Journal of Applied Developmental Psychology 31 (2010) 109–117

Contents lists available at ScienceDirect

Journal of Applied Developmental Psychology

Effortful control and adaptive functioning of homeless children: Variable-focused andperson-focused analyses

Jelena Obradović ⁎Stanford University, School of Education, 485 Lasuen Mall, Stanford, CA 94305-3096, United States

⁎ Tel.: +1 650 725 1250; fax: +1 650 725 7412.E-mail address: [email protected].

0193-3973/$ – see front matter © 2009 Elsevier Inc. Aldoi:10.1016/j.appdev.2009.09.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 October 2008Received in revised form 22 September 2009Accepted 29 September 2009Available online 27 November 2009

Keywords:Homeless childrenEffortful controlExecutive functionsAdaptationRiskResilience

Homeless children show significant developmental delays across major domains of adaptation, yet researchon protective processes that may contribute to resilient adaptation in this highly disadvantaged group ofchildren is extremely rare. This study examined the role of effortful control for adaption in 58 homelesschildren, ages 5–6, during their transition to school. Effortful control skills were assessed using children'sperformance on four standard laboratory tasks. Adaptive functioning was assessed by teacher report ofacademic competence, peer competence, and internalizing and externalizing symptoms. Variable-focusedand person-focused results indicate that effortful control may be an important marker of school readinessand resilience. Controlling for child IQ, parenting quality, and socio-demographic risks, effortful controlemerged as the most significant predictor of all four salient developmental domains of adaptation as well asof resilient status of homeless children. Implications of these findings are discussed for future research anddesign of interventions.

l rights reserved.

© 2009 Elsevier Inc. All rights reserved.

Introduction

Twenty years of research on homeless children and families hasproduced a substantial body of literature showing that homelesschildren living in poverty are at a very high risk for delays in multipledomains of adaptive functioning, including academic, social, emo-tional, and behavioral problems (Buckner, 2008; Buckner, Bassuk,Weinreb, & Brooks, 1999; Haber & Toro, 2004; Masten, 1992; Masten,Miliotis, Graham-Bermann, Ramirez, & Neemann, 1993; Rafferty &Shinn, 1991). In addition to the cumulative risks associated withpoverty, such as traumatic life experiences, parental psychopathology,and lack of support systems (Luthar, 1999; McLoyd, Aikens, & Burton,2006), homeless children face specific threats to development fromresidential instability and broken bonds with potentially positivesources of security and opportunity (Rafferty, Shinn, & Weitzman,2004; Rog & Buckner, 2007). However, recent studies have demon-strated that homeless children are not a homogenous group(Huntington, Buckner, & Bassuk, 2008; Obradović et al., 2009).Evidence of positive adaptation in homeless children, even at veryhigh levels of risk, suggests that there is significant variability in thepromotive and protective factors influencing the lives of thesechildren. Understanding the processes that support successfuladaptation in children exposed to homelessness and its attendantrisks are important for preventive intervention. And yet, research on

positive processes that contribute to resilient adaption in this highlydisadvantaged group of children is rare. The current study wasdesigned to fill this gap by examining how effortful control abilitiesrelate to multiple domains of adaptive functioning in 5 and 6-year-oldhomeless children during their transition to school.

Adaptive functioning in homeless children

Homeless children fall at the high end of the risk continuum ofchildren living in poverty, showing significant developmental delaysacross the major domains of adaptation. Studies of academic successhave found that a large majority of homeless children show severeacademic achievement delays and perform below grade level norms(Masten et al., 1997; Rubin et al., 1996; Zima, Wells, & Freeman,1994). Masten et al. (1993) found that a significantly higherpercentage of homeless children have clinical levels of social problemswhen compared to national norms. Moreover, homeless childrenshow elevated levels of externalizing and internalizing behaviorproblems, often above clinical thresholds (Masten et al., 1997).

Further, many studies have shown that homeless children presenta unique risk group that is distinct from the larger group of childrenliving in poverty. When compared to economically disadvantagedchildren living at home, homeless children show lower levels ofliteracy and arithmetic skills (Rubin et al., 1996; Obradović et al.,2009), are more likely to report having no close friends (Masten et al.,1997), and are at higher risk for clinical levels of psychopathology(Masten et al., 1993; Rescorla, Parker, & Stolley, 1991; Vostanis,Grattan, Cumella & Winchester, 1997). These differences often remain

110 J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

after controlling for important confounds, such as race, social class,and family composition (Rubin et al., 1996).

Nevertheless, housing status has been an inconsistent predictor ofadaptation (Buckner, 2008). Homeless children tend to face higherlevels of adversity exposure, and housing status per se might not bethe most important marker of risk. For example, Buckner et al. (1999)found that housing status predicted internalizing symptoms but notexternalizing symptoms, controlling for the significant contribution ofage, ethnicity, maternal distress, stressful life events, and history ofabuse. However, homeless children had significantly higher levels ofstressful life events and history of sexual abuse. Similarly, Masten et al.(1993) reported that housing status did not predict overall symptomlevels over and above parental distress, demographic risk, andnegative life events, but homeless children on average were exposedto nearly twice as many negative life events. In a recent review ofstudies on homeless children, Buckner (2008) argued that incon-sistencies in the literature may be largely due to historical, contextual,and policy-related factors. For example, he suggests that recentstudies reveal fewer differences between homeless and low-income,housed children in part due to better social policies and servicescurrently in place for homeless families that did not exist in the 1980s.More importantly, Buckner (2008) advised that additional studiesexamining differences between homeless and housed children maynot meaningfully advance our understanding of what can improvehomeless children's adaption.

While homelessness presents a marker of high cumulative risk,there are hints of considerable variation in the adaptation of homelesschildren, with some children showing signs of positive development(Obradović et al., 2009). However, little is known about the protectivefactors that may account for resilient functioning among homelesschildren. In one of the few studies focused on protective factors,Masten and colleagues identified closeness of the parent–childrelationship and parental involvement in education as significantpredictors of homeless children’s school success (Masten & Sesma,1999; Milotis, Sesma & Masten, 1999). More research is needed toidentify other protective processes that contribute to the variability ofadaption in homeless children and to differentiate subgroups ofchildren who do well despite adversity.

As Buckner (2008) points out, the next generation of studies onhomeless children needs to have clear implications for policy andintervention program designs. One key focus of this research shouldbe early childhood, as developmental delays appear to emerge earlyand persist over time. In a recent study of achievement trajectories ina large urban school district, homeless and highly mobile studentsshowed lower initial levels of achievement than low-incomeclassmates who were more stably housed as early as second grade,controlling for sex, ethnicity, English language proficiency, andattendance (Obradović et al., 2009). Moreover, this gap persistedover three school years, underscoring the importance of examiningfactors that may prevent early school failures in homeless children.School entry may present a particularly vulnerable time for homelesschildren, as residential instability may jeopardize the consistency oftheir educational experiences and interfere with the formation of newpeer groups.

Effortful control and adaptive functioning

Scientists examining processes that promote successful develop-ment in high-risk children have emphasized the importance ofinvestigating systems that (a) bridge multiple levels of analysis,(b) are implicated in the salient domains of adaptation, and (c) areamenable to change (Luthar, 2006; Masten & Obradović, 2006). Onesystem that meets all three criteria is effortful control, a set ofexecutive functions aimed at the intentional, internal manipulation ofone's attention and behavior. Effortful control is frequently defined asthe ability to inhibit a dominant response in order to execute a

subdominant response (Rothbart & Bates, 1998). It includes skills suchas inhibitory control, attention shifting, and attention focusing.Together, these abilities play an important role in achieving thedevelopmental tasks of the school years: learning, forming friend-ships, and following the rules of classroom and society.

Effortful control has been linked to multiple domains of adaptivefunctioning. Recent studies show the significant influence of attentionregulation and inhibitory control on early academic competence, suchas reading and math achievement in preschoolers and kindergarten-ers (Blair & Razza, 2007; Howse, Lange, Farran, & Boyles, 2003; Senn,Espy, & Kaufmann, 2004). Effortful control has also been linked tosocial competence in community samples of 8 to 12-year-olds(Lengua, 2002, 2003) and preschool children (Lengua, Honorado, &Bush, 2007). Further, a robust relation has been documented betweeneffortful control skills and externalizing behavior problems inpreschoolers (Eisenberg et al., 1997; Kochanska & Knaack, 2003;Olson, Sameroff, Kerr, Lopez, & Wellman, 2005), and this relationpersists even when the high longitudinal stability and concurrentintercorrelations of both constructs are taken into account (Valienteet al., 2003; Eisenberg et al., 2004). Initial levels as well asimprovements in effortful control skills have been found to predictshort-term changes in externalizing and internalizing behaviorproblems (Eisenberg et al., 2005; Riggs, Blair, & Greenberg, 2003).Together, these studies suggest that in community samples effortfulcontrol can be a powerful correlate and predictor of adaptation acrossthe salient developmental domains of childhood. However, most ofthe research on effortful control and adaptive functioning has focusedon Caucasian, middle-class samples (Eisenberg, Hofer, & Vaughan,2007).

Development of effortful control is shaped by both biological andenvironmental factors (Eisenberg et al., 2003; NICHD Early Child CareResearch Network, 2005; Rothbart, Sheese, & Posner, 2007; Tarullo,Obradović & Gunnar, 2009). Recent studies suggest that risk andadversity exposure significantly undermine effortful control. Eco-nomically disadvantaged 5 to 7-year-olds performed significantlyworse on attention regulation and attention shifting tasks than theirmore affluent age-mates (Howse et al, 2003, Mezzacappa, 2004). In acommunity sample of preschoolers, indices of poverty, mobility, andfamily problems were related to lower performance on effortfulcontrol tasks, and a cumulative index of demographic and psychoso-cial risks negatively predicted 6-month change in effortful control(Lengua et al., 2007). Similarly, Li-Grining (2007) reported that socio-demographic and residential risks negatively predicted effortfulcontrol in low-income preschoolers.

Given that risk and adversity exposure undermine effortful controlskills, it is surprising that few studies have examined how effortfulcontrol contributes to adaptive functioning in at-risk children. In theHead Start sample, Blair and colleagues found some evidence ofrelation between performance on effortful control tasks and teacherreport of classroom behaviors (Blair, Granger, & Razza, 2005; Blair &Peters, 2003). In another study, effortful control tasks predicted mathand literacy skills in kindergarten (Blair & Razza, 2007). Althoughthese studies provide initial evidence that effortful control isimportant for the adaptation of economically disadvantaged children,the researchers did not control for family sources of risk andresources. Moreover, homeless children living in an emergencyshelter tend to face higher levels of adversity than low-incomehoused children, which may further undermine the associationbetween effortful control and adaptation.

One published study to date has examined self-regulation inchildren living in extreme poverty and exposed to high adversity,including homelessness. Buckner, Mezzacappa, and Beardslee (2003)showed that self-regulation differentiated groups of children whoshowed resilient and maladaptive adaptation and was also asignificant predictor of a continuous measure of resilience, controllingfor important confounds such as negative life events, chronic strains,

111J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

abuse history, IQ, self-esteem, parental monitoring, and emotionalsupport. While this study established the importance of studying self-regulation in highly disadvantaged children, the self-regulationconstruct combined ratings of effortful control skills with measuresof reactivity, motivation, and emotion regulation; thus, the distinctcontributions of specific elements were not evident. Moreover,children's ages varied considerably from 8 to 17 years, making itdifficult to assess the effects of effortful control in early childhood,especially during the transition to school.

Current study

The current study was designed to examine the variability offunctioning within the homeless children population (rather thancomparing it to low-income housed children) and to identifyprocesses that promote homeless children’s positive adaptation. Itaims to extend the findings of Buckner and colleagues (2003) bytesting the effect of effortful control skills, as measured by laboratorytasks, on various domains of adaptation during an importantdevelopmental transition period. At age 5 to 6, children showsignificant developmental improvements in effortful control, just asthey begin to face new challenges in the school environment. Asadvocated by Buckner (2008), this study incorporated both variable-focused and person-focused analytic approaches. The first goal was toexamine the separate relations between effortful control and foursalient developmental domains of adaption. The second goal was toexamine whether effortful control identifies a group of children whodemonstrate resilience across all four domains of adaptation, asindexed by average or better functioning. In accordance with existingliterature, effortful control was expected to relate to indices ofacademic achievement, peer competence, and internalizing andexternalizing symptoms, and to differentiate resilient from maladap-tive homeless children. All analyses were tested controlling for sexand age. In addition, the unique effects of effortful control wereexamined over and above child IQ, parenting quality, and familysocio-demographic risks, three key correlates of adaptation amongboth low- and high-risk samples.

Methods

Participants and procedures

Participants for this study were recruited from one of the largesthomeless shelters in the upper Midwest. Children in this shelterrepresent approximately 40% of all children living in homelessshelters who enroll at any point during the school year in the urbanpublic school district in which it is located. Families with children whowere scheduled to enter kindergarten or first grade in the fall of 2006were invited to participate in the study. Due to the nature of theassessments, families who did not speak English were not eligible forthe study, and five families were excluded on this basis. As the studyprogressed, it became evident that the current census at the shelterhad more eligible boys and more first graders; thus, if a family had twoor more eligible children, females and kindergartners were recruited.

The sample consisted of 58 children (20 females, 38 male). The meanage of the children was 6.09 (SD = 0.54 range 4.97–7.23). The sampleincluded 81.0% African-American, 1.7% American Indian, 3.4% Anglo-American, and 13.8% mixed-ethnicity children. The ethnic compositionof this sample is very similar to previous studies of homeless childrenliving in the same city (Masten et al., 1993, 1997). About 75% of childrenlived with a single caregiver; 90% of primary caregivers were mothers,7% were fathers, and 3% were grandmothers. For simplicity, from thispoint forward primary caregivers are referred to as parents.

Although an effort was made to schedule sessions at least two daysafter the family's arrival at the shelter, in order to allow them toacclimate to the new environment, one family participated in the

study before their second night at the shelter. The duration ofhomelessness is difficult to estimate, as many families live “doubledup” with their relatives or friends prior to arriving at the shelter, whileothers bounce between different emergency shelters. At the time ofparticipation, the parent-reported length of families' current shelterstay ranged from 1 to 365 days (Mdn = 9 days; M = 26; SD = 64).Forty-three percent of parents reported being without their ownhousing (rented or owned by the parent or a partner) for less than amonth, 22% for 1–3 months, 19% for 3 months to 1 year, 14% for morethan a year, and one family never had their own housing. Moreover,46.6% of parents reported being homeless at other times in their lives.Due to the nature of homelessness, it was not feasible to determinewhen these families ended being homeless.

Children and parents separately participated in a 90-minutesession conducted on shelter premises. Children completed a seriesof standardized intelligence subtests and a battery of effortful controltasks, while parents were interviewed and administered variousquestionnaires. In the fall of 2006, 22 of the children enteredkindergarten and 36 entered the first grade. Teachers of the 54children who attended area schools were invited to complete aquestionnaire about the child's adaptive functioning, and 100% of theteachers returned the questionnaire, representing 93% of the originalsample. On average, teacher completed their assessments approxi-mately 4 months after the testing session was conducted at the shelter(range: 20–229 days; M = 115, SD = 52).

It is difficult to ascertain participation rates accurately for studiesin emergency shelters, because the families continuously move in andout. At the outset, some families move out before they learn of thestudy and others move before they can be scheduled to participate.Based on participation rates during the middle week of each of thethree full months when assessments were conducted, approximately90% of families with eligible children participated in the study,suggesting a highly representative sample of families residing in thisemergency shelter.

Measures

Effortful controlChildren's effortful control skills were assessed using a battery of

four standard laboratory tasks that were selected based on theirdevelopmental appropriateness and ability to tap different aspects ofeffortful control (Carlson, 2005). In the Simon Says task (Strommen,1973; Kochanska, Murray, & Coy, 1997) children were instructed toimitate the experimenter's behavior during commands prefaced bythe words “Simon says”, while inhibiting imitations when commandswere not prefaced with “Simon says”. After practice trials, childrenwere administered a series of 20 counterbalanced trials (10 activation“Simon says” trials and 10 inhibition trials). Two trained coders, blindto children's performance on all other effortful control tasks as well asall other family or child data, coded the degree of transgression ormovement during the 10 inhibition trials. Interrater reliability forinhibitory trials was calculated based on 25% of the sample (weightedκ = .94). There was no variability in activation scores as all childrenperformed well on activation trials. In the Peg-tapping task (Diamond& Taylor, 1996), children were instructed to tap twice with a woodendowel when the experimenter tapped once and to tap once when theexperimenter tapped twice. After practice trials, the children werepresented with 16 counterbalanced test trials. Performance on thePeg-tapping task was measured by the number of correct taps on ascale from 0 to 16 which were recorded during the task. In thecomputerized pointing Stroop task (Berger, Jones, Rothbart, & Posner,2000), children were trained first to point to a picture matching ananimal sound (e.g., congruent trials: point to the picture of a cat afterhearing ‘meow’), and then to point to a picture not matching theanimal sound (e.g., incongruent trials: point to the dog after hearing‘meow’). After practice trials, children were presented with a series of

112 J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

16 congruent trials followed by a series of 16 incongruent trials.Performance on incongruent Stroop trials was recorded by computersoftware on a 0 to 16 scale. Two children performed at or belowchance on congruent trials, indicating that they did not understandthe task, and their incongruent trials were not analyzed. During theDimensional Change Card Sort task (DCCS; Zelazo, 2006), childrenwere taught first to play the “color game”, placing all red cards in a boxwith a red rabbit and all blue cards in a box with a blue boat. After theylearned this rule, they were asked to play the “shape game”, placing allcards with blue or red rabbits in the box with the red rabbit and allcards with blue or red boats in the box with the blue boat, whichrequired them to inhibit the previously established pattern of sortingby color. DCCS consisted of 6 “color game” trials before the rule switchand 6 “shape game” trials after the rule switch. Performance wasscored during the task on a scale from 0 to 6, reflecting the number ofcorrectly sorted cards during the post-switch, “shape game” trials.

The number of correct trials on the Simon Says task, Peg-tappingtask, and Stroop task were transformed into percentage scores toaccount for any missing data due to procedural irregularities (e.g., thechild was off camera during one Simon Says trial). A composite indexof effortful control skills (EC) was calculated by averaging standard-ized scores on the four effortful control tasks. The four effortful controltasks scores satisfied Patterson and Bank's (1986) criteria by showinghigh internal consistency as a scale (α = .75), high item-totalcorrelations (r N .70, p b .001), and loading on a single factor with allloadings over .70.

Adaptive functioningFour different domains of children's adaptive functioning were

assessed using the teacher form of The MacArthur Health BehavioralQuestionnaire (HBQ; Armstrong, Goldstein, & The MacArthur WorkingGroup on Outcome Assessment, 2003). All the subscale and scalereliability statistics are presented for this sample. Academic Compe-tence was assessed using the HBQ Academic Functioning Scale,which is an average of two subscales (r = .33, p b .05): (1) theeight-item School Engagement subscale (α = .87) and (2) the five-item Academic Competence subscale (α = .96). Peer Competence wasassessed using the HBQ Peer Relations Scale, which is an average of twosubscales (r = .63, p b .001): (1) the eight-item Peer Acceptance/Rejection subscale (α = .94) and (2) the three-item Bullied by Peerssubscale (α = .80). Internalizing Behavioral Problems was assessedusing the HBQ Internalizing Symptoms Scale, which is an average oftwo subscales (r = .70, p b .001): (1) the six-item Depression subscale(α = .83) and (2) the eight-item Overanxious subscale (α = .69).Externalizing Behavioral Problems was assessed using the HBQExternalizing Symptoms Scale, which is an average of four subscales(α = .89): (1) the nine-item Oppositional Defiant subscale (α = .92),(2) the eleven-item Conduct Problem subscale (α = .85), (3) the four-item Overt Hostility subscale (α = .80), and (4) the six-item RelationalAggression subscale (α = .83).

Homeless children were classified as resilient if they scored all of thefollowing: (1) 3 or higher on a 5-point academic functioning scale,which corresponds to average academic competence and some schoolengagement, (2) 3 or higher on a 4-point peer competence scale, whichcorresponds to somewhat likely to be accepted and not very likely to berejected and bullied by peers, (3) lower than the clinical threshold forthe depression subscale (b .86) and anxiety subscale (b .91), and (4)lower than the clinical threshold for the oppositional defiant subscale (b.84) and conduct problem subscale (b .79). Clinical thresholds weredetermined by Luby and colleagues (2002), who evaluated perfor-mance of parent-reported HBQ symptoms in identifying DSM-IVinternalizing and externalizing symptoms in referred and non-referred4 to 8-year-olds. Clinical thresholds derived by Luby et al. (2002) wereemployed over the more recent Lemery-Chalfant et al. (2007) thresh-olds because the latter study focused on children older than the currentsample and did not provide subscale level data for internalizing

symptoms. However, classifications based on the two set of thresholdsoverlap considerably (93% for externalizing symptoms).

Risk and resourcesA Cumulative Risk Index was created by summing five well-

established socio-demographic risk factors (M = 1.94; SD = 1.10):single parent household (74%), maternal age at first birth younger than18 years old (37%), 3 or more siblings living with the family (33%),primary caregiver's education less than a high school degree (26%), andno current income (27%). Continuous risk factors (i.e., maternal age)were dichotomized according to meaningful cutpoint standards com-monly applied in risk studies (Sameroff, Seifer, Zax, & Barocas, 1987).

Children's General Intellectual Functioning (IQ) was measured usingthe Block Design, Matrix Reasoning, and Vocabulary subtests of theWechsler Preschool and the Primary Scale of Intelligence-Third Edition(WPPSI-III; Wechsler, 2002). The raw scores on each subtest weretransformed to age adjusted scaled scores. According to WPPSI-IIIstandardization norms, each subset produces a scaled score with a meanof 10 and a standard deviation of 3. Scaled scores on the Block Designand the Matrix Reasoning subtests were first composited to create anon-verbal measure of IQ (r = .38, p b .01), which was then compositedwith the Vocabulary subscale scaled score (r = .49, p b .001) to yield ameasure of IQ.

Parenting Quality (PQ) was scored by research assistants whocompleted five behavioral ratings immediately following a 90-minutestructured interview with a parent, assessing how positively andwarmly the parent spoke of the child, how hostile or rejecting theparent was in describing the child, closeness of parent and child, andthe overall quality of the parent–child relationship (α = .89). Thesebehavioral ratings were based on overall impressions of the parent'sbehavior and tone during the interview as well as specific reports ofparenting practices, joint activities, displays of affections, schoolinvolvement, and many aspects of the child's behavior. In addition, theparent interviewer observed parent and child interactions during theshort consent procedure. Parent interviewers were blind to children'sperformance on EC and IQ tests. Analogous rating scales showed highinternal consistency, interrater agreement, and predictive validity in aprevious study of homeless children (Milotis et al., 1999).

Data preparation procedures

Missing data at the level of subscale scores (i.e., HBQ, IQ) andindividual indices (i.e., effortful control tasks, PQ, risk) were handledusing the recommended maximum likelihood estimation procedure formissing data, specifically the expectation-maximization (EM) algorithm(Schafer & Graham, 2002). Percentage of missing data was as follows:teacher report (6.9%) except for academic competence and depressionsubscales (8.6%), PQ (3.4%), Simon Says task (13.8%), Peg-tapping task(1.7%), and Stroop task (6.9%). All other variables had complete data.

Data was examined for significant outliers and deviations from anormal distribution. Only one variable value fell beyond 3 SD from themean (i.e., one child had a standardized externalizing symptoms scoreof 3.27). Since review of the scatter plots revealed that no outlierswere artificially inflating the relations between predictor andoutcome variables, all data were included in the analyses. Lastly,evaluation of skewness, kurtosis, and normal p–p plots revealed nogross deviation from normal distribution of outcome variables.

Results

Descriptive statistics

Table 1 presents range, mean, and standard deviation for all variablesin the study. Effortful control tasks captured considerable variability ofeffortful control skills in homeless children. Only 26% of children passedthe Simon Says task by correctly performing at least 80% of the

Table 1Descriptive statistics.

Min Max M SD

Simon says task 0.00 100.00 42.99 36.50Peg-tapping task 6.25 100.00 80.29 24.53Stroop task 25.00 100.00 83.74 16.75DCCS task 0.00 6.00 4.62 2.46IQ–BD subtest 2.00 12.00 5.97 2.38IQ–MR subtest 3.00 16.00 7.86 2.24IQ–V subtest 3.00 14.00 7.83 1.96Parenting quality 2.00 5.00 4.04 0.84Cumulative risk index 0.00 5.00 1.95 1.10Academic functioning 1.63 4.60 3.20 0.66

School engagement 2.25 5.00 4.16 0.74Academic competence 1.00 4.20 2.24 0.88

Peer competence 1.38 4.00 3.28 0.63Peer acceptance/rejection 1.13 4.00 3.07 0.77Bullied by peers (reversed) 1.00 4.00 3.49 0.63

Internalizing symptoms 0.00 1.49 0.52 0.37Depression 0.00 1.83 0.50 0.45Anxiety 0.00 1.25 0.53 0.35

Externalizing symptoms 0.00 1.49 0.35 0.35Oppositional defiant 0.00 1.78 0.44 0.47Relational aggression 0.00 1.17 0.29 0.33Conduct problems 0.00 1.50 0.30 0.32Overt hostility 0.00 2.00 0.39 0.46

Notes. N = 58; BD: Block design; DCCS: Dimensional Change Card Sort; MR: MatrixReasoning; V: Vocabulary.

113J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

inhibitory trials. On other hand, 72% and 67% of children correctlycompleted at least 80% of trials during the Peg-tapping and Stroop tasks,respectively. In addition, 76% of children passed the DCCS task by sortingat least 5 out of 6 post-switch cards correctly. Children showedsignificant delays on the IQ test, with average performance on allthree subsets well below WPPSI-III scale score norms. On average,children showed higher levels of school engagement than academiccompetence. A large majority of students (78%) were rated as havingunsatisfactory academic competence (b 3.0), and 36% of students wererated as having poor academic competence, well below grade level(b 2.0). On average, children showed higher levels of internalizingsymptoms than externalizing symptoms.

Variable-focused analyses

Bivariate correlationsTable 2 presents the correlations among all variables included in

the study. Children's performance on effortful control tasks was

Table 2Bivariate correlations.

1 – Age 2 – Sex 3 – SS 4 – PEG 5 – Stroop 6 – DCCS

l Age –2 Sex .05 –3 SS .54⁎⁎⁎ .28⁎ –4 PEG .39⁎⁎ .21 .46⁎⁎⁎ –5 Stroop .34⁎⁎ .22 .52⁎⁎⁎ .38⁎⁎ –6 DCCS .42⁎⁎⁎ .14 .37⁎⁎ .39⁎⁎ .47⁎⁎⁎ –7 EC .56⁎⁎⁎ .28⁎ .78⁎⁎⁎ .74⁎⁎ .78⁎⁎⁎ .74⁎⁎⁎8 IQ .40⁎⁎ .37⁎⁎ .32⁎ .33⁎ .51⁎⁎⁎ .55⁎⁎⁎9 PQ .26⁎ .19 .42⁎⁎ .33⁎ .38⁎⁎ .27⁎10 Risk –.26† –.07 –.18 –.23† .29⁎ –.1811 ACA .08 .00 .24† .26⁎ .40⁎⁎ .40⁎⁎12 PEER .02 .04 .11 .26⁎ .34⁎⁎ .25†

13 LNT .16 .05 –.09 –.08 –.16 –.1514 EXT –.06 .02 –.28⁎ –.37⁎⁎ –.46⁎⁎⁎ –.34⁎⁎

Notes. N = 58; ACA: academic competence; DCCS: Dimensional Change Card Sort; EC: effcompetence; PEG: Peg-tapping task; PQ: Parenting quality; Risk: cumulative risk index SS:

†p b .10.⁎p b .05.⁎⁎p b .01.

⁎⁎⁎p b .001.

significantly correlated with age and sex. Older children and girlsdemonstrated higher levels of effortful control skills. The effortfulcontrol composite (EC) was significantly related to three domains ofadaptation: academic competence, peer competence, and externaliz-ing behavior problems. Children with higher levels of EC showedhigher levels of competence and lower levels of symptoms. Inaddition, EC was positively related to IQ and PQ, and negativelyrelated to socio-demographic risk. Resources and risk were alsorelated to adaptation domains in expected directions.

Linear regression analysesFour separate hierarchical regression analyses were conducted to

examine the association between EC and adaptive functioning (seeTable 3). Age and sex were controlled in step 1. The effect of EC onadaptation domains was examined in step 2. Finally, the significanceof EC effect over and above IQ, PQ, and the cumulative risk index wastested in step 3 by introducing the controls variables into the model.Controlling for age and sex, EC significantly predicted teacher reportof all four domains of adaptive functioning. EC uniquely explained 24%of variance in academic competence, 14% of variance in peercompetence, 11% of variance in internalizing symptoms, and 33% ofvariance in externalizing symptoms. Moreover, EC remained asignificant predictor of all four domains after controlling for thecontribution of IQ, PQ, and risk. IQ also significantly predictedacademic competence and internalizing symptoms and marginallypredicted peer competence and externalizing symptoms. Childrenwith lower IQ demonstrated lower competence and higher levels ofsymptoms. PQ was marginally related to internalizing symptoms,with higher parenting quality related to more symptoms. Risk did notpredict adaptation over and above other variables in the model. In thefinal step, IQ, PQ, and risk explained an additional 14% of variance inacademic competence, 7% of variance in peer competence, 10% ofvariance in internalizing symptoms, and 7% of variance in external-izing symptoms.

Person-focused analyses

Resilient group compared to maladaptive groupBased on the thresholds described in the methods section, 36% of

children showed maladaptive levels of academic functioning, whereas29% showed maladaptive levels of peer competence. In terms ofpsychopathology, 19% and 24% of children showed clinical levels ofinternalizing and externalizing symptoms, respectively, with 9%having comorbid symptoms. Specifically, 16% and 14% of children

7 – EC 8 – IQ 9 – PQ 10 – Risk 11 – ACA 12 – PEER 13 – INT

–

.56⁎⁎⁎ –

.46⁎⁎⁎ .40⁎⁎ –

.29⁎ .31⁎ –.23† –

.43⁎⁎⁎ .42⁎⁎⁎ .40⁎⁎ –.29⁎ –

.32⁎ .30⁎ .16 .06 .32⁎ ––.16 –.21 .09 .01 .29⁎ –.19 ––.48⁎⁎⁎ –.36⁎⁎ –.18 .26⁎ –.47⁎⁎⁎ –.44⁎⁎⁎ .47⁎⁎⁎

ortful control; EXT: externalizing behaviors; INT: internalizing behaviors; PEER: peerSimon Says task.

Table 3Linear regressions: predictors of adaptive functioning in homeless children.

Academic Peer Internalizing Externalizing

β ΔR² β ΔR² β ΔR² β ΔR²

Step 1 .01 .00 .03 .00Age .08 .02 .16 –.06Sex .00 .04 .04 .02

Step 2 .24⁎⁎⁎ .14⁎⁎ .11⁎ .33⁎⁎⁎Age –.26† –.24 .39⁎ .34⁎Sex –.16 –.08 .15 .21†

EC .62⁎⁎⁎ .47⁎⁎ –.41⁎ –.73⁎⁎⁎Step 3 .14⁎ .07 .10† .07

Age –.34⁎ –.26† .44⁎⁎ .40⁎⁎Sex –.26⁎ –.15 .22 .27⁎EC .38⁎ .40⁎ –.38⁎ –.63⁎⁎⁎IQ .32⁎ .29† –.36⁎ –.26†

PQ .20 .01 .24† .09Risk –.13 .19 –.03 .14

Adjusted R² .32⁎⁎⁎ .12⁎ .15⁎ .33⁎⁎⁎Total R² .39⁎⁎⁎ .21⁎ .24⁎ .40⁎⁎⁎

Notes. N = 58; ACA: academic competence; EC: effortful control; EXT: externalizingbehaviors; INT: internalizing behaviors; PEER: peer competence; PQ: parenting quality;Risk: cumulative risk index.

†p b .10.⁎p b .05.⁎⁎p b .01.

⁎⁎⁎p b .001.

Table 5Logistic regression: predictors of resilience in homeless children.

Exp(B) 95% CI

Age .12⁎ (.02, .65)Sex 1.85 (.42, 8.24)EC 5.40⁎⁎ (1.68, 17.35)IQ 1.85 (.74, 4.58)PQ .90 (.43, 1.87)Risk 1.02 (.52, 2.00)

Notes. N = 58; EC: effortful control; PQ: parenting quality; Risk: cumulative risk index.⁎p b .05.

⁎⁎p b .01.

114 J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

showed clinical levels of depression and anxiety symptoms, respec-tively, whereas 10% and 22% of children showed clinical levels ofconduct problems and oppositional defiant disorder, respectively. Insum, 34 children (59%) showed maladaptive adaptation in at least onedomain, with 36% showing maladaptation in more than one domain.In contrast, 24 children (41%) demonstrated average levels ofacademic functioning and peer competence and had no clinical levelsof psychopathology, and thus were identified as resilient. Table 4presents means and standard deviations for resilient and non-resilientchildren across four domains that were used as defining criteria, aswell as for EC, IQ, PQ and risk variables. As expected, resilient childrenshowed higher levels of competence and lower levels of psychopa-thology. Although the two groups did not significantly differ in socio-demographic risk or PQ, resilient children also showed higher levels ofIQ and EC.

Logistic regression analysisLogistic multiple regression analysis was conducted to identify

factors that distinguish resilient homeless children from non-resilienthomeless children as defined by the criteria described in the methodssection. The same variables that were used in linear regression analyses

Table 4Differences between resilient and non-resilient homeless children.

Resilient Non-resilient

N = 24 N = 34

M SD M SD F-test

ACA 3.68 0.43 2.86 0.58 34.84⁎⁎⁎PEER 3.69 0.21 2.99 0.67 24.71⁎⁎⁎INT 0.30 0.18 0.67 0.39 18.27⁎⁎⁎EXT 0.12 0.14 0.52 0.36 26.74⁎⁎⁎EC 0.46 0.60 –0.33 1.10 10.06⁎⁎IQ 7.93 1.45 6.98 1.72 4.81⁎PQ 4.24 0.77 3.90 0.87 2.37Risk 1.83 0.96 2.03 1.19 0.44

Notes. N = 58; ACA: academic competence; EC: effortful control; EXT: externalizingbehaviors; INT: internalizing behaviors; PEER: peer competence; PQ: parenting quality;Risk: cumulative risk index.

⁎p b .05.⁎⁎p b .01.

⁎⁎⁎p b .001.

to predict continuous indices of adaptation domains were used inlogistic regression to predict categorical measure of resilience. Oddsratios and confidence intervals are listed in Table 5. Age and EC emergedas the only significant predictors of resilience. With each increase of 1 SDin EC skills in this sample of children who were homeless, children were5.4 times more likely to be identified as resilient.

Discussion

This study identifies effortful control skills as an importantpotential indicator of adaptation and school readiness among childrenwho are homeless. Performance on effortful control tasks wassignificantly related to school adjustment across four of the mostsalient childhood domains of adaptation, as judged by the teacher, anindependent observer of the child's functioning. Children's effortfulcontrol emerged as a unique predictor of early academic success,which can have long lasting effects on future adaptation (Mastenet al., 2005). Interestingly, even though parenting quality wassignificantly related to academic competence, it did not emerge as asignificant predictor children's academic functioning once the effect ofeffortful control was accounted for. Moreover, effortful control in thissample of homeless children showed a unique relation to academicachievement that was independent of general intellectual abilities,despite the significant correlation between effortful control and IQ.

Effortful control also predicted peer competence, internalizingsymptoms and externalizing symptoms independent of intelligence.In a school setting, the ability to focus attention on relevant stimuliand inhibit unconstructive behaviors, together with problem-solvingand verbal skills, may be particularly relevant for homeless children'searly experience with peers and their rule-abiding conduct. Homelesschildren with lower effortful control and lower cognitive skills seemto be at greater risk for social, emotional, and behavior problemswithin the school context than homeless children who demonstratehigher levels of effortful control and intelligence.

Children's effortful control performance explained considerablevariability in adaptive functioning. Although these effects areconsistent with the handful of studies conducted with economicallydisadvantaged groups (Blair et al., 2005; Howse et al., 2003), themagnitude of these effects is larger than previously found and isevident across academic, social, and behavioral indices of adaptationin the same population. Person-focused analysis also revealed that thelevel of effortful control skills differentiated children who wereresilient across all four domains by demonstrating adequate compe-tence and lack of considerable psychopathology from those whoshowed major delays or problems in at least one domain offunctioning. Effortful control emerged as the only significant predictorof resilient status controlling for the two most established protectivefactors in the resilience literature: IQ and parenting quality (Masten &Obradović, 2006). Thus, it appears that effortful control plays animportant role in the adaptation of homeless children, whose chaotic,stressful, and unstable lives place them at greater risk for develop-mental delays than other economically disadvantaged children.Effortful control may represent a good marker of how homeless

115J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

children are doing in terms of overall developmental expectationsthat society holds for children entering primary school and mayprovide an ecologically valid indication of children's capacity forfunctioning in the structured school setting.

Even though homeless children occupy the high end of the socio-demographic risk continuum, the cumulative risk index was nega-tively related to effortful control skills. Moreover, homeless childrenin this study demonstrated lower levels of effortful control thanchildren from predominantly middle-class samples, performingworse than age-mates, or at the level of younger children, on tasksfor which comparison could be made with other studies. Thepercentage of 5 to 6-year-old homeless children who passed DCCSin this study was equivalent to the percentage of older 4-year-oldsfrom a primarily middle-class sample that passed the same task in adifferent study (Carlson, 2005). Similarly, while less than a third of 5-to 6-year-old homeless children in this study passed the Simon Saystask, nearly half of middle-class 5-year-olds passed the same task in adifferent study (Carlson, 2005). Thus, it will be important to furtherexamine which aspects of homeless children's lives tend to under-mine development of effortful control skills.

Similarly, future studies should identify specific processes in thelives of homeless children that may promote development of effortfulcontrol, a process influenced by various environmental factors innormative populations. One such factor may be parenting quality. Thecurrent study indicates that parenting quality was related to all foureffortful control tasks. This finding is consistent with a recent studyshowing that higher levels of parental limit setting and scaffoldingpredicted an increase in effortful control in a community sample ofpreschoolers and mediated the effect of cumulative risk on effortfulcontrol (Lengua et al., 2007). However, another recent study foundthat mother–child connectedness did not predict effortful controlover and above indices of risk in low-income preschoolers (Li-Grining,2007), and more research is needed to identify which particularaspects of parenting may influence the development of effortfulcontrol in at-risk children.

Finally, this study demonstrated that effortful control skills can bemeasured in a highly disadvantaged population of homeless childrenin an emergency shelter using a battery of standard behavioral tasksdeveloped primarily with mainstream, middle-class samples. Alleffortful control tasks were positively associated with age, supportingthe notion that effortful control abilities show age-related improve-ments in 5 to 6-year-olds (Carlson, 2005; Greenberg, Riggs, & Blair,2006). Moreover, effortful control tasks were positively related tostandardized measures of children's IQ, providing additional validitygiven that effortful control skills are also important for goodperformance on the IQ test. Valid measurement of effortful controlin homeless children is an important prerequisite for future studies ofeffortful control training and interventions.

Implications

Identifying protective processes that contribute to variability ofadaptation in homeless children is a crucial step toward designingsuccessful prevention and intervention programs (Buckner, 2008).This study indicates that laboratory measures of children's effortfulcontrol skills can be used as early markers of homeless children'sadaptive functioning during the transition to school. Since effortfulcontrol has also been found to promote academic gains (Howse et al.,2003), it may be an important predictor of future academictrajectories, particularly for homeless children who show very earlysubstantial achievement delays (Obradović et al., 2009; Rubin et al.,1996; Zima et al., 1994).

Moreover, the current findings have important implications forfostering resilience in homeless children, as recent studies demon-strate considerable plasticity of effortful control in response to variousinterventions. Effortful control skills in community samples have been

improved with training programs and specialized classroom curricula.Children enrolled in a preschool program with curriculum specificallydesigned to improve cognitive control performed significantly betteron effortful control tasks than children exposed to a literacy-basedcurriculum (Diamond, Barnett, Thomas, & Munro, 2007). In addition,participation in a training program designed to enhance effortfulcontrol skills in 4 and 6-year-olds was associated with improvedperformance on intelligence tests and more mature brain activationduring subsequent performance on an effortful control task (Rueda,Rothbart, McCandliss, Saccomanno, & Posner, 2005). Further, theeffect of a preventive intervention known as Promoting AlternativeThinking Strategies on levels of externalizing and internalizingsymptoms at a 1-year follow-up has been linked to improvedinhibitory control (Riggs, Greenberg, Kusche, & Pentz, 2006). Theconsequences of such training interventions could be far-reaching,due to the significant effects that effortful control appears to have onearly cognitive and socio-emotional development. Homeless children,who are at greater risk for developmental delays and problems thanother economically disadvantaged children, may particularly benefitfrom programs designed to foster better effortful control, as low levelsof effortful control skills have been shown to undermine theirsuccessful achievement of core age-salient developmental tasks.

Limitations

Many challenges surround the study of homeless children living inemergency shelter. First, this study is based on a modest sample of 58children. The lack of power due to the small sample size needs to beconsidered when interpreting the results. Nonetheless, the large effectsizes underscore the possible significance of the results, whilemarginal findings may be informative to future, larger studies.Second, while about 90% of eligible children from the largest familyhomeless shelter in the area participated in this study, it is difficult toknow how representative this sample is of the larger homelesspopulation, particularly in rural areas. Third, the assessments wereconducted in homeless shelter facilities, which were adequatelytransformed to meet the needs of assessment procedures (i.e.,children were tested in a library at the end of a quiet hallway).However, the level of noise and disruption outside the assessmentspace was slightly higher than in a regular research lab. On thenegative side this set-up may have contributed to eliciting lowerlevels of effortful control, but on the positive side it also probablyincreased the ecological validity of the measure. Fourth, although thestrength of this study incorporates many assessment methods andinformants, the measure of parenting quality is based solely on theinterviewer's clinical judgment of the parent's relationship with thechild based on information provided by the parent and observationsof their interaction incidental to the consenting process. Futurestudies of homeless children should incorporate systematic observa-tional measures of parent–child interactions. Fifth, clinical thresholdsare only approximations of clinical symptom level, as they are basedon studies using parent report. Sixth, the measures of effortful controlmay reflect concurrent instability and stress rather than children'sabsolute level of skills. While this distinction may not be relevant forchildren exposed to chronic adversity, future longitudinal studiesneed to examine how effortful control waxes and wanes withexposure to adversity. Finally, the correlational nature of this studyprevents researchers from drawing causal conclusions regardingassociations between effortful control and adaptation. It is quitepossible that these children had developmental problems before theybegan school, along with lower effortful control. However, given thesignificance of effortful control skills for salient domains of adaptationas well as its plasticity in response to simple training procedures,effortful control may present a better target for intervention effortsthan specific competence domains.

116 J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

Conclusions

The current study showed that effortful control skills may promoteresilience processes in homeless children during their transition toschool, as performance on effortful control tasks predicted achievementof early childhood salient developmental tasks over and above otherimportant riskand protective factors. Withrecentstudies demonstratingplasticity of effortful control skills in community samples, it is imperativeto investigate the malleability of these skills in homeless children. Thetransition to the school environment provides an important window ofopportunity for preventing early achievement delays and behavioralproblems in homeless children. It also represents a time when effortfulcontrol skills develop rapidly due to biological maturational processes.By encouraging healthy development of effortful control, programs mayhelp homeless children engage in school, learn basic reading and mathskills, form friendships, and inhibit aggressive behavior. The conse-quences of such training could be substantial, but more research will beneeded, including randomized experiments with longitudinal follow-ups, in order to rigorously test the possibilities of strategic interventionto promote success in school through the improvement of effortfulcontrol skills.

Acknowledgments

This study is based on a doctoral dissertation and was supported by aNational Institute of Mental Health NRSA pre-doctoral training grantaward to the author and by research awards to Ann S. Masten from theCenter on Urban and Regional Affairs at the Hubert H. HumphreyInstitute and a Distinguished McKnight University Professorship, both atthe University of Minnesota. I want to express my deep appreciation tothe children and their families for participating in this project. I wouldlike to acknowledge the extraordinary support of the staff of PeopleServing People shelter and Minneapolis Public Schools. I would like tothank the amazing graduate and undergraduate students at theUniversity of Minnesota who contributed to this study, especiallyJanette Herbers, J. J. Cutuli, Kristin Wiik, Theresa Lafavor, and DanielleVrieze. I am also grateful to Megan Gunnar, Bonny Donzela, and AmandaTarullo for their helpful advice. I especially want to acknowledge AnnMasten for her support and guidance during this project. Lastly, I wanttothank Josh Wallaert for all his support and encouragement.

References

Armstrong, J. M., & Goldstein, L. H. The MacArthur Working Group on OutcomeAssessment. (2003). Manual for the MacArthur Health and Behavior Questionnaire(HBQ 1.0). MacArthur Foundation Research Network on Psychopathology andDevelopment (David J. Kupfer, Chair), University of Pittsburgh.

Berger, A., Jones, L., Rothbart, M. K., & Posner, M. I. (2000). Computerized games tostudy the development of attention in childhood. Behavioral Research Methods andInstrumentation, 32, 297–303.

Blair, C., Granger, D., & Razza, R. P. (2005). Cortisol reactivity is positively related toexecutive function in preschool children attending head start. Child Development,76, 554–567.

Blair, C., & Peters, R. (2003). Physiological and neurocognitive correlates of adaptivebehavior in preschool among children in Head Start. Developmental Neuropsychology,24, 479–497.

Blair, C., & Razza, R. P. (2007). Relating effortful control, executive functioning, and falsebelief understanding to emerging math and literacy ability in kindergarten. ChildDevelopment, 78, 647–663.

Buckner, J. C. (2008). Understanding the impact of homeless on children: Challengesand future research directions. American Behavioral Scientist, 51, 721–736.

Buckner, J. C., Bassuk, E. L., Weinreb, L. F., & Brooks, M. G. (1999). Homelessness and itsrelation to mental health and behavior of low-income school-age children. Develop-mental Psychology, 35, 246–257.

Buckner, J. C., Mezzacappa, E., & Beardslee, W. R. (2003). Characteristics of resilientyouths living in poverty: The role of self-regulatory processes. Development andPsychopathology, 15, 139–162.

Carlson, S. M. (2005). Developmentally sensitive measures of executive function inpreschool children. Developmental Neuropsychology, 28, 595–616.

Diamond, A., Barnett, W. S., Thomas, J., & Munro, S. (2007). Preschool program improvescognitive control. Science, 318(5855), 1387–1388.

Diamond, A., & Taylor, C. (1996). Development of an aspect of executive control:Development of the abilities to remember what I said and to “do as I say, not as I do”.Developmental Psychobiology, 29, 315–334.

Eisenberg, N., Fabes, R. A., Shepard, S. A., Murphy, B. C., Guthrie, I. K., Jones, S., et al.(1997). Contemporaneous and longitudinal prediction of children's socialfunctioning from regulation and emotionality. Child Development, 68, 642–664.

Eisenberg, N., Hofer, C., & Vaughan, J. (2007). Effortful control and its socioemotionalconsequences. In J. J. Gross (Ed.), Handbook of emotion regulation (pp. 287–306).New York: Guilford.

Eisenberg, N., Sadovsky, A., Spinrad, T. L., Fabes, R. A., Losoya, S. H., Valiente, C., et al. (2005).The relations of problem behavior status to children's negative emotionality, effortfulcontrol, and impulsivity: Concurrent relations and prediction of change. Develop-mental Psychology, 41, 193–211.

Eisenberg, N., Spinrad, T. L., Fabes, R. A., Reiser, M., Cumberland, A., Shepard, S. A., et al.(2004). The relations of effortful control and impulsivity to children's resiliency andadjustment. Child Development, 75, 25–46.

Eisenberg, N., Zhou, Q., Losoya, S. H., Fabes, R. A., Shepard, S. A., Murphy, B. C., et al.(2003). The relations of parenting, effortful control, and ego control to children'semotional expressivity. Child Development, 74, 875–895.

Greenberg, M. T., Riggs, N. R., & Blair, C. (2006). The role of preventive interventions inenhancing neurocognitive functioning and promoting competence in adolescence.In D. Romer, & E. Walker (Eds.), Adolescent psychopathology and the developingbrain: Integrating brain and prevention science. New York: Oxford University Press.

Haber, M. G., & Toro, P. A. (2004). Homelessness among families, children, andadolescents: An ecological–developmental perspective. Clinical Child and FamilyPsychology Review, 7, 123–164.

Howse, R. B., Lange, G., Farran, D. C., & Boyles, C. D. (2003). Motivational and self-regulation as predictors of achievement in economically disadvantaged youngchildren. Journal of Experimental Education, 71, 151–174.

Huntington, N., Buckner, J. C., & Bassuk, E. L. (2008). Adaptation in homeless children:An empirical examination using cluster analyses. American Behavioral Scientist, 51,737–755.

Kochanska, G., & Knaack, A. (2003). Effortful control as a personality characteristic ofyoung children: Antecedents, correlates, and consequences. Journal of Personality,71, 1087–1112.

Kochanska, G., Murray, K. T., & Coy, K. C. (1997). Inhibitory control as a contributor toconscience in childhood: From toddler to early school age. Child Development, 68,263–277.

Lemery-Chalfant, K., Schreiber, J. E., Schmidt, N. L., Van Hulle, C. A., Essex, M. J., &Goldsmith, H. H. (2007). Assessing internalizing, externalizing, and attentionproblems in young children: Validation of the MacArthur HBQ. Journal of theAmerican Academy of Child and Adolescent Psychiatry, 46, 1315–1323.

Lengua, L. J. (2002). The contribution of emotionality and self-regulation to theunderstanding of children's responses to multiple risk. Child Development, 73, 144–161.

Lengua, L. J. (2003). Associations among emotionality, self-regulation, adjustmentproblems, and positive adjustment in middle childhood. Applied DevelopmentalScience, 24, 595–618.

Lengua, L. J., Honorado, E., & Bush, N. R. (2007). Contextual risk and parenting aspredictors of effortful control and social competence in preschool children. Journalof Applied Developmental Psychology, 28, 40–55.

Li-Grining, C. P. (2007). Effortful control among low-income preschoolers in threecities: Stability, change, and individual differences. Developmental Psychology, 43,208–221.

Luby, J. L., Heffelfinger, A., Measelle, J. R., Ablow, J. C., Essex, M. J., Dierker, L., et al.(2002). Differential performance of the MacArthur HBQ and DISC-IV in identifyingDSM-IV internalizing psychopathology in young children. Journal of the AmericanAcademy of Child and Adolescent Psychiatry, 41, 458–466.

Luthar, S. S. (1999). Poverty and children's adjustment. Newbury Park, CA: Sage.Luthar, S. S. (2006). Resilience in development: A synthesis of research across five

decades. In D. Cicchetti, & D. J. Cohen (Eds.), Developmental psychopathology, vol. 3,2nd., (pp. 739–795). Hoboken, NJ: Wiley.

Masten, A. S. (1992). Homeless children in the United States: Mark of a nation at risk.Current Directions in Psychological Science, 1, 41–44.

Masten, A. S., Long, J. D., Roisman, G. I., Burt, K. B., Obradović, J., Roberts, J. M., et al. (2005).Developmental cascades: Linking academic achievement, externalizing and internal-izing symptoms over 20 years. Developmental Psychology, 41, 733–746.

Masten, A. S., Miliotis, D., Graham-Bermann, S. A., Ramirez, M., & Neemann, J. (1993).Children in homeless families: Risks to mental health and development. Journal ofConsulting and Clinical Psychology, 61, 335–343.

Masten, A. S., & Obradović, J. (2006). Competence and resilience in development. Annalsof the New York Academy of Sciences, 1094, 13–27.

Masten, A. S., & Sesma, A. (1999). Risk and resilience among children homeless inMinneapolis. Center for Urban and Regional Affairs Reporter, 29, 1–6.

Masten, A. S., Sesma, A., Si-Asar, R., Lawrence, C., Miliotis, D., & Dionne, J. A. (1997).Educational risks for children experiencing homelessness. Journal of SchoolPsychology, 35, 27–46.

McLoyd, V. C., Aikens, N. L., & Burton, L. M. (2006). Childhood poverty, policy,and practice. In K. A. Renninger, & I. E. Sigel (Eds.), Handbook of child psychology:Vol. 4. Child psychology in practice 6th ed., (pp. 700–775). Hoboken, NJ: Wiley &Sons.

Mezzacappa, E. (2004). Alerting, orienting, and executive attention: Developmentalproperties and sociodemographic correlates in an epidemiological sample ofyoung, urban children. Child Development, 75(5), 1373–1386.

Milotis, D., Sesma, A., & Masten, A. S. (1999). Parenting as a protective process for schoolsuccess in children from homeless families. Early Education & Development, 10,111–133.

117J. Obradović / Journal of Applied Developmental Psychology 31 (2010) 109–117

NICHD Early Child Care Research Network. (2005). Predicting individual differences inattention, memory, and planning in first graders from experiences at home, childcare, and school. Developmental Psychology, 41, 99–114.

Obradović, J., Long, J. D., Cutuli, J. J., Chan, A., Hinz, E., Heistad, D., et al. (2009). Academicachievement of homeless and highly mobile children in an urban school district:Longitudinal evidence on risk, growth, and resilience. Development and Psychopa-thology, 21, 493–518.

Olson, S. L., Sameroff, A. J., Kerr, D. C. R., Lopez, N. L., & Wellman, H. M. (2005).Developmental foundation of externalizing problems in young children: The role ofeffortful control. Development and Psychopathology, 17, 25–45.

Patterson, G. R., & Bank, L. (1986). Bootstrapping your way in the nomological thicket.Behavioral Assessment, 8, 49–73.

Rafferty, Y., & Shinn, M. (1991). The impact of homelessness on children. AmericanPsychologist, 46, 1170–1179.

Rafferty, Y., Shinn, M., & Weitzman, B. C. (2004). Academic achievement among formerlyhomeless adolescents and their continuously housed peers. Journal of SchoolPsychology, 42, 179–199.

Rescorla, L., Parker, R., & Stolley, P. (1991). Ability, achievement, and adjustment inhomeless children. American Journal of Orthopsychiatry, 61, 210–220.

Riggs, N. R., Blair, C. B., & Greenberg, M. T. (2003). Concurrent and 2-year longitudinalrelations between executive function and the behavior of 1st and 2nd gradechildren. Child Neuropsychology, 9, 267–276.

Riggs, N. R., Greenberg, M. T., Kusche, C. A., & Pentz, M. A. (2006). The mediational role ofneurocognition in the behavioral outcomes of a social–emotional prevention programin elementary school students: Effects of the PATHS curriculum. Prevention Science, 7,91–102.

Rog, D. J., & Buckner, J. C. (2007, March). Homeless families and children. NationalSymposium on Homelessness Research. Paper presented at the National Home-lessness Research Symposium, Washington, D.C.

Rothbart, M. K., & Bates, J. E. (1998). Temperament. In W. Damon, & N. Eisenberg (Eds.),Handbook of childhood psychology: Social, emotional, personality development, Vol. 3.(pp. 105–176), New York: Wiley.

Rothbart, M. K., Sheese, B. E., & Posner, M. I. (2007). Executive attention and effortfulcontrol: Linking temperament, brain networks, and genes. Child DevelopmentPerspectives, 1, 2–7.

Rubin, D. H., Erickson, C. J., Augustin, M. S., Cleary, S. D., Allen, J. K., & Cohen, P. (1996).Cognitive and academic functioning of homeless children compared with housedchildren. Pediatrics, 97, 289–294.

Rueda, M. R., Rothbart, M. K., McCandliss, B. D., Saccomanno, L., & Posner, M. I. (2005).Training, maturation, and genetic influences on the development of executiveattention. Proceedings of the National Academy of Sciences of the USA, 102(41),14931–14936.

Sameroff, A. J., Seifer, R., Zax, M., & Barocas, R. (1987). Early indicators of developmentalrisk: The Rochester Longitudinal Study. Schizophrenia Bulletin, 13, 383–393.

Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.Psychological Methods, 7, 147–177.

Senn, T. E., Espy, K. A., & Kaufmann, P. M. (2004). Using path analyses to understandexecutive function organization in preschool children. Developmental Neuropsy-chology, 26, 445–464.

Strommen, E. A. (1973). Verbal self-regulation in a children's game: Impulsive errors on“Simon says”. Child Development, 44, 849–853.

Tarullo, A. R., Obradović, J., & Gunnar, M. R. (2009). Self-control and the developingbrain. Zero To Three, 29, 31–37.

Valiente, C., Eisenberg, N., Smith, C. L., Reiser, M., Fabes, R. A., Losoya, S., et al. (2003).The relations of effortful control and reactive control to children's externalizingproblems: A longitudinal assessment. Journal of Personality, 71, 1171–1196.

Vostanis, P., Grattan, E., Cumella, S., & Winchester, C. (1997). Psychosocial functioningof homeless children. Journal of the American Academy of Child and AdolescentPsychiatry, 36, 881–889.

Wechsler, D. (2002). WPPSI-III administration and scoring manual. San Antonio, TX:Psychological Corporation.

Zelazo, P. D. (2006). The Dimensional Change Card Sort (DCCS): A method of assessingexecutive function in children. Nature Protocols, 1, 1–6.

Zima, B. T., Wells, K. B., & Freeman, H. E. (1994). Emotional and behavioral problems andsevere academic delays among sheltered homeless children in Los Angeles County.American Journal of Public Health, 84, 260–264.