Child Abuse & Neglect 33 (2009) 353–361
Contents lists available at ScienceDirect
Child Abuse & Neglect
Executive function performance and trauma exposure in a community
sample of children夽
Anne P. DePrincea,∗, Kristin M. Weinzierla, Melody D. Combs b
aUniversity of Denver, Denver, CO, USA
bUniversity of Colorado Denver, USA
Received 8 June 2007
Received in revised form 25 July 2008
Accepted 12 August 2008
Available online 28 May 2009
Objective: Though children exposed to familial violence are reported to have difﬁculties
with a range of emotional and behavioral problems (e.g., lower school achievement) that
implicate executivefunction (EF) deﬁcits, relatively little research has speciﬁcally examined
EF as a function of trauma exposure in children.
Methods: Based on parent report of children’s exposure to potentially traumatic events,
Mean = 10.39)from an ethnically diverse community sample were com-
pared across three trauma-exposure groups: familial trauma, non-familial trauma, and
no trauma. Children completed a battery of tests to assess working memory, behavioral
inhibition, processing speed, auditory attention, and interference control.
Results: Familial trauma (relative to non-familial and no trauma exposure) was associated
with poorer performance on an EF composite (composed of working memory, inhibition,
auditory attention, and processing speed tasks); the effect size was medium. Both trauma-
exposure status and dissociation symptoms explained unique variance in EF performance
after controlling for anxiety symptoms, socio-economic status, and potential traumatic
brain injury. While IQ and EF performance were related, SES predicted unique variance
in IQ (and not EF) scores, while familial-trauma exposure did not.
Conclusions: The contribution of trauma exposure to basic executive functioning held
after taking into account symptoms (anxiety and dissociation), socio-economic status, and
possible traumatic brain injury exposure. EF problems may provide one route via which
maltreated children become at risk for peer, academic, and behavior problems relative to
Practice implications: EF problems may provide one route via which maltreated children
become at risk for peer, academic, psychological, and behavior problems relative to their
peers. Recently, intervention strategies have emerged in the anxiety and mood disorder
treatment literatures that appear to effectively target EFs. As future research continues to
specify the relationship between child trauma exposure and EF performance, these inno-
vative treatments may have important practice implications for addressing EF deﬁcits.
© 2009 Elsevier Ltd. All rights reserved.
Executive functions (EFs) are comprised of such diverse abilities as directing attention (including shifting, inhibiting, and
focusing attention), manipulating information in working memory, and self-monitoring. These functions are critical to goal-
directed behavior, allowing us to maintain, update, and integrate information to “navigate our ever-changing environmental
context” (Willcutt et al., 2005, p. 185). Disruptions in EF have been consistently replicated in samples of adults exposed to
夽This project was funded by the University of Denver PROF Award to DePrince.
∗Corresponding author address: Department of Psychology, University of Denver, 2155 S. Race Street, Denver, CO 80208, USA.
0145-2134/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
354 A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361
trauma (El-Hage, Gaillard, Isingrini, & Belzung, 2006; Navalta, Polcari, Webster, Boghossian, & Teicher, 2006; Stein, Kennedy, &
Twamley,20 02), including those with PTSD (e.g., Foa, Feske,Murdock, & Kozak, 1991; Kremenet al., 20 07;McKenna & Sharma,
1995; Parslow & Jorm, 2007; Uddo, Vasterling, Brailey, & Sutker, 1993) and dissociative (e.g., DePrince & Freyd, 1999; Simeon
et al., 2006) symptoms. However, relatively less research has examined EFs among children exposed to familial trauma (e.g.,
sexual abuse, physical abuse, witnessing domestic violence). Because EFs are central to many of the developmental tasks
children face—from navigating peer relationships to performance in academic settings and behavioral control—the extension
of systematic study of EFs and trauma exposure to children is particularly important.
To date, most approaches to studying EF correlates among children exposed to familial trauma have emphasized attention
to emotional (particularly threat-related) stimuli. For example, children diagnosed with PTSD or high levels of dissociative
symptoms appear to process threat-related information differently than control groups (e.g., Becker-Blease, Freyd, & Pears,
2004; Dalgleish, Moradi, Taghavi, Neshat-Doost, & Yule, 2001; Moradi, Taghavi, Neshat-Doost, Yule, & Dalgleish, 1999).
Similarly, children exposed to severe physical abuse show attention biases towards negatively valenced emotion stimuli
(e.g., Pollak, Cicchetti, Hornung, & Reed, 2000). While biases to emotionally salient information have important implications
for the development and maintenance of psychopathology, this body of work does not address potentially important links
between trauma exposure and other EFs that are critical to goal-directed behaviors, such as working memory, inhibition, and
processing speed (see Willcutt et al., 2005). Further, EFs in response to neutral (i.e., not emotionally salient or threat-related)
stimuli have been understudied and may have important implications for developmental tasks generally.
Indeed, deﬁcits in basic EFs are likely to disrupt a range of developmental tasks, including those with which children
exposed to familial trauma have demonstrated difﬁculties (e.g., school achievement; Eckenrode, Laird, & Doris, 1993; Kendall-
Tackett & Eckenrode, 1996; Shonk & Cicchetti, 2001). For example, working memory, processing speed, and inhibitory skills
are all required to maintain, update, and integrate information (Willcutt et al., 2005). To date, very few studies have examined
basic EF performance in trauma-exposed children. Cromer, Stevens, DePrince, and Pears (2006) reported that higher levels
of dissociation in preschool aged children in foster care (N=24) were associated with deﬁcits in tasks requiring inhibition,
but not with tasks requiring primarily planning, strategy, or multiple rule sets; however, data on children’s trauma exposure
were unavailable. In another study, children diagnosed with maltreatment-related PTSD performed more poorly than their
non-maltreated peers on several EF measures, such as freedom from distractibility and sustained visual attention tasks
(N= 29; Beers & De Bellis, 2002). While this study points to EF problems associated with maltreatment-related PTSD, the
lack of a maltreated no-PTSD group conﬂates PTSD and maltreatment status, making it difﬁcult to determine whether the
observed deﬁcits are related to maltreatment status, PTSD symptoms, or both.
The paucity of research on basic EFs, as well as the inferential problems caused by conﬂating PTSD and violence exposure,
may mask important relationships between familial-trauma exposure (including sexual or physical victimization or wit-
nessing domestic violence) and EF performance in children. For example, chronic stress in the violent family environment
could have an impact on brain regions responsible for EFs, such as the medial prefrontal cortex (mPFC), thus affecting EF
performance. Child maltreatment is associated with dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis (Tarullo
& Gunnar, 2006). Regions that are particularly vulnerable to HPA axis dysregulation, such as the hippocampus, make inputs
to the mPFC (e.g., Ishikawa & Nakamura, 2003) and could, therefore, inﬂuence EF performance. Insults to brain regions
involved in EF caused by mild traumatic brain injuries (TBIs) could also impact EF performance. In adults, at least, family
violence is associated with TBIs. For example, 92% of women in domestic violence shelters reportb eing struck in the head; the
majority complained of EF problems, including distractibility, difﬁculty dividing attention, difﬁculty concentrating, forgetting
appointments, and confusion (Jackson, Philp, Nuttall, & Diller, 2002).
In addition to insults to brain regions responsible for EF, various cognitive strategies to avoid threat-related cues may
contribute to global changes in information processing (DePrince, 2005), including EF performance. Because children exposed
to familial traumas are generally powerless to control the violence or leave the relationship (for a review of related issues,
see Freyd, DePrince, & Gleaves, 2007), ongoing awareness of threat may result in deleterious consequences, such as increased
stress, decreased attachment to caregivers, or increased conﬂict with caregivers. Thus, the ability to decrease attention to
threat cues may help children navigate environments characterized by inescapable harm (Freyd et al., 2007). For example,
children may engage in distraction to avoid threat cues (DePrince & Freyd, 1999) that may have a negative impact on the
development of EFs more generally.
Other environmental stressors could also underlie risk for EF problems among children exposed to familial vio-
lence, such as socio-economic status (SES). For example, previous research has demonstrated that shared environment
explains signiﬁcant variability in IQ scores among children living in poverty (e.g. Turkheimer, Haley, Waldron, D’Onofrio,
& Gottesman, 2003). IQ and EF represent related, though separable, constructs (Friedman et al., 2006). Thus, to the extent
that low SES inﬂuences IQ scores, studies of other related constructs—in this case, EF performance—should take SES into
Current study. The current study tests the prediction that children exposed to familial trauma (including physical abuse,
sexual abuse, and/or witnessing domestic violence) will show poorer EF performance relative to children exposed to non-
familial traumas (e.g., natural disaster, motor vehicle accident) and children exposed to no trauma. Notably, we include a
non-familial-trauma group because of the cross-sectional natureof the study. If, as predicted, familial trauma has unique links
with EF performance, the familial-trauma group will perform differently from the non-familial and no-trauma groups, which
should perform similar to one another. We present and test our hypotheses in terms of weights assigned to corresponding
hypothesized means (see Loftus, 1996). To test the prediction that the familial-trauma group would show deﬁcits in EF
A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361 355
performance relative to children exposed to non-familial traumas or no trauma, planned contrast weights were assigned
as follows: familial trauma= −2, non-familial trauma = 1, no trauma= 1. After testing our initial hypothesis, we examined
the relative contributions of several factors to EF performance, including: familial-trauma-exposure status, anxiety and
dissociation symptoms; presence of mild traumatic brain injuries; SES.
One hundred and fourteen school-aged children were recruited for a two-session study on stress and attention through
ﬂyers in social service and mental health agencies, community centers, and local businesses in a large western city in
the United States. Parents who called about the study (after having seen a ﬂyer that advertised the “Children’s Attention
Research” study) were told that they would be asked to complete questionnaires about their child and family, including the
child’s experience of potentially stressful events. They were also told that their child would be asked to complete a variety
of school-like tasks and a few brief questionnaires.
Of the initial sample, 111 were retained for bothsessions. At session 2, one guardian did not answer questions on children’s
exposure to potentially traumatic events. Thus, data analyses are based on the 110 for whom we had guardian-reported
trauma exposure. Of these 110 children (Age Mean: 10.39; SD: 1.19) whose guardians reported their child’s gender (N= 104),
58% were female. Of the 109 whose guardians reported on child ethnicity and race, 3.7% of children were described as Asian,
30.3% as Black or African-American, 33.7% as Hispanic, 8.3% as Native American, 49.5% as White or Caucasian, and 3.7% as
members of another racial or ethnic group (percentages total over 100% because 23% of guardians reported that their child
was a member of at least two racial or ethnic groups). While the convenience sampling methods used impede our ability
to evaluate the representativeness of our sample, a comparison to the overall ethnic and racial diversity in the study city
suggests we attained reasonable representation on this dimension (Denver residents identify as belonging to the following
groups: 65.3% White, 1.3% Native American, 2.8% Asian, .1% Hawaiian, 19.3% Other or multi-racial, 31.7% Hispanic or Latino;
U.S. Census Bureau, 2008). Table 1 provides demographic variables by trauma-exposure group. Groups did not differ on
demographic variables, with the exception of age [F(2,107)= 3.20, p= .045]. A post hoc Tukey’s HSD test revealed that the
familial-trauma group was older than the non-familial-trauma group (p= .04).
Materials: Cognitive variables
Drawing on other studies of EF in children (e.g., Willcutt et al., 2001, 2005), we administered a battery of tasks to assess
working memory, inhibition, processing speed, interference control, and auditory attention. The speciﬁc tasks used are
described below. Because previous studies have reported that interference control measures such as the Stroop task used in
the current study do not consistently load with other EF tasks (e.g., Willcutt et al., 2001), we used a factor analytic strategy
to examine the underlying factor structure before creating an EF composite.
Several scales of the Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV; Wechsler, 2003), scaled for child
age, were used. To assess working memory, the arithmetic, letter-number sequencing, and digit span subscales were admin-
istered. Processing speed was assessed with the Symbol Search scale. Finally, full scale IQ scores were estimated using the
Block Design and Vocabulary scales.
Behavioral inhibition was assessed with the Gordon Diagnostic System (GDS; Gordon & Barkley, 1998), which requires
children to press a key each time they see a “1” followed by a “9” under two conditions. In the ﬁrst (vigilance) condition,
children see a single stream of numbers. In the second (distractibility) condition, children see three strings of numbers and
must make the key press only after seeing the correct number sequence in the center column of numbers. The task was
scored by calculating the number of commission and omission errors in each condition.
Auditory attention was assessed using the Brief Test of Attention (BTA; Schretlen, Bobholz, & Brandt, 1996), which
requires participants to listen to a recording of several series of letters and numbers being read aloud; following each
series, the child is asked to indicate how many numbers were in the series. Children are not permitted to count on their
Interference control was assessed using a Stroop task. Children were asked to make a key press with their left index ﬁnger
if words appeared in green and with their right index ﬁnger if words appeared in red. They were instructed to ignore the
word meaning and focus only on the color of the words. All children completed a practice block of 10 trials with names as
the stimuli (e.g., ron, sally, kate, bob, and danny). They then completed the test block. Words appeared for 1700 ms with a
Demographic variables by trauma-exposure group.
No traumaaNon-familial traumabFamilial traumacTukey’s HSD test
% Female 67% 57% 53%
% Belonging to racial/ethnic minority group 61% 50% 70%
Age 10.36 (1.34) 10.05 (1.09) 10.70 (1.11) b,c
356 A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361
2000 ms inter-trial interval. Test trials included several word types (e.g., neutral, negative, positive, incongruent). For the
purposes of the current paper, we were concerned only with neutral and incongruent trials. Ten incongruent trials included
the word “red” appearing in green or the word “green” appearing in red. Five neutral trials included the following words:
coffee, hat, curtain, farmer, and button.
Materials: Guardian report
Guardians reported on children’s exposure to potentially traumatic events and current posttraumatic symptoms using
the UCLA PTSD Index (Pynoos, Rodriguez, Steinberg, Stuber, & Frederick, 1998), which reﬂects the symptom criteria for
PTSD from the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV;American Psychiatric Association, 1994). The
measure has been shown to have good reliability (Roussos et al., 2005) and validity (Rodriguez, Steinberg, Saltzman, &
Pynoos, 2001). Guardians were asked to indicate potentially traumatic events to which their children were exposed; the
severity of PTSD symptoms experienced in response to the most distressing traumatic event. Guardians rated symptoms
on a 5-point scale (0 = none of the time; 4= most of the time). A total PTSD symptom score was calculated by summing the
severity score for each symptom. Internal consistency for PTSD symptoms scores was excellent in this sample (Cronbach’s
alpha = .88).
Dissociation was assessed using the Child Dissociative Checklist (CDC; Putnam, 1997), a 20-item guardian-report measure
that assesses multiple types of observable, dissociative behaviors. The CDC has been demonstrated to have high reliability
and validity (Putnam, 1997). Guardians also reported on children’s anxiety problems using the DSM-Anxiety Problems Scale
of the Child Behavior Checklist (CBCL; Achenbach, 1991). The CBCL has been shown to have excellent test-retest reliabil-
ity as well as very good construct validity (Putnam, 1997). Internal consistency was excellent in this sample (Cronbach’s
alpha = .87).
A demographic form asked guardians to report on their occupation, marital status, years of education, and estimated
family income. Occupational prestige was coded based on Hollingshead (1975). If guardians were married, the partner’s
whose education and occupation scores were highest were used in the composite. After transforming occupation, education
and income ratings to z-scores, an SES composite was created by averaging the z-scores. To screen for TBIs, parents were
asked “...has your child ever experienced a head injury (for example after having fallen and hit his/her head)?” Because we
did not have access to medical records to conﬁrm the presence of TBIs, “yes” responses to this question are referred to as
Prior to data collection, all procedures were approved by the University of Denver Institutional Review Board. All partic-
ipants completed an extensive informed consent process; testing took place only after the parent consented and the child
assented. Understanding of consent/assent materials was assessed with a “quiz.” Both child and adult participants had to
answer all quiz questions correctly in order to take part in the study.After the consent/assent procedures, guardians answered
questionnaires in a quiet room. A research assistant was present to answer questions the guardians had while completing
measures, but the assistant did not observe guardians’ responses.
All of the measures of child EF were completed at session 1. Children, tested in a separate room by a graduate research
assistant, were encouraged to take breaks as needed. Children ﬁrst completed the WISC-IV scales. Next, they completed GDS
blocks in the following order: Practice, Vigilance, and Distractibility. The Stroop task was then administered via computer.
Finally, children and guardians completed an extensive debrieﬁng process that involved reporting on their responses to
According to guardian report, 44 children were exposed to physical maltreatment at home, sexual maltreatment by an
adult, and/or witnessing domestic violence (familial-trauma group); 38 children were exposed to non-familial traumas only,
such as natural disasters, motor vehicleaccidents, and/or community/peer violence (non-familial-trauma group); 28 children
were not exposed to trauma (no-trauma group). Of the 44 children categorized in the familial-trauma group, 6 were reported
to have experienced only sexual maltreatment by an adult, but the victim-perpetrator relationship was not speciﬁed. These
children were assigned to the familial-trauma group because sexual maltreatment by an adult was assumed to be more
similar to the experiences of children in the familial-trauma than non-familial-trauma groups.
To help describe the trauma exposure in the sample, Table 2 details the average number of familial and non-familial
traumas reported. Table 2 also provides guardian-report of symptoms on the CDC, UCLA PTSD Index, and DSM-Anxiety
Problems subscale of the CBCL as well as SES and child history of potential TBI. The relationship between the predicted
and observed patterns of means for symptoms is reported as reffectsize (Furr, 2004; Loftus, 1996)inTable 2. Children in the
familial-trauma group were more likely to have experienced a potential TBI relative to their peers [2(1)= 4.12, p= .04].
A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361 357
Descriptive statistics (reported as Mean (SD) unless otherwise noted) for factors relevant to the development of EF problems by group.
No trauma Non-familial trauma Familial trauma reffectsize
Dissociation (CDC) .25 (.34) .24 (.19) .38 (.31) −.24
PTSD (UCLA) – 12.58 (9.54) 16.60 (9.55) −.21
Anxiety (CBCL raw scores) 2.32 (2.16) 2.61 (2.39) 2.73 (2.34) −.05
SES z-score −.03 (.86) .24 (.86) −.16 (.86) .17
IQ estimate 95.75 (17.24) 96.63 (12.46) 89.02 (13.33) .25
% reporting “yes” to TBI screen 7% 19% 30%
Sum familial trauma (0–3) 0 0 1.36 (.61)
Sum non-familial trauma (0–10) 0 1.61 (.79) 1.61 (1.53)
Note: For continuous measures, effect sizes (reffectsize ) were calculated based on the following contrast weights assigned to each group: no trauma= 1,
non-familial trauma = 1, family trauma = −2).
GDS (total commission and omission errors), BTA (total correct), and WISC-IV scales were scored per standard instruc-
tions provided by task developers (see Gordon & Barkley, 1998; Schretlen et al., 1996; Wechsler, 2003). Stroop data were
cleaned to delete all error trials and trials where reaction time was greater than 2000 or less than 200 ms. Reaction times
were brought back to 2.5 SD above each individual’s mean in each condition (DePrince & Freyd, 1999). Mean reaction time
for neutral and incongruent conditions were then calculated. To calculate Stroop interference, the mean reaction time
to neutral words was subtracted from the mean reaction time to incongruent (i.e., red appears in green) words for each
To reduce data to test the EF hypothesis, all measures were transformed to z-scores and scaled such that lower scores
indicated poorer performance. Working memory (arithmetic, letter-number sequencing, and digit span) and inhibition (GDS
distractibility and vigilance errors) composite scores were created by averaging z-scores. Descriptive statistics for these
measures and reffect size (which captures the correspondence between predicted and observed patterns of means; Furr, 2004;
Loftus, 1996) are reported in Table 3. A principal-components analysis (PCA) with the direct oblimin rotation method was
conducted on the ﬁve EF measures (working memory composite, inhibition composite, BTA, Stroop interference, and pro-
cessing speed). The working memory, inhibition, BTA, and processing speed scores resulted in loadings above .60 on a single
component; however the Stroop interference measure did not load above .30. This pattern is consistent with other research
that has found Stroop performance does not load with other EF measures (Willcutt et al., 2001). When the PCA was re-run
without the Stroop, a single EF component emerged. The same solution was obtained using the orthogonal method of rota-
tion. Based on the factor analysis, an EF composite was created by averaging the working memory composite, inhibition
composite, interference control, and processing speed scores. The planned contrast comparing the familial-trauma group
to non-familial-trauma and no trauma groups on EF composite scores was signiﬁcant [t(107) = 3.20, p<.01] and revealed a
medium effect size (reffect size =.30).
Testing multiple contributions to EF performance
Table 4 details bivariate correlations among variables used in the multiple regression analysis. Table 5 provides betas
and t-values for individual contributors for each of the multiple regression analyses reported here. Having established a
link between familial-trauma-exposure status (using the planned contrast weights) and EF performance, we next tested
the relative contributions of familial-trauma-exposure status contrast weights, SES, potential TBI, dissociation, and anxiety
to EF performance. The full model was signiﬁcant [F(5, 101)= 4.15; p= .002, R2= .17]. Familial-trauma status and dissoci-
ation made unique contributions to the prediction of the EF composite scores. Notably, analyses were also conducted
without the six children exposed to only sexual maltreatment by an adult for whom we did not know the victim-
Executive function (Mean (SD)) performance by trauma-exposure group.
No trauma Non-familial trauma Familial trauma reffectsize
Working memory composite .11 (.81) .24 (.72) −.26 (.86) .27
Inhibition composite .18 (.54) .10 (.62) −.25 (1.16) .22
Interference control: Stroopa.14 (.87) .06 (1.12) −.09 (1.00) .09
Auditory attention: Brief Test of Attention .03 (1.13) .20 (.94) −.24 (.96) .18
Processing speed: Symbol Search .16 (1.09) .21 (.83) −.27 (1.07) .22
EF composite .12 (.58) .19 (.53) −.25 (.76) .30
Note: Measures of EF were transformed to z-scores. The Stroop and inhibition composite were multiplied by −1 so that lower scores on all measures
indicate poorer performance. Mean (SD) of transformed scores are reported. In addition, reffectsize to indicate the size of the relationship between the
planned contrast weights (familial trauma= −2; non-familial trauma group= 1; no trauma group= 1) and cognitive factors. Note, positive reffectsize values
indicate greater impairment for the familial-trauma group relative to the non-familial- and no-trauma groups.
aMeasure is not included in EF composite based on factor analysis.
358 A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361
Bivariate correlations among variables used in multiple regression analysis.
Familial-trauma-exposure status Dissociation Anxiety SES Potential TBI PTSD
EF composite .30** −.36*** −.18†.20*−.05 −.29**
Familial-trauma-exposure status −.24*−.05 .17†−.20*−.21†
Dissociation .58*** −.21*.15 .59***
Anxiety −.05 .08 .47***
SES .04 −.03
Potential TBI .06
Note: Correlations with PTSD include participants from the familial- and non-familial-trauma groups only.
perpetrator relationship. Results were comparable when these children were excluded; therefore, they were retained in the
A separate model that included PTSD scores was run with trauma-exposed children only. Because we used a subsample of
children and thus decreased power, SES and potential TBI were dropped from this regression because they failed to explain
unique variance in the previous analysis. The full model was signiﬁcant [F(3, 77) = 6.10; p= .0 01], R2= .19. PTSD symptom
severity did not explain unique variance in EF scores.
To examinewhether EF and IQ were predicted by different variables, which would suggest theytap dif ferentconstructs, we
repeated the ﬁrst regression to examine the relative contribution of our ﬁve theoretically relevant variables to estimated IQ
scores. While the full model was signiﬁcant [F(5, 100) = 4.83; p= .001], R2= .19, a different picture emerged among individual
predictors relative to the model predicting EF composite scores. Speciﬁcally, SES and dissociation (but not familial-trauma
status) made unique contributions to the prediction of estimated IQ.
We did not have information on the age of onset, chronicity, recency, or severity of the potentially traumatic events
to which children were exposed. We reasoned, however, that exposure to different types of potentially traumatic events
reﬂects at least one aspect of trauma-exposure severity. Therefore, we examined the relative contributions of the number of
familial trauma (0–3) and non-familial traumas (0–10) to EF performance in exposed children. The full model was signiﬁcant
[F(2, 79) = 7.53, p=.001, R2= .16]; however, only the number of familial-trauma event types explained unique variance in EF
performance (see Table 5).
Regression coefﬁcients for models predicting EF and IQ performance.
Variable BSE(B)Beta t
Model predicting EF composite (all participants)
Trauma status .10 .04 .21 2.22*
Dissociation −.64 .27 −.28 −2.34*
Anxiety −.003 .03 −.01 −.10
SES .08 .07 .10 1.04
Potential TBI .05 .15 .03 .30
Model predicting EF composite (trauma-exposed participants only)
Trauma status .11 .05 .25 2.29*
Dissociation −.59 .34 −.23 −1.76†
PTSD −.01 .01 −.10 −.82
Model predicting estimated IQ scores (all participants)
Trauma status 1.33 .91 .14 1.46
Dissociation −11.70 5.83 −.23 −2.01*
Anxiety −.24 .70 −.04 −.35
SES 4.15 1.53 .25 2.71**
Potential TBI −.33 3.29 −.01 −.10
Model predicting EF composite scores (trauma-exposed participants only)
Number of familial events −.30 .09 −.36 −3.44**
Number of non-familial events −.09 .06 −.15 −1.48
A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361 359
The current study revealed a medium effect size for the relationship between familial-trauma-exposure status and EF
performance, as assessed by a composite of working memory, inhibition, auditory attention, and processing speed mea-
sures. The direct effect of exposure to familial trauma was maintained even when the contributions of dissociation, anxiety
symptoms (including PTSD for trauma-exposed youth only), socio-economic status, and potential TBI exposure were con-
sidered. Among these factors, only familial-trauma-exposure status and dissociation contributed unique variance to the
prediction of EF performance. Previous research on children has tended to focus on alterations in information processing
associated with emotion stimuli and internalizing symptoms. Strikingly, the current study demonstrates that the relation-
ship between familial-trauma exposure and basic executive functioning (that is, the absence of emotional content) holds
even after taking into account internalizing symptoms, environmental stressors (via SES), and potential TBIs. Thus, children
exposed to family violence show poorer EF performance relative to their peers, even in the absence of trauma-relevant
IQ was signiﬁcantly related to both familial-trauma-exposure status and EF performance; however we did not control
for IQ in analyses of EF for several reasons. To the extent that IQ and the EF composite share variance because of common
measurement method in this study, controlling for IQ would remove that shared variance, which is actually variance of
interest. Further, some researchers have argued against the use of covariates generally in quasi- and non-experimental
designs where groups differ at the outset (e.g., Miller & Chapman, 2001). In lieu of controlling for IQ, we conducted an
additional multiple regression analysis to test whether the pattern of factors that predicted IQ differed from those that
predicted EF. Differences in predictors, we posited, would contribute to our understanding of how (and if) EF and IQ are
separable in this current study. While dissociation yielded comparable effect sizes across the two analyses, two notable
differences were observed. First, SES predicted unique variance in IQ, but not EF. Second, trauma status predicted unique
variance in EF, but not IQ. Thus, these ﬁndings point to a unique relationship between trauma status and EF that is not
observed in, and therefore separable from, general IQ performance.
In the face of relatively little research on dissociation and information processing in children (Cromer et al., 2006), the
observation of medium effect sizes for the relationship with both EF and IQ is particularly noteworthy. While guardians,
teachers, and clinicians may be more likely to notice or look for anxiety-related problems in children exposed to violence,
the current study suggests that dissociation may be especially important to consider in academic settings. This study is
among the ﬁrst in the literature examining basic information processing and dissociation in children.
Because the present study relied on a parent report of child dissociation, it is possible that parents’ recognition of their
children’s difﬁculties in EF inﬂuenced their report of dissociation. Some researchers (e.g., Bruce, Ray, Bruce, Arnett, & Carlson,
2007) have already begun to disentangle the measurement of dissociation and EF, observing that adult high dissociators report
more EF difﬁculties but do not appear to show impaired performance on EF tasks relative to low dissociators. However, further
research will be necessary in order to clarify the role of dissociation and to understand the measurement of dissociation and
EF among children.
Regardless of whether weexamined anxiety symptoms in the full sample (with maximal power) or PTSD symptoms among
the trauma-exposed children, anxiety symptoms did not explain unique variance in EF performance. The ﬁrst regression that
used the full sample included 107 participants and ﬁve predictors; 91 participants would be required to detect a medium
effect size with power = .80. The trauma-exposed only analysis included 81 participants with three predictors; 76 participants
would be required for power= .80. Thus, we had adequate power to detect an effect of anxiety and PTSD in the respective
analyses. At least two possibilities should be considered in interpreting these ﬁndings. First, anxiety and PTSD symptoms
would be more strongly related to EF in a clinical sample. While that may be the case, it is notable that dissociation explained
unique variance in both EF and IQ performance in this non-referred sample. Thus, it does not appear to be the case that
this sample simply did not report the degree of internalizing symptoms necessary to see relationships with information
processing. Second, the shared variance between PTSD and dissociation symptoms may mask either the contributions of
PTSD or the contributions of those PTSD symptoms that are related to dissociation. Future studies should evaluate the
relative contributions of individual PTSD clusters and dissociation to EF performance.
Several limitations should be taken into consideration. The current study relied on guardian-report of child trauma history
and thus may include false negatives given guardians’ potential concerns about mandated reporting. Every effort was made
to minimize inaccurate reports by developing procedures to allow guardians to report on trauma history anonymously;
however, the relationship between familial-trauma exposure and dissociation scores may have been decreased because of
error variance caused by false negatives. Further, given the structure of the UCLA PTSD Index, we did not have data on age
of onset, severity, recency, frequency and/or chronicity of exposure. In exploratory analyses, we did ﬁnd that the number of
familial (but not non-familial) events explained unique variance in EF composite scores. Though a very rudimentary measure
of one aspect of severity (that is, simply the number of different types of events reported), this ﬁnding is consistent with
research on cumulative risk gradients (e.g., Masten & Wright, 1998) and points to the need for future research to carefully
assess and evaluate the contributions of event characteristics to cognitive performance.
The lack of relationship between EF and the potential TBIs may reﬂect that our screen only included a single item. With
more extensive screening, TBI exposure may have related to EF performance. Given that work with women exposed to
domestic violence has demonstrated relationships between mild TBIs and self-reported attention problems (Jackson et al.,
2002), future research with children should include more thorough screening of TBI history. Future research should also
360 A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361
include other factors relevant to healthy EF development. For example, prenatal alcohol/drug exposure could be a factor in
both increasing risk of EF problems and violence exposure.
This cross-sectional study cannot determine the causal direction of the relationship between EF and familial-trauma-
exposure status. It might be the case that poorer EF performance increases risk of trauma exposure, rather than trauma
exposure leading to poorer EF performance. That the non-familial-trauma group performed differently than the family
violence group suggests that poor EF performance did not increase the chances of trauma exposure generally. We cannot
rule out, though, that impaired EF abilities increase risk of violence in the home.
Future directions and implications
Additional research should focus on several areas. First, efforts should be made to better understand possible modiﬁers
of the relationship between trauma exposure and EF, including trauma characteristics (e.g., severity, recency), anxiety, and
dissociation. Second, future research should continue to focus on disentangling the measurement of dissociation and EF.
Third, these ﬁndings must be leveraged for intervention. In recent years, two major intervention strategies that effectively
target EF performance have emerged in the anxiety and mood disorder treatment literatures: attention control training and
mindfulness-based interventions (e.g., Ma & Teasdale, 2004; Mohlman, 2004; Papageorgiou & Wells, 2000; Segal, Williams,
& Teasdale, 2002). Both approaches teach clients better self-regulation of attention, thus demonstrating that EF performance
is malleable. To the extent that risk for externalizing behavior and peer problems are associated with EF deﬁcits, it is critical
to explore treatments that may address executive weakness.
Further, the absence of anxiety and PTSD symptoms as unique contributors to the prediction of EF points to the need
for future research to continue to disentangle violence exposure from PTSD status. When violence-exposure and PTSD are
conﬂated in studies comparing PTSD groups to healthy controls, we cannot be sure whether deleterious consequences are
associated with the experience of violence or internalizing symptoms. In the current non-referred sample, trauma-exposure
status alone was related to poorer EF performance and suggests a different approach to interventions. While resources are
more likely to be dedicated to youthdiagnose d with a problem,the current study provides an explanation for the more general
ﬁnding that children exposed to family violence have greater school difﬁculties, in terms of both academic achievement and
behavioral problems. That is, poorer EF performance may underlie problems regulating behavior and performing in school
settings. Thus, these ﬁndings point to the need for academic interventions that may be targeted broadly, and not just to
children who meet criteria for PTSD.
The current study demonstrated links between EF performance and trauma-exposure status in a community sample of
children. Children exposed to familial-trauma performed more poorly than children exposed to non-familial or no traumas.
Dissociation explained unique variance in EF scores while controlling for other relevant variables. Thus, greater attention
to dissociation and information processing is warranted. Further, the current study highlights the importance of assessing
trauma-exposed children for basic (not threat-related) cognitive problems that may contribute to more general behavioral
and achievement problems in school.
We wish to thank Ann Chu, Rheena Pineda, Jackie Rea, Julia Burrell-Smith, Reilly Anderson and undergraduate research
assistants for project assistance; community agencies for assistance with recruitment; Drs. Kathy Becker-Blease, Martin
Dorahy, Daniel McIntosh, Bruce Pennington, Martha Wadsworth, and Erik Willcutt for helpful feedback and consultation.
Achenbach, T. M. (1991). Manual for the child behavior checklist/4–18 and 1991 proﬁle. Burlington, VT: University of Vermont Department of Psychiatry.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association.
Becker-Blease, K. A., Freyd, J. J., & Pears, K. C. (2004). Preschoolers’ memory for threatening information depends on trauma history and attentional context:
Implications for the development of dissociation. Journal of Trauma and Dissociation,5, 113–131.
Beers, S. R., & De Bellis, M. D. (2002). Neuropsychological function in children with maltreatment-related posttraumatic stress disorder. American Journal of
Bruce, A. S., Ray, W. J., Bruce, J. M., Arnett, P. A., & Carlson, R. A. (2007). The relationship between executive functioning and dissociation. Journal of Clinical
and Experimental Neuropsychology,29(6), 626–633.
Cromer, L. D., Stevens, C., DePrince, A. P., & Pears, K. (2006). The relationship between executive attention and dissociation in children. Journal of Trauma and
Dalgleish, T., Moradi, A. R., Taghavi, M. R., Neshat-Doost, H. T., & Yule, W. (2001). An experimental investigation of hypervigilance for threat in children and
adolescents with post-traumatic stress disorder. Psychological Medicine,31, 541–547.
DePrince, A. P. (2005). Social cognition and revictimization risk. Journal of Trauma and Dissociation,6, 125–141.
DePrince, A. P., & Freyd, J. J. (1999). Dissociation, attention and memory. Psychological Science,10, 449–452.
Eckenrode, J., Laird, M., & Doris, J. (1993). School performance and disciplinary problems among abused and neglected children. Developmental Psychology,
El-Hage, W., Gaillard, P., Isingrini, M., & Belzung, C. (2006). Trauma-related deﬁcits in working memory. Cognitive Neuropsychiatry,11, 33–46.
A.P. DePrince et al. / Child Abuse & Neglect 33 (2009) 353–361 361
Foa, E. B., Feske, U., Murdock, T. B., & Kozak, M. J. (1991). Processing of threat-related information in rape victims. Journal of Abnormal Psychology,100 ,
Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., DeFries, J. C., & Hewitt, J. K. (2006). Not all executive functions are related to intelligence. Psychological
Freyd, J. J., DePrince, A. P., & Gleaves, D. (2007). The state of betrayal trauma theory: Reply to McNally—Conceptual issues and future directions. Memory,15,
Furr, R. M. (2004). Interpreting effect sizes in contrast analysis. Understanding Statistics,3, 1–25.
Gordon, M., & Barkley,R. A. (1998). Psychological testing and observational measures. In R. A. Barkley (Ed.),Attention deﬁcit hyperactivity disorders: A handbook
for clinicians. New York: Guilford Publications.
Hollingshead, A. B. (1975). Four factor index of social status. Unpublished manuscript, Yale University, Department of Sociology.
Ishikawa, A., & Nakamura, S. (2003). Convergence and interaction of hippocampal and amygdalar projections within the prefrontal cortex in the rat. Journal
of Neuroscience,23, 9987–9995.
Jackson, H., Philp, E., Nuttall, R. L., & Diller, L. (2002). Traumatic brain injury: A hidden consequence for battered women. Professional Psychology: Research
and Practice,33, 39–45.
Kendall-Tackett, K. A., & Eckenrode, J. (1996). The effects of neglect on academic achievement and disciplinary problems: A developmental perspective. Child
Abuse & Neglect,20, 161–169.
Kremen, W. S., Koenen, K. C., Boake, C., Purcell, S., Eisen, S. A., Franz, C. E., Tsuang, M. T., & Lyons, M. J. (2007). Pretrauma cognitive ability and riskfor
posttraumatic stress disorder. Archives of General Psychiatry,64, 361–368.
Loftus, G. R. (1996). Psychology will be a much better science when we change the way we analyze data. Current Directions in Psychological Science,1996,
Ma, S. H., & Teasdale, J. D. (2004). Mindfulness-based cognitive therapy for depression:Replication and exploration of differential relapse prevention effects.
Journal of Consulting and Clinical Psychology,72, 31–40.
Masten, A. S., & Wright, M. O. (1998). Cumulative risk and protection models of child maltreatment. Journal of Aggression, Maltreatment, and Trauma,2, 7–30.
McKenna, F.P., & Sharma, D. (1995). Intrusivecognitions: An investigation of the emotional Stroop task. Journal of Experimental Psychology: Learning, Memory,
& Cognition,21, 1595–1607.
Miller, G. A., & Chapman, J. P. (2001). Misunderstanding of covariance. Journal of Abnormal Psychology,110, 40–48.
Mohlman, J. (2004). Attention training as an intervention for anxiety: Review and Rationale. Behavior Therapist,27, 37–41.
Moradi, A. R., Taghavi, M. R., Neshat-Doost, H. T., Yule, W., & Dalgleish, T. (1999). Performance of children and adolescents with PTSD on the Stroop
colour-naming task. Psychological Medicine,29, 415–419.
Navalta, C. P., Polcari, A., Webster, D. M., Boghossian, A., & Teicher, M. H. (2006). Effects of childhood sexual abuse on neuropsychological and cognitive
function in college women. Journal of Neuropsychiatry and Clinical Neuroscience,18, 45–53.
Papageorgiou, C., & Wells, A. (2000). Treatment of recurrent major depression with attention training. Cognitive and Behavioral Practice,7, 407–413.
Parslow, R. A., & Jorm, A. F. (2007). Pretrauma and posttrauma neurocognitive functioning and PTSD symptoms in a community sample of young adults.
American Journal of Psychiatry,164, 509–515.
Pollak, S. D., Cicchetti, D., Hornung, K., & Reed, A. (2000). Recognizing emotion in faces: Developmental effects of child abuse and neglect. Developmental
Putnam, F. (1997). Dissociation in children and adolescents: A developmental perspective. New York: Guilford Press.
Pynoos, R., Rodriguez, N., Steinberg, A., Stuber, M., & Frederick, C. (1998). UCLA PTSD index for DSM IV. Los Angeles: Trauma Psychiatric Services.
Rodriguez, N., Steinberg, A. S., Saltzman, W. S., & Pynoos, R. S. (2001). PTSD index: Preliminary psychometric analyses of child and parent versions. Symposium
conducted at the Annual Meeting of the International Society for Traumatic Stress Studies, New Orleans, LA.
Roussos, A., Goenjian, A. K., Steinberg, A. M., Sotiropoulou, C., Kakaki, M., Kabakos, C., Karagianni, S., & Manouras, V. (2005). Posttraumatic stress and
depressive reactions among children and adolescents after the 1999 earthquake in Ano Liosia, Greece. American Journal of Psychiatry,162 , 530–537.
Schretlen, D., Bobholz, J. H., & Brandt, J. (1996).Development and psychometric properties of the Brief Test of Attention.Clinical Neuropsychologist,10, 80–89.
Segal, Z. V., Williams, J. M. G., & Teasdale, J. G. (2002). Mindfulness-based cognitive therapy for depression: A new approach to relapse prevention.NewYork:
Shonk, S. M., & Cicchetti, D. (2001). Maltreatment, competency deﬁcits, and risk for academic and behavioral maladjustment. Developmental Psychology,37,
Simeon, D., Knutelska, M. E., Putnam, F., Yehuda, R., Schmeidler, J., & Smith, L. (2006). Attention and memory in dissociative disorders, posttraumatic stress
disorder, and healthy volunteers. Annual Meeting, International Society for Traumatic Stress Studies, Hollywood, CA, November 5–7, 2006.
Stein, M. B., Kennedy, C. M., & Twamley, E. W. (2002). Neuropsychological function in female victims of intimate partner violence with and without
posttraumatic stress disorder. Biological Psychiatry,52, 1079–1088.
Tarullo, A. R., & Gunnar, M. R. (2006). Child maltreatment and the developing HPA axis. Hormones and Behavior,50, 632–639.
Turkheimer,E., Haley, A., Waldron, M., D’Onofrio, B., & Gottesman, I. I. (2003). Socioeconomic status modiﬁes heritability of IQ in young children.Psychological
Uddo, M., Vasterling, J. J., Brailey, K., & Sutker, P. B. (1993). Memory and attention in combat-related post-traumatic stress disorder (PTSD). Journal of
Psychopathology and Behavioral Assessment,15, 43–52.
U.S. Census Bureau State and County QuickFacts for Denver, Colorado (2008). Retrieved March 5, 2008 from http://quickfacts.census.gov/qfd/states/
Wechsler, D. (2003). Wechsler Intelligence Scale for Children-Fourth Edition. San Antonio: The Psychological Corporation.
Willcutt, E. G., Brodsky, K., Chhabildas, N., Shanahan, M., Yerys, B., Scott, A., & Pennington, B. F. (2005). The neuropsychology of ADHD: Validity of the
executive function hypothesis. In D. Gozal, & D. L. Molfese (Eds.), Attention deﬁcit hyperactivity disorder: From genes to patients (3rd ed., pp. 185–213).
Totowa, NJ: Humana Press.
Willcutt, E. G., Pennington, B. F., Boada, R., Ogline, J. S., Tunick, R. A., Chhabildas, N. A., & Olson, R. K. (2001). A comparison of the cognitive deﬁcits in reading
disability and attention-deﬁcit/hyperactivity disorder. Journal of Abnormal Psychology,110, 157–172.