A Model of Vulnerability for Adult Sexual

Victimization: The Impact of Attachment, Child

Maltreatment, and Scarred Sexuality
Joan A Reid, Christopher J Sullivan. Violence and Victims. New York:
2009. Vol. 24, Iss. 4; pg. 485, 17 pgs

Abstract (Summary)
Extending previous research, this study utilized structural equation modeling
to examine the effects of poor mother/child attachment, child neglect,
juvenile sexual victimization (JSV), and Finkelhor and Browne's (1985)
proposed construct of traumatic sexualization on vulnerability to adult
sexual victimization. The proposed model was assessed using data drawn
from a sample of African American females involved in a prospective study
of child sexual abuse survivors. This group was matched to similar others
without such history. Findings suggest that child neglect worsens with poor
mother/child attachment, resulting in a greater likelihood of JSV. Both
neglect and JSV impact shaming sexual beliefs and behaviors, contributing
to the risk for adult sexual victimization. This set of variables accounted for
27% of variance in adult sexual victimization. [PUBLICATION
ABSTRACT]

» Jump to indexing (document details)

Full Text
(7504 words)
Copyright Springer Publishing Company 2009

[Headnote]
Extending previous research, this study utilized structural equation modeling to examine the effects
victimization (JSV), and Finkelhor and Browne's (1985) proposed construct of traumatic sexualizatio
assessed using data drawn from a sample of African American females involved in a prospective stu
others without such history. Findings suggest that child neglect worsens with poor mother/child attac
impact shaming sexual beliefs and behaviors, contributing to the risk for adult sexual victimization. T
victimization.
Keywords: sexual abuse; traumatic sexualization; shame; revictimization

Sexual victimization has one of the highest repeat victimization rates of all
types of crime, with a history of juvenile sexual victimization (JSV) placing
individuals at higher risk for adult sexual victimization relative to nonvictims
(for review, see Classen, Palesh, & Aggarwal, 2005). The current study
explores the mechanisms driving this elevated risk of sexual revictimization.
Assessing the dynamics that lead to heightened vulnerability is an important
precursor to providing effective interventions to avert further victimization.
The present study uses structural equation modeling (SEM) to extend
previous research by examining the effects of familial risk factors, JSV, and
Finkelhor and Browne's (1985) construct of traumatic sexualization on
vulnerability for adult sexual victimization. The use of SEM rather than a
single equation model makes it possible to provide extended specification of
key mechanisms by not only identifying key risk markers but also by
examining their antecedent and consequent interrelations.

THEORIES OF REVICTIMIZATION

The underlying mechanism, whether psychological, ecological, or

behavioral, that aligns against victims to elevate their risk for further harm is
not sufficiently understood (for reviews, see Arata, 2002; Classen et al.,
2005; Messman-Moore & Long, 2003). The damage of revictimization,
however, has been empirically documented. It is frequently associated with
increased posttraumatic stress symptoms, greater shame and self-blame for
the abuse, using drugs and alcohol to cope, sexually acting out, isolating
from others, and difficulty with affect regulation (Classen et al., 2005; Filipas
& Ullman, 2006; Messman-Moore, Long, & Siegfried, 2000; Miner, Klotz-
Flitter, & Robinson, 2006). Classen et al. (2005) concluded that cumulative
trauma (e.g., physical abuse, witnessing domestic violence, or emotional
abuse coupled with childhood sexual abuse [CSA]) increases the likelihood
of sexual revictimization. Based on information gathered with the Juvenile
Victimization Questionnaire, Finkelhor, Ormrod, Turner, and Hamby (2005)
found that experiencing multiple victimizations within one year is the norm
for victimized children. According to Finkelhor, Ormond, and Turner (2007),
the clustering of juvenile victimizations most likely has multiple explanations,
including the impact of prior victimizations on later vulnerability. Findings
from Finkelhor et al. (2007) suggest that child maltreatment and sexual
abuse could serve as "gateway" victimizations, leaving a child highly
susceptible to similar future experiences.

Family Functioning

Theorists and empirical researchers, seeking to identify risk factors for

sexual abuse, have examined the effects of family functioning to determine
which conditions result in greater maladjustment in victims of abuse and/or
lead to increased rates of revictimization (for review, see Merrill, Thomsen,
Sinclair, Gold, & Milner, 2001). Researchers have found that the long-term
detrimental effects of sexual abuse are better explained by familyrelated
conflict, decreased cohesion, and reduced family support than by the
severity of the sexual abuse (Classen et al., 2005; Fassler, Amodeo, Griffin,
Clay, & Ellis, 2005).
Emotional Attachment Between Child and Caregiver. Capacity for healthy
attachment is first acquired in infancy as a child interacts with loving and
attentive caregivers (Heide & Solomon, 2006; Perry, 2001). Bowlby (1973,
1980) asserted that this primary attachment would serve as an emotional
template for future relationships. The stability of this initial template across
time and type of relationship, including parental, romantic, and partner
relationships, was considered in a meta-analysis by Fraley (2002) and found
to be moderately stable for the first 19 years of life. Several studies on the
role of attachment in lower-income, African American children have shown
that mother/child attachment is the most important determinant of a child's
enduring pattern of attachment (Liang, Williams, & Siegel, 2006).

Insecure attachment between child and caregiver has been theorized to

precede sexual abuse (Alexander, 1992). Numerous researchers have
documented that abused individuals report poor parental bonding
experiences (Rikhye et al., 2008). Rikhye et al. (2008) reported that
maltreated children were "8.4 times more likely to endorse weak maternal-
bonding experiences than optimal-bonding experiences" (p. 27). In a review
of 19 retrospective surveys on child abuse, Finkelhor (1994) found that
children who experienced parental inadequacy, unavailability, and emotional
deprivation show elevated risk for sexual abuse.

Child Neglect. Child neglect is the most common form of child maltreatment
in the United States, accounting for 63% of all cases (U.S. Department of
Health and Human Services, 2007). Neglected children learn that they
cannot count on others to respond to their needs, have a diminished sense
of self and decreased feelings of personal value, and often perceive that
they are deserving of abuse and powerless to prevent it (Bloom, 2000; Heide
& Solomon, 2006; Solomon & Heide, 2005). Research has shown that
negligent caregivers fail to protect and guide, leading a child to seek
affection and support outside the family, resulting in vulnerability to sexual
abuse (Benedict & Zautra, 1993). According to research on the victim
selection patterns of sex offenders by Beauregard, Proulx, and Rossmo
(2007), "the victim's vulnerability is (an) important factor associated with
choice of specific victims . . . a child with family problems, without
supervision, always on the street and in need of help" (p. 455).

Traumagenic Dynamics

In the same way that primary attachment is theorized to form the emotional
template for future relationships (Bowlby, 1973, 1980; Fraley, 2002; Perry,
2001), Finkelhor and Browne (1985) theorized that a comprehensive
dysfunctional belief system due to CSA may increase the odds of later
abuse in life through the formation of an unhealthy orientation to love or sex.
The traumagenic dynamics theory highlights four factors unique to CSA,
including traumatic sexualization, betrayal, powerlessness, and
stigmatization. Traumatic sexualization may lead individuals to associate sex
with affection or attention, thereby promoting precocious sexual behavior
and increased risk for revictimization. Finkelhor and Browne (1985) describe
traumatic sexualization as the "process in which a child's sexuality . . . is
shaped in a developmentally inappropriate and interpersonally dysfunctional
fashion as a result of sexual abuse" (p. 531). The betrayal component,
resulting from abuse by a trusted individual, leads to social isolation, feelings
of guilt and shame, and relationship difficulties (Finkelhor & Browne, 1985).
Powerlessness is thought to occur following CSA because of the child's
inability to control the abusive situation, evoking fear and anxiety, the need
to feel in control, an inability to cope, and the expectation of being victimized
again (Finkelhor & Browne, 1985). Finally, according to Finkelhor and
Browne (1985), stigmatization describes the development of a negative self-
identity as bad, damaged, impure, guilty, or shamed.

Finkelhor and Browne's (1985) theorized effects of CSA may provide insight
into findings showing that CSA is a risk factor for sexual victimization in
adolescence (Humphrey & White, 2000; Messman-Moore & Long, 2003).
Adolescents face the highest risk for sexual victimization, typically assaulted
by caregivers, acquaintances, or boyfriends (Snyder, 2000), yet research on
adolescent sexual victimization lags behind research on CSA (Kaukinen &
DeMaris, 2005). Adolescent victims are more likely than younger children to
blame themselves for sexual victimizations because of the perception that
they should have been able to prevent the victimization from occurring
(Hunter, Goodwin, & Wilson, 1992).

Previous Sexual Victimization Studies Using SEM or Path Analysis

Three previous studies investigating the etiology of sexual revictimization,

utilizing SEM or path analysis, found support for Finkelhor and Browne's
theory of traumagenic dynamics. Van Bruggen, Runtz, and Kadlec (2006),
using data collected from college students, estimated that CSA, child
psychological maltreatment, and sexual self-esteem and behaviors
explained 14% of the variance in sexual assault. Arata (2000) found that the
relationship between adult sexual assault and CSA was mediated by the
level of physical severity of CSA, self-blame, posttraumatic symptoms, and
consensual sexual behavior. A path analysis model tested by Livingston,
Testa, and VanZile-Tamsen (2007) detected support for a reciprocal
relationship between sexual victimization and low sexual assertiveness (i.e.,
difficulty refusing unwanted sexual advances), providing empirical evidence
for the role of Finkelhor and Browne's (1985) construct of powerlessness.
The Current Study

The model estimated in the current study builds on previous research by

including familial risk factors and psychological and behavioral vulnerabilities
and by using data drawn from a community sample of older African
American women (see Arata, 2000; Livingston et al., 2007; Van Bruggen et
al., 2006). As shown in Figure 1, the study model postulates that poor
mother/child attachment influences child neglect, which then increases the
likelihood of JSV. Subsequently, child neglect and JSV increase shaming
sexual beliefs and behaviors. Finally, adult sexual victimization is directly
affected by JSV and shaming sexual beliefs and behaviors and indirectly
affected by poor mother/child attachment and child neglect.

First, a measurement model was used to assess the presence of two latent
variables: mother/child attachment and child neglect. Next, the full model,
including both latent and observed variables, was examined in terms of its
overall fit. Finally, the structural paths linking key latent and observed
variables relevant to the model hypotheses were examined to assess the
significance, direction, and strength of the relationships proposed here.

METHOD

Participants and Procedures

The data analyzed in this study were drawn from a prospective study of
urban, lowincome African American women who were victims of CSA and a
matched comparison group of women with no history of CSA (Siegel &
Williams, 2000, 2001a, 2001b). McCahill, Meyer (Williams), and Fischman
(1979) conducted the first wave of data collection from 1973 to 1975 on 206
girls ranging in age from 10 months to 12 years who were victims of reported
cases of CSA, underwent forensic examinations, and received treatment at a
particular municipal hospital. Follow-up interviews were conducted in 1990
and 1991 and then again in 1996 and 1997 to investigate the adult
consequences of CSA. Comparable groups of women were identified and
matched to the original victims on the basis of race, age, and date of hospital
visit for the two waves of follow-up interviews. Matched comparisons were
identified by researchers selecting similar girls who were treated in the
emergency room during the same time period for a reason other than CSA
(Siegel & Williams, 2003a).

Researchers contacted 238 women directly during the data collection wave
used for the current study (1996-1997). Of these women, 174 were
interviewed; half the interviewees were from the original sample, and half
were from the matched comparison group (Siegel & Williams, 2003b).
Analyses conducted by Siegel and Williams (2001a) indicated that a
significantly greater percentage of the women from the abused group (42%)
were located and interviewed than from the comparison group (30%). Siegel
and Williams (2001a, 2003b) reported that the dissimilar participation
percentages might stem from the availability of more current addresses for
80 women from the original sample. The face-to-face interviews, lasting
approximately 3 hours, were conducted by four interviewers trained in
developing rapport and asking sensitive questions (Liang et al., 2006; Siegel
& Williams, 2003b). The interview included a series of questions regarding
the participants' victimization history (Siegel & Williams, 2003b).

As would be expected, because of the rate of self-reported CSA in the

general female population (Briere & Elliot, 2003), Siegel and Williams (2000)
reported that within those selected for the matched comparison group, 31%
reported CSA. In this study, as in the previous research by Siegel and
Williams (2003b), rather than eliminating those participants with self-reported
CSA from the study because they are inappropriate for inclusion in the
matched comparison group, both "official" (based on hospital exam) and
self-reports are considered in denoting victims of CSA. In comparing the
official victims and the selfreported victims, no significant differences were
found in terms of the nature of the victimization (Siegel & Williams, 2003b).
For more information on data collection procedures, see Siegel and Williams
(2001a).

The full sample consisted predominantly of African American women (86%).

The average age at the time of the hospital examination was 8.4 years old;
their average age during the 1996-1997 follow-up interview was 31.6 years
old (Siegel & Williams, 2003b). In examining the two groups, the original
researchers concluded that there were no significant differences between
the CSA victims and the matched comparison group in age, race, date of
hospital visit, or socioeconomic status (Siegel & Williams, 2001a).

Measures

Mother/Child Attachment. Liang et al. (2006) developed and utilized a

mother/child attachment index based on four self-reported items, which
recalled the mother/daughter relationship. Drawing on the same four self-
report items, a latent mother/child attachment variable was specified. These
four indicators of attachment were based on participants' responses to these
items: (a) how close I felt to my mother, rated from 1 to 5 with scale
responses ranging from very close to not close; (b) mother was tender-
always, sometimes, or never; (c) I went to mother for help-always,
sometimes, or never; and (d) mother took an interest in how I felt-always,
sometimes, or never. The responses were coded so that higher scores
indicate lower levels of mother/child attachment. Cronbach's alpha for this
scale was .83. Father/child attachment was not included in the model
because of the high amount of missing data, which would have resulted in a
loss of more than 20% of the sample (Siegel & Williams, 2001a).

Child Neglect. The definition of child neglect in this study is based on that
used by the majority of states in the United States as deprivation of
adequate food, clothing, shelter, medical care, or supervision (Child Welfare
Information Gateway, 2007). Based on this definition, one indicator, parental
inability to express love, which had been collected in the original study and
included in previous analyses as part of a parental neglect scale, was not
used in the current case. Consequently, the measurement of child neglect
was based on four questions, asking each subject whether her parents ever
(a) had to leave her home alone, even when they thought an adult should be
there; (b) were unable to make sure she got the food she needed; (c) were
not able to make sure she got to a doctor or hospital when she needed to;
and (d) were so drunk or high they had a problem taking care of her. No was
coded "0" and yes was coded "1."

Cronbach's alpha for this scale was .66. Frequencies of the four observed
behaviors used as indicators for the latent neglect variable revealed a fair
amount of variation, for example, inadequate supervision (35%), inadequate
food (11%), and inadequate medical care (7%). This variation is due to the
fact that certain types of neglectful behavior, such as not taking a child to the
doctor or hospital, occur infrequently and because neglectful behaviors are
very diverse and not necessarily correlated with each other. Straus and
Kantor (2005) suggested that measures of internal consistency, such as
Cronbach's alpha, may be low for child neglect scales. In addition, as shown
in Figure 2, although these items show a slightly lower than adequate
Cronbach's alpha of .70 (Nunnally, 1978), each of the indicators produced a
high item loading (from .78 to .87) when analyzed with confirmatory factor
analysis (CFA).

Shaming Sexual Beliefs and Behaviors. Siegel and Williams (2001a, 2003b)
constructed a measure of Finkelhor and Browne's (1985) concept of
traumatic sexualization from items on Jehu's (1998) Belief Inventory. The
scale measures the level of selfdenigratory or shaming sexual beliefs and
behaviors. Participants were asked whether the following statements were
true or false for them all or most of the time: (a) you get into trouble because
of your sexual behavior; (b) you control others through the use of sex; (c)
you use sex to get something you want or need; (d) in your opinion, no man
would care for you without a sexual relationship; (e) in your opinion, only
bad, worthless guys would be interested in you; and (f) you find yourself in
awkward sexual situations. Responses were summed to create a scale with
values ranging from 0 to 6. Higher scores on the scale indicate higher levels
of shaming sexual beliefs and behaviors. Cronbach's alpha for this scale
was .79 (Siegel & Williams, 2001a). Although this particular variation of
Jehu's (1988) measure was previously used only by Siegel and Williams
(2003b), the full Belief Inventory or variations thereof have been commonly
used in research involving sexual abuse victims to measure self-denigratory
beliefs resultant of sexual abuse (Edmond, Rubin, & Wamback, 1999;
Leach, Freshwater, Aldridge, & Sunderland, 2001; Pitts & Waller, 1993;
Price, Hilsenroth, Petretic-Jackson, & Bonge, 2001; Waller, Ruddock, &
Pitts, 1993).

JSV. As in previous research, sexual victimization occurring before

adulthood was combined into one variable representing JSV (Muehlenhard,
Highby, Lee, Bryan, & Dodrill, 1998; Senn, Carey, Vanable, Coury-Doniger,
& Urban, 2007). Child sexual abuse "included incidents involving genital
contact (including fondling), force, or sexual contact with someone who was
five years older than the respondent when she was younger than 13" (Siegel
& Williams, 2003b, p. 912). "Incidents, between ages 13 and 17 that involved
sexual contact and force or that involved genital contact that she considered
nonconsensual with someone five or more years older" were defined as
adolescent sexual victimization (Siegel & Williams, 2003b, p. 912). Although
some researchers have not included sexual victimization by a peer when
defining adolescent sexual victimization, others researchers have, insisting
that some sex offenders are minors and that forced sex is still traumatic and
abusive, even if perpetrated by a peer (Muehlenhard et al., 1998). Incidents
that did not involve genital contact were not counted as JSV in this study
(Banyard, Williams, & Siegel, 2003).

Abundant research has explored and substantiated the link between

childhood and/or adolescent sexual victimization and revictimization in
adulthood, and both are considered risk factors (Classen et al., 2005). At
which life stage sexual victimization is more detrimental or results in a higher
likelihood of revictimization is not clear (Classen et al., 2005; Downs, 1993;
Kaukinen & DeMaris, 2005). For that reason, women never sexually
victimized before the age of 18 were coded "0," and participants reporting
either childhood or adolescent sexual victimization were coded "1."
Research has consistently found that revictimization or cumulative trauma
leads to a higher likelihood of further victimization (Arata, 2002; Classen et
al., 2005; Humphrey & White, 2000; Siegel & Williams, 2003b); therefore,
"double victims" (Siegel & Williams, 2001a, p. 84) were coded "2."

Adult Sexual Victimization. The dichotomous adult sexual victimization

variable identifies women who reported sexual victimization at 18 years of
age or older. Participants were asked a series of questions about unwanted
sexual experiences that ranged from fondling to penetration (Banyard et al.,
2003). Those who reported such unwanted adult sexual experiences were
coded "1," and those who did not report adult sexual victimization were
coded "0." As with the JSV variable, for the purposes of this study, incidents
that did not involve genital contact were not counted as adult sexual
victimization. In a review of variants used in measuring of adult sexual
victimization, the one used in this study is generally considered
"intermediate," with some researchers using a broader designation by
including noncontact sexual victimization (e.g., exhibitionism) and nongenital
contact (e.g., kissing, hugging) and others including only sexual intercourse
resulting from force or use of a weapon (Muehlenhard et al., 1998).

Analytic Strategy

The model pictured in Figure 1 was tested via SEM using the Mplus program
(Muthén & Muthén, 1998-2007) with the weighted least squares mean and
variance adjusted (WLSMV) estimator in order to ascertain how well the
proposed model "fits" or adequately replicates the observed relationships
between variables (Kline, 2005). The WLSMV estimator was utilized
because it can fit latent variable models to data sets that contain observed
ordinal and dichotomous variables (i.e., noncontinuously distributed
measures). This technique adjusts estimates with a weighting process and
also offers robust standard errors for use in hypothesis tests (DiStefano &
Hess, 2005). In the present study, this made it possible to conduct CFA and
SEM with all essential items.

As is typical in evaluating SEM results, an array of tests with designated

cutoff values was used to assess different aspects of model fit (Hu & Bentler,
1999; Kline, 2005). Chisquare is a statistic that tests the degree of misfit
between the hypothesized model and a null model where all variables are
assumed to be uncorrelated. A nonsignificant chisquare suggests that the
model fits the data adequately. Other indicators of good model fit are
normed chi-square (NC) of 2 or less, comparative fit index (CFI) of .95 or
higher, Tucker-Lewis index (TLI) of .95 or higher, root mean square of
approximation (RMSEA) of .06 or less, and weighted root mean square
residual (WRMR) of .90 or less.

Prior to testing the full SEM model, CFA was conducted. CFA estimates the
measurement model and its paths, specifying the latent variables' structures
without reference to the hypothesized paths among the key variables
(Anderson & Gerbing, 1988). This is a necessary first step, as it allows for
some assessment of the quality of the measures relative to their expected
properties prior to considering the substantive hypotheses. Adequate model
fit, together with strong and significant loadings of the manifest indicators on
their respective latent variables, provides support for the measurement
model (Kline, 2005). After establishing overall model fit, individual
measurement and structural paths were assessed in relation to the study
hypotheses.

RESULTS

Descriptive statistics of the study variables are summarized in Table 1. Adult

sexual victimization was reported by 43% of the women. Twenty-six percent
of the sample reported no JSV; 55% reported one type of JSV, either child or
adolescent sexual victimization; and 18% reported both types of JSV, child
and adolescent sexual victimization. Because of the data collection
procedures and matched research design (i.e., matching to an original
sample of all CSA victims), these values are not reflective of general
population rates for sexual victimization.

Results for all bivariate analyses are displayed in Table 1. Significant,

positive relationships were found among many key variables. For example,
inadequate parental care due to drug or alcohol use was significantly related
to 10 of the 11 variables in the model, with small to moderate effect sizes.
JSV was significantly and positively related to all four indicators of parental
neglect, with small effect sizes. Shaming sexual beliefs and behaviors were
significantly and positively related to 9 of the 11 observed variables, with
small to moderate effect sizes.

As shown in Figure 2, the CFA model for the two latent variables evidenced
an adequate fit to the data. The chi-square was nonsignificant, ÷2(11) =
16.09, p = .14, and all other indices demonstrated good fit as well (NC =
1.46; CFI = .99; TLI = .99; RMSEA = .05; WRMR = .59). The loadings of the
indicators on the latent variables were statistically significant and strong,
ranging from .78 to .92. These results provided support for the measurement
model, suggesting a reasonable foundation for testing structural
relationships. The full SEM model also evidenced adequate fit to the data
with a nonsignificant chi-square, ÷2(18) = 18.12, p = .45. Again, all indices
show good fit (NC = 1.01; CFI = 1.00; TLI = 1.00; RMSEA = .01; WRMR =
.58).

The structural paths linking key latent and observed variables relevant to the
model hypotheses were examined to assess the proposed relationships. The
standardized coefficients of each key path are displayed in Figure 2.
Standardized linear regression coefficients are reported for all dependent
variables in the model except for adult sexual victimization. Adult sexual
victimization is a dichotomous variable, and consequently the estimate takes
the form of a conditional probability based on the value(s) of predictor
variables (Liao, 1994). Probit coefficients are reported for those
relationships. For further reference, the unstandardized coefficients,
standardized coefficients, and standard errors for all model estimates are
reported in Table 2.

Dysfunctional mother/child attachment was positively related to child neglect

with a large direct effect size (standardized coefficient = .50, p < .01). The
indirect effect of dysfunctional mother/child attachment on adult sexual
victimization as mediated by neglect, JSV, and shaming sexual beliefs and
behaviors revealed a significant and positive relationship with a small effect
size (standardized coefficient = .13, p < .05).

Child neglect was positively related to JSV with a moderate effect size
(standard coefficient = .40, p < .05) and shaming sexual beliefs and
behaviors (traumatic sexualization) with a moderate effect size (standardized
coefficient = .33, p < .05). The indirect effect of child neglect on adult sexual
victimization as mediated by JSV and shaming sexual beliefs and behaviors
revealed a significant, positive relationship with a moderate effect size
(standardized coefficient = .26, p < .05).

JSV was positively related to shaming sexual beliefs and behaviors with a
small direct effect size (standardized coefficient = .20) and with a critical
value slightly lower than the predetermined significance level for the analysis
(z = 1.81, p = .08). In addition, JSV had a significant and positive direct
relationship with adult sexual victimization with a moderate effect size
(standardized coefficient = 28, p < .05). Finally, shaming sexual beliefs and
behaviors had a significant and positive direct relationship with adult sexual
victimization with a moderate effect size (standardized coefficient = .36, p <
.01).

The proportion of variance explained in the endogenous variables are noted

in Figure 2 and were as follows: child neglect, R2 = .25, JSV, R2 = .16;
shaming sexual beliefs and behaviors, R2 = .20; and adult sexual
victimization, R2 = .27. Structural paths connecting the key latent and
observed variables supported the study hypotheses, with only one proposed
path coefficient failing to achieve the predetermined significance value for a
two-tailed hypothesis test.

DISCUSSION

Overall, in the model assessed here, the proportion of variance explained in

adult sexual victimization was .27. Past models focused on the etiology of
adult sexual victimization found that variance explained proportions ranging
from .14 to .19 (Arata, 2000; Livingston et al., 2007; Van Bruggen et al.,
2006). Use of SEM, rather than single-equation models, allows for extended
specification of key mechanisms in the theory proposed by Finkelhor and
Browne (1985). In this case, such an approach led to the finding that
traumatic sexualization is a possible consequence and a predictor of sexual
victimization.

Beginning with the effects of mother/child attachment on child neglect, this

study sought to answer the question of whether disturbed mother/child
attachment impacts child neglect, which subsequently opens the gateway for
further harm. The study results revealed that a mother who lacks healthy
emotional attachment with her child is less likely to provide adequate
physical care and supervision than a mother who is closely attached to her
child. Poor mother/child attachment was also shown to have a weak indirect
effect on adult sexual victimization through neglect, JSV, and shaming
sexual beliefs and behaviors, indicating that primary attachments may form
an emotional template with enduring influences on future relationships
(Bowlby, 1973, 1980; Fraley, 2002; Perry, 2001).

Confirming the study hypotheses, child neglect influenced the occurrence of

JSV and the risk of adult sexual victimization (mediated by JSV and shaming
sexual beliefs and behaviors). Neglect emerges as a potential "gateway"
victimization that might open one up to further harm (Finkelhor et al., 2007)
and validates the prior research findings that vulnerability and neediness are
key factors in victim selection by sex offenders (Beauregard et al., 2007). As
theorized by Finkelhor and Browne (1985), this study demonstrated that JSV
generates shaming sexual beliefs and behaviors (traumatic sexualization).
Unexpectedly, child neglect affected the level of shaming sexual beliefs and
behaviors more than JSV, confirming findings that child neglect may be
more detrimental than abuse, resulting in a diminished sense of self and
decreased feelings of personal value (Bloom, 2000; Heide & Solomon, 2006;
Solomon & Heide, 2005). Moreover, this effect may indicate that a "more-
generalized abuse-related process" is operating (Van Bruggen et al., 2006,
p. 142), resulting in shamed-based relationship patterns, and that some
individuals neglected as children develop dysfunctional beliefs about self as
undeserving of love or loving relationships (Bloom, 2000).

Finally, both JSV and shaming sexual beliefs and behaviors contributed to
vulnerability to sexual victimization in adulthood. This outcome suggests that
sexual abuse may hinder self protective capability, corroborating the
disclosures of revictimized women in other studies who felt "so full of self-
blame and shame from the original assault that they felt unable to act on
their own behalf during the later sexual assault" (Anderson, 2004, p. 3). This
finding also authenticates a survey of serial rapists, in which 69% identified
vulnerability or "easy prey" as the strongest reason to attack a female; youth,
helpfulness, easy compliance, and those exhibiting "a learned helplessness"
or decreased defense capabilities were characteristics of vulnerable targets
(Stevens, 1998, pp. 55, 60).

Limitations

Despite the importance of these findings, there are some study limitations
that must further contextualize the results. The lack of cultural and gender
diversity in the sample limits the generalizability to men or all women.
Replication of these findings among other races is desirable. These results,
however, are similar to those found by Van Bruggen et al. (2006) in their
study using data drawn from predominantly European Canadian college
students. In addition, replication of these findings across races and
ethnicities would be expected, as a review of literature related to how CSA
and abuse experiences differentially affect African American women found
that, in the majority of studies, there were no differences in rates of rape,
sexual revictimization, or the resulting psychological consequences across
race (Classen et al., 2005; Kilpatrick, 2002; Miner et al., 2006; West,
Williams, & Siegel, 2000).

This study was based on secondary data analysis, which created limitations
in model specification, possibly resulting in omitted variable bias. For
example, child physical abuse has been found to increase the likelihood of
revictimization (Classen et al., 2005). However, physical abuse was not
included in this model because of the very broad definition given to child
physical abuse in this data set. In addition, measures were based primarily
on recollections from childhood, affecting the certainty of the temporal
relationships between measures; such data may not be as reliable as data
collected at the time of its occurrence (Finkelhor, 1994; Siegel & Williams,
2003b). The study was based on matched sampling design, and in spite of
the careful conducting of matching procedures, there remains the possibility
that some critical and undetected difference exists between the groups
(Rossi, Lipsey, & Freeman, 2004).

A limitation of SEM is that alternative models can adequately fit the data
(Kline, 2005). In this study, other possible specifications of relationships
between key variables were considered when selecting the study model. For
instance, poor mother/child attachment could have been specified as the
result (or a correlate) of child neglect rather than a precedent. However,
attachment between child and primary caregiver has been found to form an
emotional template beginning in the first year of life (Bowlby, 1973, 1980;
Farley, 2002; Perry, 2001). Therefore, child neglect was placed subsequent
to attachment.
Implications for Treatment Providers

Children who experience poor attachment with their primary caregiver

should be identified as being potentially at risk for child neglect and thereby
possibly susceptible to further abuse. Young children whose caregivers are
absent or neglectful because of poor parenting skills, physical or mental
illness, drug or alcohol abuse, death, or incarceration, or children sheltered
by child protective services should be protected (DiLauro, 2004). As soon as
these potentially at-risk children are able, they should be provided with
supplemental information on the prevention of sexual abuse, tactics of
sexual offenders, and information on what to do if they are abused.

Caregivers and surrogate caregivers need to be informed of the importance

of healthy emotional attachment and given suggestions on how to bond to
their children. Foster caregivers may require training to recognize and
respond therapeutically to the special needs of foster children (Cicchetti,
Rogosch, & Toth, 2006). Alternative opportunities for neglected children to
bond with surrogate caregivers (e.g., mentors, teachers, coaches), under
appropriate supervision and fully implemented abuse prevention safeguards,
could fill the gaps created by unavailable or inadequate primary caregivers.

Children and adults who have experienced victimization are worthy of

protection from further harm. Treatment for the survivors of childhood sexual
victimization or neglect should include techniques centered on reprocessing
cognitions of self-blame for past abuse, altering ideas that the self is
undeserving of love, and shaping new beliefs about the value of the self in
comparison to others. Clients would benefit from information on how the
psychological effects of childhood neglect or victimizations (e.g., feelings of
selfblame or lack of self-efficacy) may influence self-protective behaviors,
thereby placing them at an elevated risk for revictimization. To facilitate such
treatment, the development of training and information for therapists and
advocates is needed to assist them in identifying and resolving
vulnerabilities in clients.

Recommendations for Further Research

Vulnerability to juvenile victimization can be fully understood and possibly

reduced only if the interrelationships between various types of victimization
are understood. Research is needed to further illuminate the clustering or
patterns of juvenile victimization and to examine possible correlates with
certain types of victimization patterns. Results of this study indicating
potential influence of mother/child attachment on neglect and sexual
victimization may be reflective of Finkelhor and Browne's construct of
betrayal. Although Finkelhor and Browne (1985) focused primarily on the
betrayal that the victim of sexual abuse feels if a trusted person sexually
victimizes her or him, victims of sexual abuse can develop deeper feelings of
anger regarding the betrayal by the nonoffending parent (Knauer, 2002)
because of their lack of protection and belief. Further research is needed to
determine whether betrayal by the nonoffending parent is more detrimental
to a sexual abuse victim than the betrayal by the actual offender.

The current study focused primarily on how individuals' past experiences,

together with psychosocial and psychosexual functioning, impact their risk of
adult sexual victimization. Repeat victims of sexual violence can be
especially susceptible to society's tendency to victim blame given that it
replicates the core messages of blame and guilt often embedded in their
experience as juvenile victims (Finkelhor & Browne, 1985). First and
foremost, the unwavering message to victims of sexual violence is that only
a perpetrator can prevent a sexual assault from taking place. In studying the
risk of repeat victimization, the goal of both researchers and advocates is to
increase potential victims' capacity for avoidance of victimization without
implying responsibility (Boney-McCoy & Finkelhor, 1995). Empirically based
explanations regarding vulnerabilities due to undeserved life experiences
and their effects could help limit self-blame and convey accurate advice
regarding ways to enhance safeguards to avert future victimizations.

[Reference] » View reference page with links

REFERENCES
Alexander, P. C. (1992). Application of attachment theory to the study of sexual abuse. Journal of C
Anderson, J. (Ed.). (2004). Sexual assault revictimization. Research and Advocacy Digest WCSA, 6
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and rec
Arata, C. M. (2000). From child victim to adult victim: A model for predicting sexual assault. Child Ma
Arata, C. M. (2002). CSA and sexual revictimization. Clinical Psychology: Science and Practice, 9(2
Banyard, V. L., Williams, L. M., & Siegel, J. A. (2003). The impact of complex trauma and depressio
Child Maltreatment, 8(4), 334-349.
Beauregard, E., Proulx, J., & Rossmo, K. (2007). A descriptive model of the hunting process of seria
22(6), 449-463.
Benedict, L. W., & Zautra, A. J. (1993). Family environmental characteristics as risk factors for child
Bloom, S. L. (2000). Email from America: The neglect of neglect: Part I. Psychotherapy Review, 2(5
Boney-McCoy, S., & Finkelhor, D. (1995) Prior victimization: A risk factor for child sexual abuse and
Abuse and Neglect, 19, 1401-1421.
Bowlby, J. (1973). Attachment and loss: Vol. II, Separation: Anxiety and anger. New York: Basic Bo
Bowlby, J. (1980). Attachment and loss: Vol. III, Loss: Sadness and depression. New York: Basic Bo
Briere, J. N., & Elliot, D. M. (2003). Prevalence and psychological sequence of self-reported childho
Neglect, 27(10), 1205-1222.
Child Welfare Information Gateway. (2007). State statutes series: Definitions of child abuse and neg
http://www.childwelfare.gov/systemwide/laws_policies/statutes/define.cfm
Cicchetti, D., Rogosch, F. A., & Toth, S. L. (2006). Fostering secure attachments in infants in maltre
Psychopathology, 18, 623-649.
Classen C. C., Palesh, O. G., & Aggarwal, R. (2005). Sexual revictimization: A review of the empiric
DiLauro, M. D. (2004). Psychosocial factors associated with types of child maltreatment. Child Welfa
DiStefano, C., & Hess, B. (2005). Using confirmatory factor analysis for construct validation: An emp
Downs, W. R. (1993). Developmental considerations for the effects of childhood sexual abuse. Jour
Edmond, T., Rubin, A., & Wamback, K. G. (1999). The effectiveness of EMDR with adult female sur
Fassler, I. R., Amodeo, M., Griffin, M. L., Clay, C. M., & Ellis, M. A. (2005). Predicting long-term outc
severity versus family environment. Child Abuse and Neglect, 29, 269-284.
Filipas, H. H., & Ullman, S. E. (2006). Childhood sexual abuse, coping responses, self-blame, posttr
Interpersonal Violence, 21(5), 652-672.
Finkelhor, D. (1994). Current information on the scope and nature of child sexual abuse. The Future
Finkelhor, D., & Browne, A. (1985). The traumatic impact of CSA: A conceptualization. American Jo
Finkelhor, D., Ormrod, H. A., & Turner, H. A. (2007). Poly-victimization and trauma in a national long
Finkelhor, D., Ormond, H. A., Turner, H. A., & Hamby, S. L. (2005). The victimization of children and
25.
Fraley, R. C. (2002). Attachment stability from infancy to adulthood: Meta-analysis and dynamic mo
Review, 6(2), 123-151.
Heide, K. M., & Solomon, E. P. (2006). Biology, childhood trauma and murder: Rethinking justice. In
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conven
55.
Humphrey, J. A., & White, J. W. (2000). Women's vulnerability to sexual assault from adolescence t
Hunter, J. A., Goodwin, D. W., & Wilson, R. J. (1992). Attributions of blame in child sexual abuse vic
Abuse, 1(3), 75-89.
Jehu, D. (1988). Beyond sexual abuse. New York: Wiley.
Kaukinen, C., & DeMaris, A. (2005). Age at first sexual assault and current substance use and depr
Kilpatrick, D. G. (2002, March). Making sense of rape in America: Where do the numbers come from
Women Supplement Meeting, Centers for Disease Control, Atlanta, GA.
Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York: Gui
Knauer, S. (2002). Recovering from sexual abuse, addictions and compulsive behaviors: "Numb" su
Leach, C., Freshwater, K., Aldridge, J., & Sunderland, J. (2001). Analysis of repertory grids in clinica
Liang, B., Williams, L. M. & Siegel, J. A. (2006). Relational outcomes of childhood sexual trauma in
Liao, T. F. (1994). Interpreting probability models: Logit, probit, and other generalized linear models
Livingston, J. A., Testa, M., & VanZile-Tamsen, C. (2007). The reciprocal relationship between sexu
13(3), 298-313.
McCahill, T., Meyer (Williams), L., & Fischman, A. S. (1979). The aftermath of rape. Lexington, MA:
Merrill, L. L., Thomsen, C. J., Sinclair, B. B., Gold S. R., & Milner, J. S. (2001). Predicting the impac
support, and coping strategies. Journal of Consulting and Clinical Psychology, 69(6), 992-1006.
Messman-Moore, T. L., & Long, P. J. (2003). The role of childhood sexual abuse sequelae in the se
reformulation. Clinical Psychology Review, 23(4), 537-571.
Messman-Moore, T. L., Long, P. J., & Siegfried, N. J. (2000). The revictimization of child sexual abu
child sexual abuse, adult sexual assault, and adult physical abuse. Child Maltreatment, 5(1), 18-27.
Miner, M. H., Klotz-Flitter, J. M., & Robinson, B. E. (2006). Association of sexual revictimization with
21(4), 503-524.
Muehlenhard, C. L., Highby, B. J., Lee, R. S., Bryan, T. S., & Dodrill, W. A. (1998). The sexual revic
literature. Annual Review of Sex Research, 9, 177-223.
Muthén, L. K., & Muthén, B. O. (1998-2007). Mplus user's guide (4th ed.). Los Angeles: Muthén & M
Nunnally, J. (1978). Psychometric theory. New York: McGraw-Hill.
Perry, B. D. (2001). Bonding and attachment in maltreated children: Consequence of emotional neg
Trauma Academy Press. Retrieved April 9, 2008, from http://www.childtrauma.org
Pitts, C., & Waller, G. (1993). Self-denigratory beliefs following sexual abuse: Association with the s
Disorders, 13(4), 407-410.
Price, J. L., Hilsenroth, M. J., Petretic-Jackson, P. A., & Bonge, D. (2001). A review of individual psy
Clinical Psychology Review, 21(7), 1095-1121.
Rikhye, K., Turka, A. R., Kelly, M. M., Gagne, G. G., Jr., Mello, A., F., Mello, M. F., et al. (2008). Inte
effects: Impact on quality of life. Child Abuse and Neglect, 32, 19-34.
Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004). Evaluation: A systematic approach (7th ed.). T
Senn, T. E., Carey, M. P., Vanable, P. A., Coury-Doniger, P., & Urban, M. (2007). Characteristics of
behavior in adulthood. Archives of Sexual Behavior, 36(5), 637-645.
Siegel, J. A., & Williams, L. M. (2001a). Risk factors for violent victimization of women: A prospectiv
Justice, National Institute of Justice.
Siegel, J. A., & Williams, L. M. (2001b). Risk factors for violent victimization of women in a major no
v1. Camden, NJ: Rutgers University; Wellesley, MA: Wellesley College [producers]. Ann Arbor, MI:
Siegel, J. A., & Williams, L. M. (2003a). The relationship between child sexual abuse and female de
and Delinquency, 40(1), 71-94.
Siegel, J. A., & Williams, L. M. (2003b). Risk factors for sexual victimization of women: Results from
Snyder, H. N. (2000). Sexual assault of young children as reported to law enforcement: Victim, incid
of Justice Statistics, U.S. Department of Justice.
Solomon, E. P., & Heide, K. M. (2005). The biology of trauma: Implications for treatment. Journal of
Stevens, D. J. (1998). Inside the mind of a serial rapist. Bethesda, MD: Austin & Winfield.
Straus, M. A., & Kantor, G. K. (2005). Definition and measurement of neglectful behavior: some prin
U.S. Department of Health and Human Services, Administration on Children, Youth and Families. (2
Printing Office.
Van Bruggen, L. K., Runtz, M. G., & Kadlec, H. (2006). Sexual revictimization: The role of sexual se
131-145.
Waller, G., Ruddock, A., & Pitts, C. (1993). When is sexual abuse relevant to bulimic disorders? The
143-151.
West, C. M., Williams, L. M., & Siegel, J. A. (2000). Adult sexual revictimization among black women
consequences of abuse. Child Maltreatment, 5, 49-57.

[Author Affiliation]
Joan A. Reid, MA
University of South Florida
Christopher J. Sullivan,
PhD
University of Cincinnati

[Author Affiliation]
Correspondence regarding this article should be directed to Joan A. Reid, MA, University of South F