abused females. Development and Psychopathology, 6, 183-194. Vaillant, G.E. & Davis, T. (2000). Social/Emotional intelligence and midlife resi lience in schoolboys with low tested intelligence. American Journal of Orthopsychiatry, 70 (2), 215-222. Veltman, M.W.M. & Browne, K.D. (2001). Three decades of child maltreatment research: Implications for the school years. Trauma, Violence, & Abuse, 2, (3), 215-239. Watkins, B., & Bentovim, A. (1992). The sexual abuse of male children and adolescents: A review of current research. Journal of Child Psychology and Psychiatry, 33 197-248. Weinstein, D., Staffelbach, D., & Biaggio, M. (2000). Attention-deficit hyperactivi ty disorder and post-traumatic stress disorder: Differential diagnosis in childh ood sexual abuse. Clinical Psychology Review, 20 (3), 359-378. Werner, E. (1994). Overcoming the odds. Developmental and Behavioral Pediatrics, 15 (2), 131-136. Wyatt, G.E. (1990). Sexual abuse of ethnic minority children: identifying dimensions of victimization. Professional Psychology: Research and Practice, 21, (5), 338343.
About the Author Kelli-Lee Harford is originally from Kingston, Jamaica and re ceived her Bachelor of Science in Psychology and Management Studies from the University of the West Indies, Mona Campus. After completing her degree she moved to the United St ates and began to work in the field of psychology in Virginia. To further develop her skills and knowledge, she pursued a Masters degree in Psychology at Teachers College, Columbia University in New York, and later her Doctoral degree at the University of South Florida. Her primary research interest is in the area of resilience in survivors of childhood sexual abuse. Her clinical area of interest is working with children and families particularly those with chronic or life threatening illnesses. After completing her Doctor al degree, she will be assuming a clinical postdoctoral fellowship in pediatric psychology at Arkansas Childrens Hospital in Little Rock, Arkansas.