Individual differences in growth in executive function across the transition to school predict externalizing and internalizing behaviors and self-perceived academic success at 6 years of age Claire Hughes * , Rosie Ensor Centre for Family Research, University of Cambridge, Cambridge CB2 3RF, UK article info Article history: Available online 29 July 2010 Keywords: Executive function Externalizing problems Internalizing problems Academic competence Individual differences Trajectories School transition abstract Building on an existing latent variable analysis of executive func- tion (EF) in children (N = 191, 57% boys and 43% girls) making the transition to school (Hughes et al. (2010), Developmental Neuro- psychology, vol. 35, pp. 20–36), the current study both documented average developmental improvements from 4 to 6 years of age and examined individual differences in EF growth in relation to latent factors for two sets of child outcome measures at 6 years: (a) first-grade teachers’ ratings of emotional symptoms, hyperactivity, and conduct/peer problems and (b) children’s self-perceived aca- demic and social competencies. With effects of concurrent verbal ability and EF controlled, variation in EF slopes across the transi- tion to school predicted variation in latent constructs for (a) all four problem behavior subscales and (b) children’s self-reported aca- demic (but not social) competence. These findings underscore the clinical and educational significance of early individual differences in EF and highlight the value of adopting a developmental perspective. Ó 2010 Elsevier Inc. All rights reserved. Introduction Executive function (EF) is an umbrella term that encompasses the set of higher order processes (e.g., inhibitory control, working memory, attentional flexibility), associated with the prefrontal cor- 0022-0965/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jecp.2010.06.005 * Corresponding author. E-mail addresses: [email protected], [email protected](C. Hughes). Journal of Experimental Child Psychology 108 (2011) 663–676 Contents lists available at ScienceDirect Journal of Experimental Child Psychology journal homepage: www.elsevier.com/locate/jecp
Transcript
Journal of Experimental Child Psychology 108 (2011) 663–676
Contents lists available at ScienceDirect
Journal of Experimental ChildPsychology
journal homepage: www.elsevier .com/locate/ jecp
Individual differences in growth in executive functionacross the transition to school predict externalizing andinternalizing behaviors and self-perceived academicsuccess at 6 years of age
Claire Hughes *, Rosie EnsorCentre for Family Research, University of Cambridge, Cambridge CB2 3RF, UK
Building on an existing latent variable analysis of executive func-tion (EF) in children (N = 191, 57% boys and 43% girls) makingthe transition to school (Hughes et al. (2010), Developmental Neuro-psychology, vol. 35, pp. 20–36), the current study both documentedaverage developmental improvements from 4 to 6 years of age andexamined individual differences in EF growth in relation to latentfactors for two sets of child outcome measures at 6 years: (a)first-grade teachers’ ratings of emotional symptoms, hyperactivity,and conduct/peer problems and (b) children’s self-perceived aca-demic and social competencies. With effects of concurrent verbalability and EF controlled, variation in EF slopes across the transi-tion to school predicted variation in latent constructs for (a) all fourproblem behavior subscales and (b) children’s self-reported aca-demic (but not social) competence. These findings underscore theclinical and educational significance of early individual differencesin EF and highlight the value of adopting a developmentalperspective.
� 2010 Elsevier Inc. All rights reserved.
Introduction
Executive function (EF) is an umbrella term that encompasses the set of higher order processes(e.g., inhibitory control, working memory, attentional flexibility), associated with the prefrontal cor-
664 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
tex, that govern goal-directed action and adaptive responses to novel, complex, or ambiguous situa-tions (Hughes, Graham, & Grayson, 2005). The past two decades have seen massive growth in theavailability of child-friendly EF tasks (e.g., Carlson, 2005; Diamond, Prevor, Callender, & Druin,1997; Espy, Kaufmann, McDiarmid, & Glisky, 1999; Hughes, 1998; Zelazo & Müller, 2002), leadingto dramatic improvements in our understanding of the development of EF. For instance, it is nowknown that EF is a unitary construct with partially dissociable components (Garon, Bryson, & Smith,2008) that begins to emerge during the first few years of life (e.g., Diamond, 1991), continues to de-velop through to adulthood (Huizinga, Dolan, & van der Molen, 2006), and shows strong associationswith cognitive characteristics such as language ability and understanding of false beliefs (e.g., Carlson,Moses, & Breton, 2002; Hughes, 1998; Hughes & Ensor, 2007).
However, concerns have been raised about the validity of findings from studies of children’s devel-oping EF skills. For example, in their recent review, Garon et al. (2008) noted that simplifying adulttasks to make them age appropriate for young children carries the danger of losing the critical EF com-ponent. In addition, the construct validity of children’s EF scores may be jeopardized by the fact thatperipheral cognitive constraints (e.g., in verbal comprehension) significantly affect children’s task per-formances. Similarly, Zelazo and Müller (2002) noted that correlations among measures can result notonly from similarities in the mechanisms underlying task performance but also from age-specific ef-fects and shared method variance (e.g., many EF tasks have a similarly strong verbal load). Fortunately,effects of measurement error can be partitioned out statistically (e.g., via latent variable modeling),and recent years have seen an increase in the accessibility of these techniques. As a result, researchersare increasingly conducting confirmatory factor analyses to test different models of EF development(e.g., Brookshire, Levin, & Song, 2004; Gathercole, Pickering, Ambridge, & Wearing, 2004; Lehto, Juujär-vi, Kooistra, & Pulkkinen, 2003).
The current study builds on existing latent variable analyses of EF in a socially diverse sample ofchildren at 4 and 6 years of age across the transition to school (Hughes, Ensor, Wilson, & Graham,2010). The results of this prior study added to the literature on associations between individual differ-ences in preschoolers’ EF skills and individual differences in family socioeconomic status (Hughes &Ensor, 2005; Mezzacappa, 2004) and verbal ability (e.g., Hughes, 1998; Hughes & Ensor, 2007). Specif-ically, latent growth models showed that although both family income and verbal ability predictedindividual differences in EF at 4 years (i.e., EF intercepts), only verbal ability predicted gains in EFacross the transition to school (i.e., EF slopes). In particular, verbally less able children showed greatergains in EF than their peers. This may be because the changes in children’s linguistic environmentsthat accompany the transition to school are likely to be particularly marked for verbally less able chil-dren and may well contribute to developmental improvements in children’s EF skills. Alternatively,the transition to school may foster children’s EF skills by increasing their participation in structuredactivities; this effect may be particularly important for verbally less able children who are more relianton external support. Support for this latter proposal comes from the recent finding that family chaospredicts reduced gains in EF across the preschool years (Hughes & Ensor, 2009).
The current study extends the above analyses (reported by Hughes et al., 2010) in three ways. Spe-cifically, we document average improvements in EF from 4 to 6 years of age and examine children’s EFgrowth in relation to both first-grade teachers’ ratings of problem behaviors (emotional symptoms,hyperactivity, and conduct/peer problems) and children’s self-reported academic and social successin first grade. Below, the background to each of these aims is presented in turn.
Average developmental change in EF from 4 to 6 years of age
Evidence to date suggests age-related contrasts in EF across the whole developmental span (Brocki& Bohlin, 2004; Carlson, Mandell, & Williams, 2004; Hughes & Ensor, 2007; Huizinga et al., 2006).However, most previous studies have been cross-sectional in design and have not used a latent vari-able approach. As a result, these findings are confounded by both cohort effects and developmentalchanges in how children cope with peripheral task demands. Recent research findings on the ‘‘Flynneffect,” an eponymous term used to refer to the striking population gains on standardized intelligencetests seen over the past few decades (Flynn, 1984), provide a salutary reminder of the need to considersuch potential confounds. Specifically, Wicherts et al. (2004) showed that measurement invariance
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 665
across cohorts is untenable, indicating that gains in IQ scores across birth cohorts may simply reflect ameasurement artifact. Compared with standardized IQ tests, the psychometric properties of many EFtasks are relatively weak (e.g., Bishop, Aamodt-Leeper, Creswell, McGurk, & Skuse, 2001); as a result,age-related contrasts in EF performance may also reflect measurement artifacts.
This possibility is especially salient for comparisons of preschoolers and school-aged children be-cause the transition to school is characterized by improvements in numerous salient nonexecutivefactors, including language ability, pragmatic understanding, and compliance. Reassuringly, however,Hughes et al. (2010) were able to demonstrate that a model of a unitary EF construct at 4 and 6 yearsof age showed a good fit to the data even when potential effects of measurement artifacts were re-moved by holding the metric constant across time points. However, space limitations precludedreporting on further analyses to estimate the magnitude of change in the EF latent factors from 4 to6 years of age. The first (preliminary) aim of the current study was to rectify this omission so as toquantify the average improvement in EF between 4 and 6 years of age.
Does variation in EF growth predict teacher-rated problem behaviors?
For adults, there is robust evidence that deficits in EF are associated with problems of antisocialbehavior. In their meta-analytic review of this literature, Morgan and Lilienfeld (2000) reported thatthe average EF performance of antisocial groups fell 0.62 standard deviations below that of controlgroups (although there was only inconsistent evidence that this deficit was specific to EF). For chil-dren, evidence linking EF impairments and psychopathology was reviewed in an influential articleby Pennington and Ozonoff (1996), who concluded that EF deficits are consistently found in bothattention deficit/hyperactivity disorder (ADHD) and autism but not in conduct disorder (withoutADHD) or in Tourette’s syndrome. Confirming the value of considering childhood disorders as ‘‘normaldevelopment gone awry” (Rutter & Quinton, 1984), several studies have shown that young childrenwith elevated rates of problem behaviors demonstrate clear deficits in EF (e.g., Hughes & Dunn,2000; Hughes, Dunn, & White, 1998; Moffitt, 1993; Pennington & Bennetto, 1993; Seguin, Boulerice,Harden, Tremblay, & Pihl, 1999).
More recently, a recent longitudinal study (involving just over half of the children participating inthe current study) has used multi-informant, multi-measure, multi-setting aggregate measures ofproblem behaviors at 2, 3 and 4 years of age to examine links with individual differences in youngchildren’s performances on tests of EF, verbal ability, and theory of mind (i.e., emotion understandingand false belief comprehension) at each time point (Hughes & Ensor, 2008) and produced three keyconclusions. First, by 4 years, individual differences in problem behaviors showed robust and specificassociations with individual differences in EF (but not in theory of mind or verbal ability). Second, poorEF performance at 2 years predicted worsening problem behaviors from 2 to 4 years of age. Third, indi-vidual differences in EF at 3 years fully mediated the influence of language deficits at 2 years on prob-lem behaviors at 4 years. The current study builds on these findings in two ways: by examiningpredictive relations between early EF and later problem behaviors across the transition to schooland by using a latent growth model to assess whether variation in EF growth predicts individual dif-ferences in teacher-rated problem behaviors.
Does variation in EF growth predict children’s self-perceived academic and social success?
Our third aim was to build on recent findings that highlight the educational impact of preschool EF.In particular, Blair and colleagues showed that individual differences in EF before the start of schoolpredicted both school readiness and children’s success in numeracy and literacy (Blair & Diamond,2008; Blair & Peters, 2003; Blair & Razza, 2007; Razza & Blair, 2009). To date, however, no studyhas examined whether individual differences in early EF are related to children’s own perceptionsof their success at school, even though developmentally appropriate assessments have been availablefor some time (e.g., Harter & Pike, 1984; Measelle, Ablow, Cowan, & Cowan, 1998). This is an importantomission because theorists have long argued that self-perceptions play a key role in shaping children’sbehavior (e.g., Cicchetti, 1993; Rutter, 1989). Much of the interest in children’s self-perceptions hasbeen generated by the finding that siblings often differ markedly from each other (in many
666 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
dimensions or phenotypes) despite the fact that they live in the same family and typically share 50% oftheir genes (Dunn & Plomin, 1991). These contrasts between siblings suggest that differences in howchildren perceive themselves and their environments may be as important as ‘‘objective” differencesbetween children and their environments. Consistent with this view, there is evidence that whereaspositive self-perceptions predict favorable outcomes such as academic achievement (Marsh, Ellis, &Craven, 2002) and peer acceptance (Boivin & Bégin, 1989), negative self-perceptions are associatedwith depression (e.g., Harter & Jackson, 1993) and peer rejection (Coplan, Findlay, & Nelson, 2004).
In addition, at least one study of adults has shown that individual differences in self-esteem showrobust associations with individual differences in self-control, a construct that is closely related to EF(Tangney, Baumeister, & Boone, 2004). Similar indirect support for the proposal that early EF promotespositive self-perceptions is available from least one study of children. Specifically, Maszk, Eisenberg,and Guthrie (1999) found that 4- to 6-year-olds rated by peers and teachers as high in behavioraland emotional self-control became increasingly popular over the school year. This finding suggeststhat individual differences in self-control may be meaningful for how children are viewed by othersand, hence, for how they view themselves. Therefore, the current study adds to the literature by exam-ining whether EF growth across the transition to school predicts individual differences in children’sself-perceptions, an important aspect of children’s school readiness.
Given that existing findings suggest positive associations between early EF and academic success(Blair & Razza, 2007; Razza & Blair, 2009) and between early self-control and popularity (Maszket al., 1999), we decided to use Harter and Pike’s (1984) Pictorial Scale to assess whether EF growthpredicted children’s self-perceptions in relation to two salient but distinct aspects of school success:academic competence and social competence (the physical competence subscale was also adminis-tered but is not analyzed here). We hypothesized that each of these predictive relationships wouldbe significant. We also expected that these relationships would be distinct and independent from eachother for at least two reasons. The first reason is that theoretical models have, over time, increasinglyemphasized the multifaceted nature of children’s self-perceptions (e.g., Harter, 1998), such that chil-dren may hold very positive views of their academic ability coupled with more negative views of theirsuccess with peers (or vice versa). The second is that different aspects of EF are likely to be particularlysalient for distinct aspects of adjustment and, hence, for the corresponding distinct facets of children’sself-perceptions. For example, in a recent theoretical review of factors underpinning children’s devel-oping social skills, Beauchamp and Anderson (2010) argued that, with regard to influences of EF, atten-tional control (i.e., self-monitoring, response inhibition, and self-regulation) is especially critical. Incontrast, other aspects of EF, such as working memory, appear to be critical for performance in aca-demic tests of arithmetic or reading (e.g., Adams & Hitch, 1997; Cain, Oakhill, & Bryant, 2004; Waber,Gerber, Turcios, Wagner, & Forbes, 2006).
In sum, the current study used latent variable analyses to chart the magnitude of average EF growthacross the transition to school and to examine predictive relations between EF growth and teacher-rated problem behaviors as well as children’s self-perceptions of their academic and social success.
Methods
Sample
The 191 children (108 boys and 83 girls) in this study all were participating in a larger investigationaimed at predicting success in the transition to school. Recruitment took place via schools in andaround Cambridge, United Kingdom, with priority given to schools serving low-income areas (Cam-bridge is a relatively prosperous city that has pockets of deprivation and is surrounded by high levelsof rural poverty). A participation rate of 81% was achieved by repeated visits to schools to facilitateface-to-face conversations with parents (40% of families returned signed consent forms via the teach-ers and 60% signed consent forms after meeting the research team). At Time 1, the mean age for thechildren was 4 years 3 months (SD = 5 months, range = 3 years 7 months to 4 years 11 months). AtTime 2, the mean age for the children was 6 years 0 month (SD = 4 months, range = 5 years 5 monthsto 6 years 11 months). The sample showed wide variation in verbal mental age, assessed using the
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 667
British Picture Vocabulary Scale (BPVS) (Dunn, 1997). Specifically, at the first time point the mean ver-bal mental age was 4.02 years (SD = 1.08, range = 1.58–7.50), and at the second time point the meanverbal mental age was 5.92 years (SD = 1.43, range = 2.42–10.17 years). Reflecting the local popula-tion, the sample was ethnically very homogeneous (97% of the children were Caucasian).
Data on total family income were available for 183 families, with the following distribution: lessthan £10,000 for 31 families, £10,000 to £20,000 for 47 families, £20,000 to £30,000 for 37 families,£30,000 to £40,000 for 29 families, and more than £40,000 for 39 families. Note that the median totalfamily income in the United Kingdom is £20,000, making the poverty line (defined as 60% of the med-ian family income) £12,000; current statistics suggest that 14% of British children live in poverty(Cooke & Lawton, 2008). Thus, 40% of our sample had a total family income below the national median,and at least 16% were living in poverty. Data on head-of-household occupational status were availablefor 185 families: 34 unskilled, 54 skilled/semiskilled, 22 technical/secretarial, and 75 managerial/pro-fessional. In total, 186 families provided data on maternal education, showing that 19 mothers had noeducational qualifications at all, 58 mothers had elementary (16-year-old) qualifications, 50 mothershad higher (18-year-old) qualifications, and 59 mothers had degrees. Family income, occupational sta-tus, and maternal education were substantially intercorrelated, with r values ranging from .46 to .53(and a mean of .50). All three measures were also correlated with verbal mental age, with mean r = .20at 4 years of age and mean r = .33 at 6 years of age.
Procedures
At each time point (4 and 6 years of age), researchers visited each child at the nursery or school for2 h. During this visit, researchers conducted face-to-face interviews with teachers to complete theStrengths and Difficulties Questionnaire (SDQ) (Goodman, Forth, Simmons, Gatward, & Meltzer,2003). Details of the five items for each of the four subscales of this questionnaire are given in Table 1.
Table 1Descriptive statistics for teachers’ ratings of emotional symptoms, hyperactivity, and conduct/peer problems on the Strengths andDifficulties Questionnaire.
Item M SD Range
Emotional symptomsOften complains of headaches, stomach-aches, or sickness .24 .43 0–1Many worries, often seems worried .39 .49 0–1Often unhappy, downhearted, or tearful .19 .40 0–1Nervous or clingy in new situations, easily loses confidence .37 .48 0–1Many fears, easily scared .21 .41 0–1
HyperactivityRestless, overactive, cannot stay still for long .41 .49 0–1Constantly fidgeting or squirming .38 .49 0–1Easily distracted, concentration wanders .57 .50 0–1Thinks things out before actinga .66 .48 0–1Sees tasks through to the end, good attention spana .56 .50 0–1
Conduct problemsOften has temper tantrums or hot tempers .15 .36 0–1Generally obedient, usually does what adults requesta .22 .42 0–1Often fights with other children or bullies them .19 .39 0–1Often lies or cheats .21 .41 0–1Steals from home, school, or elsewhere .04 .21 0–1
Peer problemsRather solitary, tends to play alone .33 .47 0–1Has at least one good frienda .19 .39 0–1Generally liked by other childrena .19 .39 0–1Picked on or bullied by other children .08 .27 0–1Gets on better with adults than with other children .27 .44 0–1
a Reversed.
668 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
In addition, each child completed a comprehensive cognitive assessment consisting of the BPVS, Har-ter and Pike’s (1984) Pictorial Scale (for which details are given in Table 2), and the EF tasks describedbelow. Schools received gift tokens as thanks for their time.
EF tasks
Our analyses focus on three EF tasks that were administered at both 4 and 6 years of age: the Day/Night Stroop test of inhibitory control (Gerstadt, Hong, & Diamond, 1994), the Beads working memorysubtest of the Stanford Binet Intelligence Scales (Thorndike, Hagen, & Sattler, 1986), and the Tower ofLondon planning task (Shallice, 1982).
Day/Night StroopThis task (Gerstadt et al., 1994) consisted of two conditions. On each of the 12 trials in the first con-
dition, children were shown one of two abstract patterns and were asked to say either ‘‘day” or‘‘night.” Thus, this condition did not require children to inhibit a prepotent response, but over thecourse of the 12 trials these day/night responses became salient, and so this condition did require chil-dren to inhibit the incorrect response. The second condition also consisted of 12 trials (and the sameresponses), but this time the abstract patterns were replaced by pictures of the moon and the sun andchildren were asked to say ‘‘day” for a picture of the moon and ‘‘night” for a picture of the sun. In bothconditions, pictures were presented in a pseudo-random order and performance was rated by the totalnumber of correct trials.
BeadsIn this task (Thorndike et al., 1986), children were first shown a photograph of an array of 12 beads
(red, blue, and white beads in four shapes: sphere, cone, cylinder, and disk). The five trials on each ofthe first two parts of the task required children to point to the bead that exactly matched the bead (orpair of beads) shown by the experimenter for 2 s (1-bead trials) or 3 s (2-bead trials). In the third (16-trial) part of the task, children were asked to look closely at a picture of beads arranged on a stick(shown for 5 s on each trial) and then to build a matching tower of beads. Our analyses are basedon the total number of correct trials across all three parts of the task. Therefore, children could scorea maximum of 26 points.
Tower of LondonThis task (Shallice, 1982) is an adaptation of the Tower of Hanoi (Piaget, 1974/1976) and involves
generation of plans and inhibition of maladaptive responses. The apparatus includes a wooden board(100 � 250 � 10 mm) with three pegs of different sizes (240, 160, and 80 mm) and three sponge ballsof different colors (red, green, and blue), each approximately 80 mm in diameter with a plastic pipe
Table 2Descriptive statistics for children’s responses to questions about academic and social competencies.
Item M SD Range
Academic competenceGood at numbers .69 .46 0–1Knows lots of things in school .56 .50 0–1Good at reading by self .53 .50 0–1Good at writing words .62 .49 0–1Good at spelling words .63 .48 0–1Good at adding numbers .60 .49 0–1
Peer acceptanceLots of friends to play with .54 .50 0–1Kids share their toys with you .50 .50 0–1Many friends to play games with .48 .50 0–1Lots of friends to play with on the playground .56 .50 0–1Gets asked to play with other kids .47 .50 0–1Kids want to sit next to you .50 .50 0–1
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 669
through the middle such that they could be placed on the pegs. The big peg could carry all three balls,the middle peg could carry two balls, and the little peg could carry just one ball. Individual problemswere presented by showing children large pictures of goal arrangements with the number of movesrequired clearly shown, beginning with a warm-up set of simple 1-move problems (i.e., children’sarrangement of balls could be made to match the goal arrangement by moving just one ball). Childrenwere told that the aim of the task was to copy the goal arrangement by moving the balls one at a time(large sponge balls were chosen to make it difficult for children to hold more than one ball in theirhand). On each problem, children were encouraged to think about how they could solve the problembefore they began to move the balls. Children were given three problems at each of three levels: min-imum n (moves) = 2, 3, or 4. On each problem, they received 2 points for a perfect solution (i.e., a solu-tion with the minimum number of moves), 1 point for a solution within the allowed number of moves(2n + 1), and 0 points for any other response. We rated children’s performances on each of the three 2-,3-, and 4-move problems. Therefore, children could score a maximum of 18 points.
Perceived competence and peer acceptance
We used Harter and Pike’s (1984) Pictorial Scale, including all items for two subscales: academiccompetence and social competence (for details of each item, see Table 2). This structured interviewwas administered using separate sets of pictures for boys and girls, with identical activities depictedin each set of plates. Items occurred in the order of academic competence, peer acceptance, and phys-ical competence and then continued to repeat themselves in that order (for reasons of time, the fourthsubscale, tapping perceived maternal acceptance, was not included). Within each subscale, the pic-tures were counterbalanced so that the most competent/accepted child was on the left in three pic-tures and on the right in three pictures. After hearing the statement accompanying each picture,children first were asked which child they were most like and then were asked to choose betweena big circle and a little circle to indicate whether they were a lot like that girl/boy or just a little bitlike that girl/boy.
Results
Our analyses build directly on Hughes et al.’s (2010) latent variable analyses of executive functionacross the transition to school. The models included indicators of children’s performance on three EFtasks at 4 and 6 years of age: the Day/Night Stroop test of inhibitory control (Gerstadt et al., 1994), theBeads working memory subtest of the Stanford Binet Intelligence Scales (Thorndike et al., 1986), andthe Tower of London planning task (Shallice, 1982). We used MPlus 5 and a maximum likelihood func-tion to analyze the sample variance–covariance matrix. Good model fit was evaluated using the fol-lowing indexes and criteria: standardized root mean square residual (SRMR) 6 .08, root meansquare error of approximation (RMSEA) 6 .06, and comparative fit index (CFI) and Tucker–Lewis index(TLI), where values P .95 indicate good model fit and values in the range from .90 to .95 suggest ade-quate model fit.
Developmental change in EF from 4 to 6 years of age
Hughes et al. (2010) specified a model in which, at each time point (4 and 6 years of age), measuresof planning, inhibitory control, and working memory loaded onto a latent EF factor. This model fittedthe data well even when the measurement properties were constrained to be the same across-timepoints. In other words, Hughes and colleagues established an underlying construct of EF at 4 and6 years with measurement invariance (i.e., for each latent factor value, the observed values of eachpair of indicators were statistically equivalent), a fundamental prerequisite for evaluating temporalchange in a latent construct. If such factor loading and indicator intercepts invariance is untenable,it cannot be determined whether temporal change in the latent construct (EF here) is due to truechange rather than across-time changes in nonexecutive aspects of performance on EF tasks (i.e., mea-surement error).
670 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
In the current study, we extended Hughes et al.’s (2010) confirmatory factor analysis describedabove by assessing the average magnitude of changes in children’s latent EF scores between 4 and6 years of age. Specifically, we constrained the within-group variability (i.e., dispersion) of the EF la-tent factors to be equal across time to ensure the interpretability of the latent factor means. This con-straint significantly degraded the model fit, v2 difference (1) = 11.94, p < .01. That is, children’sperformance on the three EF tasks drew on different ranges of the EF latent factor at 4 and 6 years.Specifically, as shown by the original model solution, the variance of the EF latent factor at 4 years(9.73) was more than double the variance of the EF latent factor at 6 years (4.40), suggesting thatthe data showed a clear regression to the mean. When the variances of the EF latent factors con-strained to be equal across the two time points (5.97), the standardized latent factor means were3.90 at 4 years and 5.92 at 6 years. To test whether this difference was significant, we constrainedthe means of the EF latent factors to be equal across the two time points. This constraint significantlydegraded the absolute model fit, v2 difference (1) = 206.91, p < .01.
The means of the EF latent factors were set by the thresholds of first indicator, namely the Beadsworking memory task, for which possible scores ranged from 0 to 26 points. When we constrainedthe metric and variance to be equal at each time point, the mean latent factor scores were 9.65 at4 years of age and 14.32 at 6 years of age. This average increase corresponded to 2 standard deviations.Note also that children’s scores on the latent EF factor at 4 years ranged from 2 to 16, with 80% ofscores falling below the minimum score (i.e., 10) at 6 years. However, the top 20% of 4-year-oldshad latent factor scores similar to those of the average 6-year-old, highlighting the magnitude of indi-vidual differences in preschool EF.
Variation in growth in EF
Hughes et al.’s (2010) latent growth model of children’s performance on planning, inhibitory con-trol, and working memory tasks at 4 and 6 years of age showed significant mean initial levels and in-creases over time and, most important for the current study, significant variation around the interceptand slope (see Fig. 1). In this study, we explored correlates of variability of growth in EF across the
-10
-5
0
5
10
15
20
Age 4 Age 6
Fig. 1. Individual differences in growth in executive function from 4 to 6 years of age.
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 671
transition to school. Specifically, we derived latent factors from the relevant subscales of the SDQ(Goodman et al., 2003) and Harter and Pike’s (1984) Pictorial Task.
Turning first to the SDQ, for each of the 20 difficulty questions, teachers tended not to respond ‘‘cer-tainly true”; fewer than 10% of teachers responded ‘‘certainly true” to 15 or more of the questions.Therefore, we recoded teachers’ responses into binary formats (0/1) by combining ‘‘somewhat true”and ‘‘certainly true” responses. Table 1 shows the descriptive statistics. We used categorical data fac-tor analysis to specify that ratings for each of the four subscales (emotional symptoms, conduct prob-
Fig. 2. Variance in growth in executive function predicted lower teacher-rated emotional symptoms, hyperactivity, andconduct/peer problems as well as higher self-perceived academic competence.
672 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
lems, hyperactivity, and peer problems) loaded onto separate factors. Note that although 50 teacherscompleted an SDQ for only 1 child, 33 teachers provided SDQ ratings for 2–4 children and 6 teacherscompleted SDQs for 5–11 children. To account for this cluster sampling of teachers, we adjusted thevariances of the latent factors. In addition, we permitted the four latent factors to be correlated.
Children’s responses to each of the 12 questions on the Pictorial Task were also skewed. In partic-ular, for each question, fewer than 14% of children selected the first response category (e.g., ‘‘not toogood at numbers,” ‘‘hardly any friends”), whereas more than 47% of children selected the fourth re-sponse category (e.g., ‘‘really good at numbers,” ‘‘a whole lot of friends”). Therefore, once again we re-coded responses into binary formats, collapsing the first three response categories together. Table 2shows the descriptive statistics. Again using categorical data factor analysis, we specified that ratingsfor each of the two subscales (academic competence and social competence) loaded onto separate fac-tors, which were permitted to be correlated. We covaried the EF intercept (defined as children’s per-formance on the three EF tasks at 6 years of age) with each of the six latent factors to account forconcurrent associations. We regressed the EF slope onto each of these latent variables. In addition,we included the covariance between verbal ability at 6 years and both the EF intercept and slope. Notethat although the intercept is typically the initial value, for some research questions (e.g., interven-tions) the last occasion of measurement is more appropriate (Preacher, Wichman, MacCallum, & Brig-gs, 2008). In the current study, our aim was to examine whether rate of change in EF across thetransition to school predicted outcome measures over and above associations with concurrent EF.
The model was overidentified with df = 39 and v2 = 61.60, p < .05. Each index suggested adequatemodel fit: RMSEA = .06, CFI = .94, and TLI = 94. Fig. 2 shows completely standardized parameter esti-mates. All factor loadings were statistically significant, and the latent factors explained significant var-iance in all of their purported indicators (p < .01). Thus, teachers’ ratings were reliable indicators ofemotional symptoms, hyperactivity, and conduct/peer problems, and children’s ratings were reliableindicators of their perceived academic and social competencies.
The coefficients shown in bold in Fig. 2 are statistically significant. Variation in the EF slope signif-icantly predicted the latent factors of teacher-rated emotional symptoms, hyperactivity, and conduct/peer problems (p < .01). Although individual differences in the EF slope did not significantly predictthe latent factor of children’s perceived social competence (t = .32, p = .75), the predictive relation withthe latent factor of children’s perceived academic competence was significant (p = .05). In other words,greater gains in EF across the transition to school predicted lower teacher-rated emotional symptoms,hyperactivity, and conduct/peer problems as well as higher perceived academic (but not social)competence.
Discussion
In previous research with the current sample, Hughes et al. (2010) demonstrated the utility of la-tent variable analyses; across the transition to school, verbally less able children showed greater gainsin EF than their peers. The current study extended this work in three ways. First, we conducted furtheranalyses to quantify average gains in EF between 4 and 6 years of age. Second, we examined whetherindividual differences in EF growth predicted latent constructs for first-grade teachers’ ratings of chil-dren’s emotional symptoms, hyperactivity, and conduct/peer problems. Third, we examined whetherindividual differences in EF growth predicted children’s self-reported academic and social compe-tence. Here our findings demonstrated that gains in EF predicted teachers’ ratings of externalizingand internalizing behaviors as well as children’s perceptions of their academic (but not social) com-petence. Note that the multi-method, multi-informant longitudinal design of this study minimizedthe risk of inflated associations and so maximized the reliability of the study findings, which we dis-cuss in turn below.
Average gains in EF between 4 and 6 years of age
As noted in the Introduction, most previous studies of age-related improvements in EF have beencross-sectional in design and have not used a latent variable approach. As a result, these findings are
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 673
confounded by both cohort effects and developmental changes in how children cope with peripheraltask demands. In a recent study involving the current sample of children, Hughes et al. (2010) dem-onstrated the feasibility, validity, and utility of applying latent growth models to examine develop-mental change in EF across the transition to school (i.e., between 4 and 6 years of age).
A first (preliminary) aim of the current study was to extend these analyses to assess the magnitudeof changes across this interval in children’s latent EF scores. As outlined in the Introduction, the tran-sition to school is likely to elicit improvements in children’s abilities to cope with peripheral task de-mands (e.g., on verbal comprehension, attention, and motoric control). Our results suggest that evenwhen such effects are partitioned out, children’s ‘‘true” EF scores show major changes between 4 and6 years of age (with an average increase of 2 standard deviations). Individual differences in the gains inEF made by children across this period were also striking (see also Hughes et al., 2010). Below we dis-cuss the findings from analyses of how these individual differences in EF slopes related to two sets ofoutcome measures at 6 years: teachers’ ratings of children’s behavior and children’s self-perceivedacademic and social competence.
Low gains in EF across the transition to school predict teachers’ ratings of externalizing and internalizingbehaviors
As outlined earlier, there is now robust evidence that EF deficits are associated with antisocialbehavior during adulthood and externalizing problems during childhood. The current study adds tothis research by adopting a developmentally dynamic approach. Specifically, we assessed whetherindividual differences in improvements in children’s EF scores across the transition to school predictedindividual differences in latent factor scores for teacher-rated externalizing and internalizing behav-iors. Our results showed that variation in EF slopes was indeed a significant predictor of latent factorsof emotional symptoms, hyperactivity, and conduct/peer problems. Moreover, these predictive effectsof EF slopes were evident over and above effects of concurrent EF and verbal ability.
In other words, when assessing the relationship between children’s cognitive performance andtheir behavior, children’s growth (rather than their actual performance) appears to be what reallymatters. This finding is open to at least two possible interpretations. First, children who are makingrapid improvements in their cognitive performance are likely to be actively engaged with their school-work, and this engagement may explain why such children show fewer behavioral difficulties thantheir peers. We do not have any direct measures of how ‘‘on task” the study children were at school;this proposal is a direction for future research. A second related possibility is that children’s self-es-teem is likely to be enhanced when they are making rapid cognitive progress; therefore, increasedself-esteem may act as a mediator for this relationship between EF gains and children’s behavior. Test-ing this mediation model would require data from further time points; however, note that this pro-posal is consistent with the finding (discussed below) that EF gains also predicted children’s self-perceived academic competence.
High gains in EF across the transition to school predict children’s self-perceived academic competence
The third set of results from this study concerned predictive relations between EF growth acrossthe transition to school and children’s self-perceptions. Note that, as outlined in the Introduction, chil-dren’s positive self-perceptions have been found to predict both academic achievement (Marsh et al.,2002) and peer acceptance (Boivin & Bégin, 1989), suggesting that self-perceptions may be a key as-pect of children’s psychological (as opposed to practical) school readiness. As noted earlier, includingchildren’s own views about school is important because self-perceptions play an important role inshaping children’s behavior (e.g., Harter & Pike, 1984; Measelle et al., 1998; Rutter, 1989). For exam-ple, poor self-perceptions during early childhood are associated with a range of problems, includingloneliness, withdrawal, and peer exclusion (Coplan et al., 2004).
The current study contributes to this literature by addressing a possible source of influence on indi-vidual differences in children’s self-perceptions. In particular, our results demonstrate that children’sgains in EF across the transition to school predicted their self-perceived academic (but not social)competence at 6 years of age. Interestingly, recent years have seen a growth of research on children’s
674 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
self-perceived academic abilities, fueled by the consistent finding that individual differences in IQ failto account for up to 50% of the variance in academic performance (e.g., Chamorro-Premuzic &Furnham, 2005; Rhode & Thompson, 2007).
Indeed, a meta-analytic review has shown that, even controlling for previous achievement, self-perceived abilities exert small but consistent effects on later achievement (Valentine, DuBois, & Coo-per, 2004). To elucidate the nature of this influence, Greven, Harlaar, Kovas, Chamorro-Premuzic, andPlomin (2009) recently examined data from a large-scale longitudinal twin study. Their findings pro-vide evidence for temporally stable nonshared environmental and genetic influences on all three con-structs (self-perceived abilities, IQ, and achievement), with genetic influencing accounting for roughlyhalf of the variability in both self-perceived abilities and IQ. In their discussion, these authors notedthat research into the mechanisms by which genetic variation leads to phenotypic variation shouldfocus on top-down influences such as personality (cf. Marsh, Trautwein, Lüdtke, Köller, & Baumert,2006). Consistent with this emphasis on top-down processes, our findings suggest that individual dif-ferences in children’s EF growth across the transition to school provide a sensitive and specific predic-tor of variation in children’s self-perceived academic abilities.
Of course, this leaves unresolved the question of whether EF growth should itself be explained interms of nature or nurture. Our prediction is that, as for so many other phenotypes, both nature andnurture (and their interplay) will matter. In addition, even though genetic factors are important forexplaining individual differences in EF (Friedman et al., 2008), heritability does not mean immutabil-ity. Indeed, several preschool interventions have proven to be effective in improving children’s EFskills (e.g., Bierman, Nix, Greenberg, Blair, & Domitrovich, 2008; Diamond, Barnett, Thomas, & Munro,2007). Taken alongside the above literature, our current findings suggest that assisting children tomake gains in EF across the transition to school may have widespread benefits for children’s behavior,their self-concepts, and their academic achievements.
Limitations and conclusions
Although the current study was ambitious in many ways, at least two limitations deserve mention.First, to model EF developments, we could include only tasks that were applied at both 4 and 6 years ofage; as a result, we did not have sufficient tasks to develop a fractionated model to explore which as-pects of EF show particularly strong associations with teacher-rated problem behaviors and self-re-ported academic competence. Second, with only two time points, we were unable to assesswhether the shape of growth in EF matters for children’s school readiness or to conduct a multivariateanalysis of temporal change (Horn, McArdle, & Mason, 1983) such as a curve-of-factors latent growthmodel (in which a growth curve is fitted to latent factors). Despite these limitations, we believe thatthe findings from this study provide a valuable foundation for future research into the interestingquestions of how children’s growing EF skills affect how they view themselves in relation to the chal-lenges posed by the transition to school. In particular, our findings highlight the salience of EF growthfor explaining individual differences in children’s behavior and their self-perceived academiccompetence.
References
Adams, J., & Hitch, G. (1997). Working memory and children’s mental addition. Journal of Experimental Child Psychology, 67,21–38.
Beauchamp, M., & Anderson, V. (2010). SOCIAL: An integrative framework for the development of social skills. PsychologicalBulletin, 136, 39–64.
Bierman, K., Nix, R., Greenberg, M., Blair, C., & Domitrovich, C. (2008). Executive functions and school readiness intervention:Impact, moderation, and mediation in the Head Start REDI program. Development and Psychopathology, 20, 821–843.
Bishop, D., Aamodt-Leeper, G., Creswell, C., McGurk, R., & Skuse, D. (2001). Individual differences in cognitive planning on theTower of Hanoi task: Neuropsychological maturity or measurement error? Journal of Child Psychology and Psychiatry, 42,551–556.
Blair, C., & Diamond, A. (2008). Biological processes in prevention and intervention: The promotion of self-regulation as a meansof preventing school failure. Development and Psychopathology, 20, 899–911.
Blair, C., & Peters, R. (2003). Physiological and neurocognitive correlates of adaptive behavior in preschool among children inHead Start. Developmental Neuropsychology, 24, 479–497.
C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676 675
Blair, C., & Razza, R. (2007). Relating effortful control, executive function, and false belief understanding to emerging math andliteracy ability in kindergarten. Child Development, 78, 647–663.
Boivin, M., & Bégin, G. (1989). Peer status and self-perception among early elementary school children: The case of the rejectedchildren. Child Development, 60, 591–596.
Brocki, K., & Bohlin, G. (2004). Executive functions in children aged 6 to 13: A dimensional and developmental study.Developmental Neuropsychology, 26, 571–593.
Brookshire, B., Levin, H., & Song, J. (2004). Components of executive function in typically developing and head-injured children.Developmental Neuropsychology, 25, 61–83.
Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory,verbal ability, and component skills. Journal of Educational Psychology, 96, 31–42.
Carlson, S. (2005). Developmentally sensitive measures of executive function in preschool children. DevelopmentalNeuropsychology, 28, 595–616.
Carlson, S., Mandell, D., & Williams, L. (2004). Executive function and theory of mind: Stability and prediction from ages 2 to 3.Developmental Psychology, 40, 1105–1122.
Carlson, S., Moses, L., & Breton, C. (2002). How specific is the relation between executive function and theory of mind?Contributions of inhibitory control and working memory. Infant and Child Development, 11, 73–92.
Chamorro-Premuzic, T., & Furnham, A. (2005). Personality and intellectual competence. Mahwah, NJ: Lawrence Erlbaum.Cicchetti, D. (1993). Developmental psychopathology: Reactions, reflections, projections. Developmental Review, 13, 471–502.Cooke, G., & Lawton, K. (2008). Working out of poverty: A study of the low paid and the ‘‘working poor”. London: Institute of Public
Policy Research.Coplan, R., Findlay, L., & Nelson, L. (2004). Characteristics of preschoolers with lower perceived competence. Journal of Abnormal
Child Psychology, 32, 399–408.Diamond, A. (1991). Developmental time course in human infants and infant monkeys, and the neural bases of inhibitory
control in reaching. Annals of the New York Academy of Sciences, 608, 637–704.Diamond, A., Barnett, W., Thomas, J., & Munro, S. (2007). Preschool program improves cognitive control. Science, 318,
1387–1388.Diamond, A., Prevor, M., Callender, G., & Druin, D. (1997). Prefrontal cortex cognitive deficits in children treated early and
continuously for PKU. Monographs of the Society for Research in Child Development, 62(4, Serial No. 252).Dunn, L. (1997). British Picture Vocabulary Scale—Revised. Windsor, UK: NFER–Nelson.Dunn, J., & Plomin, R. (1991). Why are siblings so different? The significance of differences in sibling experiences within the
family. Family Process, 30, 271–283.Espy, K., Kaufmann, P., McDiarmid, M., & Glisky, M. (1999). Executive functioning in preschool children: Performance on A-not-B
and other delayed response format tasks. Brain and Cognition, 41, 178–199.Flynn, J. (1984). The mean IQ of Americans: Massive gains. Psychological Bulletin, 95, 29–51.Friedman, N., Miyake, A., Young, S., DeFries, J., Corley, R., & Hewitt, J. (2008). Individual differences in executive functions are
almost entirely genetic in origin. Journal of Experimental Psychology: General, 137, 201–225.Garon, N., Bryson, S., & Smith, I. (2008). Executive function in preschoolers: A review using an integrative framework.
Psychological Bulletin, 134, 31–60.Gathercole, S., Pickering, S., Ambridge, B., & Wearing, H. (2004). The structure of working memory from 4 to 15 years of age.
Developmental Psychology, 40, 177–190.Gerstadt, C., Hong, Y., & Diamond, A. (1994). The relationship between cognition and action: Performance of children 3½–7
years old on a Stroop-like Day–Night test. Cognition, 53, 129–153.Goodman, R., Forth, T., Simmons, H., Gatward, R., & Meltzer, H. (2003). Using the Strengths and Difficulties Questionnaire (SDQ)
to screen for child psychiatric disorders in a community sample. International Review of Psychiatry, 15, 166–172.Greven, C., Harlaar, N., Kovas, Y., Chamorro-Premuzic, T., & Plomin, R. (2009). More than just IQ: School achievement is predicted
by self-perceived abilities—but for genetic rather than environmental reasons. Psychological Science, 20, 753–762.Harter, S. (1998). The development of self-representations. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Vol.
3. Social, emotional, and personality development (pp. 553–617). New York: John Wiley.Harter, S., & Jackson, B. (1993). Young adolescents’ perceptions of the link between low self-worth and depressed affect. Journal
of Early Adolescence, 33, 383–407.Harter, S., & Pike, R. (1984). The Pictorial Scale of Perceived Competence and Social Acceptance for young children. Child
Development, 55, 1969–1982.Horn, J., McArdle, J., & Mason, R. (1983). When is invariance not invariant: A practical scientist’s look at the ethereal concept of
factor invariance. The Southern Psychologist, 1, 179–188.Hughes, C. (1998). Executive function in preschoolers: Links with theory of mind and verbal ability. British Journal of
Developmental Psychology, 16, 233–253.Hughes, C., & Dunn, J. (2000). Hedonism or empathy? Hard-to-manage children’s moral awareness, and links with cognitive and
maternal characteristics. British Journal of Developmental Psychology, 18, 227–245.Hughes, C., Dunn, J., & White, A. (1998). Trick or treat? Uneven understanding of mind and emotion and executive function
among ‘‘hard to manage” preschoolers. Journal of Child Psychology and Psychiatry & Allied Disciplines, 39, 981–994.Hughes, C., & Ensor, R. (2005). Theory of mind and executive function in 2-year-olds: A family affair? Developmental
Neuropsychology, 28, 645–668.Hughes, C., & Ensor, R. (2007). Executive function and theory of mind: Predictive relations from ages 2 to 4 years. Developmental
Psychology, 43, 1447–1459.Hughes, C., & Ensor, R. (2008). Does executive function matter for preschoolers’ problem behaviors? Journal of Abnormal Child
Psychology, 36, 1–14.Hughes, C., & Ensor, R. (2009). How do families help or hinder the development of executive function? New Directions in Child
and Adolescent Psychiatry, 123, 35–50.Hughes, C., Ensor, R., Wilson, A., & Graham, A. (2010). Tracking executive function across the transition to school: A latent
676 C. Hughes, R. Ensor / Journal of Experimental Child Psychology 108 (2011) 663–676
Hughes, C., Graham, A., & Grayson, A. (2005). Executive function in childhood: Development and disorder. In J. Oates (Ed.),Cognitive development (pp. 205–230). Milton Keynes, UK: Open University Press.
Huizinga, M., Dolan, C., & van der Molen, M. (2006). Age-related change in executive function: Developmental trends and alatent variable analysis. Neuropsychologia, 44, 2017–2036.
Lehto, J., Juujärvi, P., Kooistra, L., & Pulkkinen, L. (2003). Dimensions of executive functioning: Evidence from children. BritishJournal of Developmental Psychology, 21, 59–80.
Marsh, H., Ellis, L., & Craven, R. (2002). How do preschool children feel about themselves? Unraveling measurement andmultidimensional self-concept structure. Developmental Psychology, 38, 376–393.
Marsh, H., Trautwein, U., Lüdtke, O., Köller, O., & Baumert, J. (2006). Integration of multidimensional self-concept and corepersonality constructs: Construct validation and relations to well-being and achievement. Journal of Personality and SocialPsychology, 74, 403–456.
Maszk, P., Eisenberg, N., & Guthrie, I. (1999). Relations of children’s social status to their emotionality and regulation: A short-term longitudinal study. Merrill–Palmer Quarterly, 45, 468–492.
Measelle, J. R., Ablow, J. C., Cowan, P. A., & Cowan, C. P. (1998). Assessing young children’s views of their academic, social, andemotional lives: An evaluation of the self-perception scales of the Berkeley Puppet Interview. Child Development, 69,1556–1576.
Mezzacappa, E. (2004). Alerting, orienting, and executive attention: Developmental properties and sociodemographic correlatesin an epidemiological sample of young, urban children. Child Development, 75, 1373–1386.
Moffitt, T. (1993). The neuropsychology of conduct disorder. Development and Psychopathology, 5, 135–152.Morgan, A., & Lilienfeld, S. (2000). A meta-analytic review of the relation between antisocial behavior and neuropsychological
measures of executive function. Clinical Psychology Review, 20, 113–156.Pennington, B., & Bennetto, L. (1993). Main effects or transactions in the neuropsychology of conduct disorder: A commentary
on the neuropsychology of conduct disorder. Development and Psychopathology, 5, 153–164.Pennington, B., & Ozonoff, S. (1996). Executive function and developmental psychopathology. Journal of Child Psychology and
Psychiatry, 37, 51–87.Piaget, J. (1976). The grasp of consciousness: Action and concept in the young child (S. Wedgewood, Trans.). Cambridge, MA: Harvard
University Press (Original work published 1974).Preacher, K., Wichman, A., MacCallum, R., & Briggs, N. (2008). Latent growth curve modeling. Thousand Oaks, CA: Sage.Razza, R., & Blair, C. (2009). Associations among false-belief understanding, executive function, and social competence: A
longitudinal analysis. Journal of Applied Developmental Psychology, 30, 332–343.Rhode, T., & Thompson, L. (2007). Predicting academic achievement with cognitive ability. Intelligence, 35, 83–92.Rutter, M. (1989). Isle of Wight revisited: Twenty-five years of child psychiatric epidemiology. Journal of the American Academy
of Child and Adolescent Psychiatry, 28, 633–653.Rutter, M., & Quinton, D. (1984). Parental psychiatric disorder: Effects on children. Psychological Medicine, 14, 853–880.Seguin, J. R., Boulerice, B., Harden, P. W., Tremblay, R. E., & Pihl, R. O. (1999). Executive functions and physical aggression after
controlling for attention deficit hyperactivity disorder, general memory, and IQ. Journal of Child Psychology and Psychiatry,40, 1197–1208.
Shallice, T. (1982). Specific impairments in planning. Philosophical Transactions of the Royal Society of London B, 298, 199–209.Tangney, J., Baumeister, R., & Boone, A. (2004). High self-control predicts good adjustment, less pathology, better grades, and
interpersonal success. Journal of Personality and Social Psychology, 72, 271–324.Thorndike, R. L., Hagen, E. P., & Sattler, J. M. (1986). Stanford–Binet Intelligence Scales. Chicago: Riverside.Valentine, J., DuBois, D., & Cooper, H. (2004). The relation between self-beliefs and academic achievement: A meta-analytic
review. Educational Psychologist, 39, 111–133.Waber, D., Gerber, E., Turcios, V., Wagner, E., & Forbes, P. (2006). Executive functions and performance on high-stakes testing in
children from urban schools. Developmental Neuropsychology, 29, 459–477.Wicherts, J., Dolan, C., Hessen, D., Oosterveld, P., van Baal, G., Boomsma, D., et al (2004). Are intelligence tests measurement
invariant over time? Investigating the nature of the Flynn effect. Intelligence, 32, 509–537.Zelazo, P., & Müller, U. (2002). Executive function in typical and atypical development. In U. Goswami (Ed.), Handbook of
childhood cognitive development (pp. 445–469). Oxford, UK: Blackwell.
