Preschool Predictors of School-Age Academic Achievement in Autism Spectrum Disorder (2024)

As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice

Clin Neuropsychol. Author manuscript; available in PMC 2018 Feb 1.

Published in final edited form as:

Clin Neuropsychol. 2017 Feb; 31(2): 382–403.

Published online 2016 Oct 5. doi:10.1080/13854046.2016.1225665

PMCID: PMC5464727

NIHMSID: NIHMS856227

PMID: 27705180

Lauren E. Miller,¹ Jeffrey D. Burke,¹ Eva Troyb,² Kelley Knoch,³ Lauren E. Herlihy,⁴ and Deborah A. Fein¹

Author information Copyright and License information PMC Disclaimer

Abstract

Objective

Characterization of academic functioning in children with autism spectrum disorder (ASD), particularly predictors of achievement, may have important implications for intervention. The current study aimed to characterize achievement profiles, confirm associations between academic ability and concurrent intellectual and social skills, and explore preschool predictors of school-age academic achievement in a sample of children with ASD.

Method

Children with ASD (N = 26) were evaluated at the approximate ages of two, four, and ten years. Multiple regression was used to predict school-age academic achievement in reading and mathematics from both concurrent (i.e., school-age) and preschool variables.

Results

Children with ASD demonstrated a weakness in reading comprehension relative to word reading. There was a smaller difference between mathematics skills; math reasoning was lower than numerical operations, but this did not quite reach trend level significance. Concurrent IQ and social skills were associated with school-age academic achievement across domains. Preschool verbal abilities significantly predicted school-age reading comprehension, above and beyond concurrent IQ, and early motor functioning predicted later math skills.

Conclusions

Specific developmental features of early ASD predict specific aspects of school-age achievement. Early intervention targeting language and motor skills may improve later achievement in this population.

Keywords: autism spectrum disorder, academic achievement, preschool, predictors, outcome

Autism spectrum disorder (ASD) describes a group of neurodevelopmental disorders characterized by impairments in social interaction and communication, and by restricted interests or repetitive behaviors (RRBs) (American Psychiatric Association (APA), 2000; APA, 2013). ASD is highly prevalent in the United States, with current population-based estimates from the Centers for Disease Control and Prevention (CDC) suggesting that one in every 68 children meets diagnostic criteria for the disorder (CDC, 2014).

A reciprocal relationship between social and academic competence, such that they affect each other over time, has been observed in typically developing (TD) children, with potential implications for individuals with ASD (). This bidirectional theory suggests that children with social deficits often demonstrate inattention and distractibility in the school setting, contributing to academic difficulties and poor relationships with teachers and peers. Conversely, children with academic weaknesses may become frustrated in the classroom and engage in socially disruptive behaviors, causing peer rejection or stigma (Welsh et al., 2001). For children with ASD who by definition display social deficits, these social symptoms could contribute to academic challenges, which may then exacerbate further social difficulties with peers. Indeed, individuals with ASD commonly experience social and academic challenges within the school setting, often alongside cognitive, language, and motor impairments (). Despite difficulty engaging with the classroom environment, a large number of students with ASD are being mainstreamed as part of the inclusion model of special education, and teachers may be ill-equipped to meet their specific needs (; Jones et al., 2009).

Achievement Profiles in ASD

Relatively few studies have looked at academic attainment in individuals with ASD, and findings are variable (; ; Jones et al., 2009; ; ; ; ; ). Even within more hom*ogenous subgroups (e.g., high-functioning autism (HFA), Asperger’s Syndrome), achievement profiles are quite variable, with individuals’ scores ranging from impaired to above average across subtests (Griswold et al., 2002; ). Additionally, even when they display average overall achievement scores, groups of children with AS tend to have large standard deviations in specific areas, suggesting large group level variability within specific academic domains (Griswold et al., 2002).

Although the body of literature on academic functioning in ASD is relatively small, some work has been done on reading and mathematics abilities and their relationship to ASD symptoms and deficits (Jones et al., 2009; Nation et al., 2006).

Reading

Reading comprehension appears to be most consistently impaired among academic skills in ASD (; Nation et al., 2006; Ricketts et al., 2013; Troyb et al., 2014). Studies of decoding skills show mixed findings, with some reporting impairment in word recognition (Ricketts et al., 2013) and others demonstrating average to above average word reading ability (Brown et al., 2013; Minshew et al., 1994; Nation et al., 2006). Similar variability also applies to oral vocabulary, as some studies demonstrate unimpaired single word comprehension (Minshew et al., 1994) whereas other research reports relative weaknesses in basic semantic knowledge (Brown et al., 2013). For example, one study of reading skills in children with ASD showed that these individuals display word reading, nonsense word reading, and reading fluency scores within the average range, but show impaired comprehension (Nation et al., 2006). Specifically, 65 percent of the sample performed one standard deviation or more below average on a test of reading comprehension (i.e., Neale Analysis of Reading Ability – II, NARA-II), and one third of the sample showed very severe reading comprehension deficits. Additionally, in a study comparing older children and adolescents with HFA to TD peers and individuals with optimal outcome (OO) (i.e., individuals with a history of ASD who no longer meet diagnostic criteria, see Fein et al., 2013), individuals with HFA generally performed in the average range on tests of academic achievement, with a strength in reading accuracy, but with a significant relative weakness in reading comprehension (Troyb et al., 2014). Taken together, findings suggest that children and adolescents with ASD may be less impaired, or even within the average range, in rote or simple reading skills (e.g., word reading, nonsense word reading, single word comprehension), but they appear to perform poorly compared to TD peers on tests of comprehending stories or other complex text (Brown et al., 2013; Minshew et al., 1994). As suggested by Minshew et al. (1994), this pattern of performance indicates a deficit in more complex skills, particularly tasks requiring social knowledge, abstraction, inference making, integration of information, and working memory.

Mathematics

Individuals with ASD also demonstrate variable mathematics skills, with greater impairment in tasks requiring mathematical reasoning or inferential processing, rather than simple calculations (Minshew et al., 1994; Troyb et al., 2014). For example, Minshew et al. (1994) showed that adolescents with HFA display intact performance of numerical computations, providing evidence for an adequate ability to apply mechanical procedures to master rote academic tasks (e.g., basic math facts). However, individuals with ASD may have difficulty with tasks requiring problem solving abilities (; Troyb et al., 2014). Bae et al. (2015), in a study comparing school-age children with ASD to their TD peers, showed that individuals with ASD demonstrate impairment in mathematical word problem solving and application of mathematics to everyday situations. Although some individuals with AS or HFA demonstrate high average or superior mathematical abilities, a high proportion of children with HFA meet criteria for a learning disability in mathematics ().

Discrepancy from IQ

Academic performance in ASD may be discrepant from that predicted by a child’s overall cognitive ability. Within an HFA sample, Estes et al. (2011) showed that a majority of individuals demonstrated a discrepancy between IQ and academic achievement. Notably, 60 percent of children had low achievement in at least one academic domain, compared to IQ-based expectancy, yet 60 percent also had high achievement in at least one domain, suggesting that the typically strong association between achievement and IQ may be more complex in children with ASD. Additionally, the relationship between IQ and academic achievement in ASD may vary as a function of cognitive level (; ). Specifically, high functioning children with ASD (i.e., average to superior IQ) tend to display average academic achievement, indicating that they perform at or below their intellectual level on tests of academic skills, with a particular weakness in writing. Low functioning children with ASD (i.e., IQ < 80), though, appear to demonstrate math and spelling scores commensurate with IQ, and significantly higher than predicted reading decoding performance, consistent with a relative strength in rote learning even in low functioning children. Although there is evidence of academic performance discrepant from IQ in individuals with ASD, there is no single pattern of discrepancy, with some children outperforming their expected achievement level based on IQ, whereas others underperform on multiple achievement domains.

Predictors of Achievement in ASD

Very few studies have looked at early predictors of later achievement. Estes et al. (2011) explored specific factors associated with achievement outcomes in middle childhood; in their sample, social functioning at age six years was predictive of word reading ability at age nine years, after adjusting for age six nonverbal IQ. However, no published research has examined preschool predictors of academic functioning. Developmental impairments associated with ASD typically present in early childhood, with the average age of first diagnosis in the United States between two and four years (CDC, 2014). Research suggests that many children with ASD may make considerable gains in intellectual and social-behavioral functioning with early intervention services (Estes et al., 2011). By identifying preschool predictors of school-age achievement, it may be possible to further tailor early intervention and educational strategies to maximize later academic attainment in this population.

Overall, several themes have emerged from cross-sectional regression-based prediction models of reading and mathematics abilities in high functioning subgroups, with findings suggesting a role for IQ, basic cognitive processes (e.g., processing speed and working memory), executive functions (e.g., attentional switching), and motor skills in explaining the variance in academic achievement in individuals with ASD (; May et al., 2013; ).

Cognitive ability

Despite evidence of a discrepancy between IQ and academic achievement in some individuals with ASD, and in some academic areas specifically (e.g., poor writing, good word reading), global cognitive abilities are significantly related to level of achievement in this population. In children with HFA, concurrent full scale IQ (FSIQ) was the single best predictor of reading, math, and writing achievement (). Furthermore, using a cross-sectional predictor model in a sample of cognitively gifted children with ASD (i.e., IQ > 120), Assouline et al. (2012) found that working memory and processing speed accounted for significant variance (i.e., 61 percent) in reading achievement, working memory predicted written language achievement, and perceptual reasoning ability predicted oral language achievement. These findings suggest that basic domain general thinking skills (e.g., working memory and processing speed) also contribute to academic success in HFA.

Motor skills

An association between motor functioning and mathematical achievement has also been documented in TD children. A large-scale naturalistic observation study of TD toddlers aged 30 to 33 months revealed that, depending on level of fine and gross motor skills, toddlers significantly differed in their early math skills (i.e., counting, numerical series, shape and space, pattern and order, mathematical language, logical reasoning) (). Many researchers have noted impairments in graphom*otor (i.e., writing) skills and motor coordination in children with HFA and AS (; ), and these deficits appear to specifically affect mathematical abilities (). Scores on the Beery-Buktenica Developmental Test of Visual-Motor Integration (Beery VMI) and the Wechsler Block Design subtest predicted math achievement in gifted children with ASD (Assouline et al., 2012). Impaired performance on the skills needed for these measures, including fine motor skills, visual spatial ability, working memory, and speed of processing, may also affect performance on academic testing of math skills.

Social functioning

The unique impairments in social functioning associated with ASD also seem to contribute to specific difficulties in academic attainment, particularly within the domain of reading comprehension. For example, several studies have shown that the social impairments of ASD, as measured by tests of both social cognition and general social and communication functioning (i.e., Autism Diagnostic Observation Schedule, or ADOS), contribute unique variance to reading comprehension, even after accounting for vocabulary, decoding, oral language, and nonverbal IQ levels (; Ricketts et al., 2013). Furthermore, reading comprehension appears to fall increasingly below IQ with increasing social and communication impairments (Jones et al., 2009).

Although most previous research has focused on concurrent predictors of achievement in ASD, Estes et al. (2011) examined the association between early social skills at age six years and later reading abilities. After controlling for nonverbal IQ, better social functioning at age six years was significantly associated with greater academic achievement, particularly in the subdomain of word reading, at age nine years.

Taken together, these findings suggest that the generally poor reading comprehension abilities shown by individuals with ASD are not simply a product of weaknesses in core language skills (e.g., phonetic decoding, oral vocabulary). Thus, educational interventions aimed at improving these core language skills may not be sufficient to overcome reading comprehension deficits in the ASD population, and remediation of children’s social cognition and social behavior may be crucial in improving reading comprehension.

Current Study

Examination of academic achievement in children with ASD is limited and has focused almost entirely on concurrent relationships. Results of concurrent regression-based prediction models suggest that academic deficits generally appear to be linked to level of social impairment as well as performance in the domains of intellectual ability, working memory, processing speed, and motor skills. Furthermore, specific associations between social skills and reading comprehension, and motor skills and mathematics have been found. The main goals of the current study are to (1) characterize achievement profiles in a heterogeneous sample of school-age children with ASD, (2) test associations between academic achievement and concurrent intellectual functioning, communication and social skills, and ASD symptoms, and (3) explore potential preschool predictors of school-age achievement.

Methods

Participants

Participants were a subset of a sample from a larger federally funded project at the University of Connecticut. The aims of the original study focused on developing and validating the Modified Checklist for Autism in Toddlers (M-CHAT), a population-based screening measure used to detect ASD in young children (). Participants screened positive on the M-CHAT between the ages of 16 and 30 months and received a developmental and diagnostic evaluation at the approximate ages of two (i.e., Time 1, T1) and four (i.e., Time 2, T2) years. To have a well-characterized sample of children with ASD, participants were deemed eligible for the current study if they received a diagnosis on the autism spectrum (i.e., either AD, PDD-NOS, or AS) at both evaluations. They were then invited back for a follow-up evaluation in middle childhood. In the current study, only children who were between the ages of eight years and 10 years, 11 months were included. Exclusion criteria included sensory impairments (e.g., blindness) or severe deficits in motor functioning (e.g., severe cerebral palsy) that would negatively impact the participant’s ability to complete testing. Participants and their families were recruited either by telephone or letter.

One hundred two participants from the original M-CHAT study were deemed eligible for participation in the current study. Of those participants, 48 were contacted and given the opportunity to participate in the current study (47.1%); the remaining participants were unable to be contacted by either phone or mail. Of the 48 families, one moved (1%), seven declined participation (6.8%), and 40 were evaluated (39.2%).

During the Time 3 (T3) evaluation, a measure of academic achievement was given to a subset of children for whom time and cognitive level allowed such testing. Of the 40 children evaluated at T3, 27 completed at least some academic achievement testing. One participant was deemed an outlier for age at T2 (i.e., 8.1 years) and was excluded from the current study. The final sample contained 26 children with ASD, 22 males and four females. Caregivers self-identified their child’s race as White (n = 22), Black (n = 1), or Asian/Pacific Islander (n = 2); data were missing for one participant. At T1, 13 participants were diagnosed with AD, eight participants were diagnosed with PDD-NOS, and five participants were diagnosed with ASD and low mental age (ASD-LMA), a research diagnosis given to children exhibiting diagnostic features of PDD-NOS with severe cognitive impairment (i.e., nonverbal reasoning, receptive language, and expressive language scores at a 12-month level or lower). Mean age at initial evaluation was 27.3 months (SD = 3.4). Diagnosis was confirmed on follow-up at T2, when 22 participants were diagnosed with AD and four participants were diagnosed with PDD-NOS. Mean age at the time of re-evaluation was 52.9 months (SD = 5.7). At T3, diagnostic classification was based on the ADOS, with 23 participants continuing to meet criteria for AD and three participants falling within the non-spectrum score range. Mean age at T3 re-evaluation was 119.5 months (SD = 9.9), or nine years, 11 months. Participant characteristics are summarized in Table 1.

Table 1

Participant Characteristics

Variable	N	M	SD	Range
Cognitive constructs
T1 MSEL receptive language	24	37.7	17.2	12.0 – 84.9
T1 MSEL expressive language	24	40.6	16.8	10.5 – 77.8
T1 MSEL visual reception	23	61.8	18.3	19.2 – 92.0
T1 MSEL fine motor	24	68.5	13.2	44.2 – 95.2
T2 MSEL receptive language	22	53.8	27.1	19.5 – 118.4
T2 MSEL expressive language	23	51.9	26.7	15.7 – 89.5
T2 MSEL visual reception	23	64.8	31.0	24.8 – 131.0
T2 MSEL fine motor	23	61.1	22.5	17.3 – 93.3
T3 DAS-II GCA	24	80.0	28.7	25 – 130
Adaptive constructs
T1 VABS communication	26	63.9	5.5	55 – 78
T1 VABS socialization	26	66.4	6.3	56 – 77
T2 VABS communication	26	63.6	15.1	47 – 93
T2 VABS socialization	26	61.6	10.3	51 – 90
T3 VABS-II communication	25	75.2	15.6	53 – 108
T3 VABS-II socialization	25	70.1	17.6	45 – 129
ASD symptom constructs
T1 ADOS CSS	25	7.0	1.9	4 – 10
T1 CARS total	25	33.3	3.8	27 – 40
T2 ADOS CSS	26	6.7	1.6	3 – 10
T2 CARS total	25	32.7	4.9	24.5 – 44.5
T3 ADOS CSS	26	6.9	2.1	1 – 10
T3 CARS total	23	30.3	6.0	19 – 41
Academic achievement constructs
T3 WIAT-II word reading	26	87.7	25.7	40 – 121
T3 WIAT-II reading comprehension	21	74.3	26.7	40 – 120
T3 WIAT-II numerical operations	22	85.1	30.3	40 – 147
T3 WIAT-II math reasoning	20	78.6	34.0	40 – 140

Open in a separate window

Note. MSEL = Mullen Scales of Early Learning; DAS-II = Differential Ability Scales, Second Edition; VABS = Vineland Adaptive Behavior Scales; VABS-II = VABS, Second Edition; ADOS = Autism Diagnostic Observation Schedule – Generic; CARS = Childhood Autism Rating Scale; WIAT-II = Wechsler Individual Achievement Test, Second Edition. Means and standard deviations presented are based on MSEL developmental quotient (DQ) scores (M = 100, SD = 15); DAS-II, VABS, VABS-II, and WIAT-II standard scores (M = 100, SD = 15); ADOS calibrated severity scores (CSS) (non-spectrum = 1 – 3, ASD = 4 – 5, AD = 6 – 10); and CARS total scores (ASD cut-off = 25.5).

Procedures

All preschool evaluations were completed at the University of Connecticut Psychological Services Clinic. Participants who did not have transportation were provided with a free taxi service. Initial evaluations (T1) occurred when participants were between the ages of 20 and 33 months, and all children were invited back for a second evaluation (T2) when they were between the ages of 43 and 71 months. A team of clinicians, consisting of one licensed clinical psychologist or developmental-behavioral pediatrician and one clinical psychology doctoral student, completed each evaluation. All evaluations lasted approximately three hours, including a feedback session in which diagnosis and recommendations were reviewed with the child’s caregiver. Given the years in which T1 and T2 evaluations were conducted (i.e., between 2001 and 2008), all study diagnoses were assigned according to DSM-IV-TR criteria (APA, 2000), based on clinical best estimate judgment of symptoms from observation, developmental history, and testing data (i.e., performance on ADOS, CARS, and MSEL).

Children diagnosed with ASD at both preschool evaluations (T1 and T2), who were between the ages of eight years and 10 years, 11 months (T3), were recruited for the current study by letter or telephone contact. Participants were offered a free developmental evaluation that was completed, either at the University of Connecticut Psychological Services Clinic or in participants’ homes, by two clinical psychology doctoral students under the supervision of a licensed clinical psychologist. Evaluations were video-recorded and lasted up to four hours depending on the child’s abilities and tolerance.

Measures

Times 1 and 2

Cognitive ability

The Mullen Scales of Early Learning (MSEL; Mullen, 1995) is a developmental assessment of cognitive, motor, and language abilities in children aged one month to five years, eight months. It was used to assess cognitive ability at both preschool evaluations. The current study used scores in the domains of visual reception, fine motor, and receptive and expressive language. Histograms revealed that MSEL T scores in each domain were not normally distributed because a large number of children received the lowest possible standard score. In order to use parametric statistical tests without violating their assumptions, age-equivalent (AE) scores were converted to developmental quotient (DQ) scores according to the formula mental age (i.e., AE scores) divided by chronological age, multiplied by 100 (Reitzel et al., 2013).

Adaptive functioning

The Vineland Adaptive Behavior Scales: Interview Edition, Survey Form (VABS; ) is a semi-structured caregiver interview that assesses adaptive behaviors (i.e., how a child functions in his or her daily life) in the domains of socialization, communication, daily living, and motor skills. The VABS was used to quantify adaptive skills at both T1 and T2. Standard scores in the domains of communication and socialization were used in the current study.

ASD symptoms

The Autism Diagnostic Observation Schedule – Generic (ADOS; Lord et al., 2000) is a semi-structured observational assessment designed to measure symptoms of ASD in toddlerhood through adulthood. The ADOS includes four separate modules based on a participant’s expressive language level and chronological age. The current study used Modules 1 or 2 during preschool evaluations. To allow for comparison of ASD symptom severity across different ADOS modules and evaluation time points, raw scores were converted into calibrated severity scores (CSS) based on a published algorithm ().

The Childhood Autism Rating Scale (CARS; ) is a 15-item clinician rating scale measuring autism symptom severity based on observation and caregiver report, used to quantify ASD severity at both T1 and T2. A total score is calculated by summing scores from all individual items and classifies a child into one of three groups: non-autistic (total score = 15 – 30), mildly-moderately autistic (total score = 30 – 37), and severely autistic (total score = 37 – 60). Although a cut-off of 30 is typically used for AD, a cut-off of 25.5 has been proposed for ASD more broadly ().

Time 3

Cognitive ability

The Differential Ability Scales, Second Edition (DAS-II; Elliott, 2007) is a measure of cognitive ability in children between the ages of two years, six months and 17 years, 11 months. The DAS-II School-Age Cognitive Battery was used for the majority of participants in the current study based on their chronological age. However, nine participants were administered the DAS-II Early Years Cognitive Battery based on their cognitive level. The DAS-II provides extended General Conceptual Ability (GCA) standard scores, which were used in the current study.

Adaptive functioning

The Vineland Adaptive Behavior Scales, Second Edition: Survey Interview Form (VABS-II; ) was used to assess adaptive functioning during the school-age evaluation. The current study used standard scores in the domains of communication and socialization.

ASD symptoms

The Autism Diagnostic Observation Schedule – Generic (ADOS; Lord et al., 2000) and the Childhood Autism Rating Scale (CARS; ) were used to measure autism-specific symptomatology at T3 in an effort to maintain diagnostic measure consistency across all three evaluation time points. Children were administered either ADOS Modules 2 or 3 based on level of expressive language; raw scores were converted into CSS to allow for comparison across modules and time. CARS total scores were also used to quantify ASD severity at ages eight to 10 years.

Academic achievement

The Wechsler Individual Achievement Test, Second Edition (WIAT-II; Wechsler, 2001) is a structured measure of academic achievement in children and adults between the ages of four and 85 years. Standard scores for the specific subtests of word reading, reading comprehension, numerical operations, and math reasoning were used in the current study.

See Table 2 for a summary of study measures by evaluation time point and construct.

Table 2

Measures by Evaluation Time Point and Construct

Construct	Evaluation Time Point
Construct	Times 1 and 2	Time 3
Cognitive ability	MSEL	DAS-II
Adaptive functioning	VABS	VABS-II
ASD symptoms	ADOS	ADOS
	CARS	CARS
Academic achievement	-	WIAT-II

Open in a separate window

Data Analytic Plan

Achievement profile analyses

To determine whether school-age individuals with ASD demonstrated greater impairment in academic skills requiring abstraction and inferential processing, a series of paired-samples t-tests were used to compare performance on specific subtests in the domains of reading (i.e., word reading versus reading comprehension) and mathematics (i.e., numerical operations versus math reasoning).

Predictor analyses

Consistent with the aims of the study, multiple regression analyses were conducted to determine whether cognitive abilities, adaptive skills, and ASD symptoms seen during both preschool and school-age years accounted for a significant amount of variance in school-age academic achievement.

Data preparation

All data were analyzed to determine if assumptions of linear regression were violated. Based on non-normal distributions of error terms, scores were adjusted using square (for word reading) and natural log (for reading comprehension) transformations. Since no transformations of math reasoning scores satisfactorily addressed these concerns, a bootstrapping method with 500 replications was employed to estimate the true relationships between predictors and math reasoning scores. All analyses were run using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corporation, 2013) with the exception of bootstrapped regression analyses, which were conducted using Stata 14 (StataCorp, 2015).

Variable selection process

Multiple regression analysis was used to examine the possible variables that contribute to variation in academic achievement. For each regression, assumptions of collinearity were assessed through the evaluation of VIF and tolerance statistics. Conservative cut-offs of variance inflation factor (VIF) > 4 and tolerance < .20 were used, as described in Menard (1995). Due to consistent evidence of multicollinearity, MSEL expressive language and receptive language developmental quotients were averaged into a single construct (i.e., MSEL verbal ability), which was tested in subsequent T1 and T2 regression analyses. Where multicollinearity arose in other models, a strategy of evaluating reduced model subsets was successful in assessing the strength of predictors while managing threats from multicollinearity.

Because of the large number of constructs collected at T1 and T2, separate initial regression models were carried out with each of the four WIAT-II subdomains as dependent variables for each set of predictor variables, including cognitive constructs (i.e., MSEL verbal ability, visual reception, and fine motor DQ scores), adaptive constructs (i.e., VABS communication and socialization standard scores), and ASD symptom constructs (i.e., ADOS CSS, CARS total scores), for both preschool time points. A similar process was used to select predictor variables for concurrent regression models. Separate initial regression models were carried out with each of the four WIAT-II subdomains as dependent variables for each set of predictor variables, including adaptive constructs (i.e., VABS-II communication and socialization standard scores) and ASD symptom constructs (i.e., ADOS CSS, CARS total scores) collected at T3. Those variables identified within the initial regression models as the strongest, or as theoretically significant, were retained and tested together as part of final models, as reported in Results. Any significant analyses were then run in an additional regression model adding T3 IQ (i.e., DAS-II GCA) in order to determine the amount of variance in academic achievement accounted for by early and concurrent predictors above and beyond a child’s level of intellectual functioning in middle childhood.

In initial regression analyses, an alpha level of .1 was adopted to identify variables to be retained for subsequent analyses. An alpha level of .05 was adopted for statistical tests of final models. The only exception to this approach was for models in which MSEL fine motor skills were examined. Due to literature suggesting a link between fine motor ability and mathematical skills, an a priori decision was made to retain MSEL fine motor skills as a predictor throughout the variable selection process despite any evidence of initial non-significance.

Results

Differential Attrition

Analyses of variance revealed no significant differences on gender, race, cognitive ability, adaptive functioning, or ASD symptom severity at T1 and T2 evaluations between children who participated in T3 testing and those who did not (Knoch, 2014; Troyb et al., 2016).

Aim One: Achievement Profiles

A paired-samples t-test was conducted to compare performance on WIAT-II subtests of word reading and reading comprehension. There was a significant difference between scores on word reading (M = 87.6, SD = 25.3) and reading comprehension (M = 74.3, SD = 26.7), t(20) = 2.829, p = .010. A second paired-samples t-test was run to compare achievement on WIAT-II subtests of numerical operations and math reasoning. No significant difference was found between numerical operations (M = 83.7, SD = 30.3) and math reasoning (M = 78.6, SD = 34.0) scores, t(19) = 1.666, p = .112, although the difference was in the predicted direction and close to trend level significance. The absolute difference between math subtests was five standard score points, whereas the difference between reading subtests was 13 points.

Aim Two: Concurrent Predictors of Achievement

Reading

We examined the impact of T3 functional communication skills and ASD severity on concurrent word reading (see Table 3) and reading comprehension (see Table 4) abilities. Results indicated that higher VABS-II communication scores predicted higher concurrent WIAT-II word reading and reading comprehension scores, and higher CARS total scores (i.e., greater ASD severity) predicted lower WIAT-II reading comprehension scores.

Table 3

Summary of Concurrent Multiple Regression Analyses Predicting Word Reading

Variable	Model 1a			Model 1b

	B	SE_B	β	B	SE_B	β
T3 VABS-II communication	173.9	51.9	.729^**	94.3	55.3	.404
T3 CARS total	22.8	141.4	.035	96.2	138.3	.159
T3 DAS-II GCA				84.1	32.9	.561^**
R²	.498			.599
F	9.425^**			7.956^**

Open in a separate window

Note. B = unstandardized coefficient, β = standardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). VABS-II = Vineland Adaptive Behavior Scales, Second Edition; CARS = Childhood Autism Rating Scale; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability.

N_{Model 1a} = 21; N_{Model 1b} = 19

^*p < .05;

^**p < .01

Table 4

Summary of Concurrent Multiple Regression Analyses Predicting Reading Comprehension

Variable	Model 2a			Model 2b

	B	SE_B	β	B	SE_B	β
T3 VABS-II communication	.010	.004	.443^*	.008	.005	.341
T3 CARS total	−.025	.011	−.451^*	−.019	.012	−.337
T3 DAS-II GCA				.004	.003	.302
R²	.664			.712
F	14.801^**			10.705^**

Open in a separate window

N_{Model 2a} = 17; N_{Model 2b} = 16

^*p < .05;

^**p < .01

Mathematics

Results of concurrent regression analyses predicting numerical operations ability from T3 ASD severity are summarized in Table 5. Higher T3 CARS total scores predicted lower WIAT-II numerical operations scores. We then assessed the impact of school-age communication and social skills on concurrent math reasoning ability (see Table 6). More severe ASD symptomatology predicted lower WIAT-II math reasoning scores.

Table 5

Summary of Concurrent Multiple Regression Analyses Predicting Numerical Operations

Variable	Model 3a			Model 3b

	B	SE_B	β	B	SE_B	β
T3 CARS total	−2.640	.923	−.559^*	−.379	.793	−.079
T3 DAS-II GCA				.969	.200	.799^**
R²	.312			.720
F	8.180^*			20.619^**

Open in a separate window

Note. B = unstandardized coefficient, β = standardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). CARS = Childhood Autism Rating Scale; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability.

N_{Model 3a} = 19; N_{Model 3b} = 18

^*p < .05;

^**p < .01

Table 6

Summary of Concurrent Multiple Regression Analyses Predicting Math Reasoning

Variable	Model 4a		Model 4b

	B	SE_B	B	SE_B
T3 VABS-II communication	.595	.695	−.126	.532
T3 CARS total	−2.804^*	1.230	−1.220	1.114
T3 DAS-II GCA			1.049^**	.331
R²	.454		.807
Wald Χ²	13.43^**		36.66^**

Open in a separate window

Note. B = unstandardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). VABS-II = Vineland Adaptive Behavior Scales, Second Edition; CARS = Childhood Autism Rating Scale; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability.

N_{Model 4a} = 17; N_{Model 4b} = 16

^*p < .05;

^**p < .01

For each of the models described above, we ran additional multiple regression models adding T3 DAS-II GCA. Concurrent IQ was the best individual predictor of academic achievement in all cases except for reading comprehension ability, for which no individual predictor variable contributed significantly to the multiple regression model (i.e., Model 2b; see Table 4). These findings suggest that some of the variance in word reading, numerical operations, and math reasoning scores initially explained by communication skills and level of ASD severity was likely accounted for by variability in concurrent IQ.

Aim Three: Preschool Predictors of Achievement

Reading

Results of initial regression models predicting word reading from T1 cognitive, adaptive, and ASD symptom constructs did not reveal any significant predictors; thus, a final regression model was not tested. We explored the impact of T2 (i.e., four-year-old) communication skills and ASD severity on school-age word reading ability (see Table 7). Taken together, both variables accounted for significant variance in T3 word reading ability, but neither individual predictor significantly contributed to the multiple regression model.

Table 7

Summary of Preschool Multiple Regression Analyses Predicting Word Reading

Variable	Model 5a			Model 5b

	B	SE_B	β	B	SE_B	β
T2 VABS communication	92.2	64.8	.360	14.4	65.2	.060
T2 CARS total	−207.9	203.0	−.259	−82.7	181.1	−.112
T3 DAS-II GCA				78.6	32.2	.583^*
R²	.332			.498
F	5.475^*			6.292^**

Open in a separate window

N_{Model 5a} = 24; N_{Model 5b} = 22

^*p < .05;

^**p < .01

Regression analyses predicting reading comprehension ability from preschool variables are summarized in Table 8. Greater verbal ability at T1, as measured by the MSEL, predicted higher WIAT-II reading comprehension scores at T3. The inclusion of MSEL verbal ability and VABS communication as predictors in a model resulted in multicollinearity concerns (VIF > 16). As a result, these predictors were separately analyzed as part of final regression models, along with other variables deemed significant in initial regression analyses (see Table 8). Higher communication abilities at age four years, as measured by both T2 MSEL verbal ability and T2 VABS communication scores, predicted higher WIAT-II reading comprehension scores at school age.

Table 8

Summary of Preschool Multiple Regression Analyses Predicting Reading Comprehension

Variable	Model 6a			Model 6b

	B	SE_B	β	B	SE_B	β
T1 MSEL verbal ability	.014	.004	.605^**	.005	.004	.230
T3 DAS-II GCA				.009	.002	.660^**
R²	.366			.657
F	9.822^**			14.385^**

Variable	Model 7a			Model 7b

	B	SE_B	β	B	SE_B	β

T2 MSEL verbal ability	.008	.003	.551^*	.005	.004	.339
T2 CARS total	−.032	.021	−.340	−.031	.023	−.339
T3 DAS-II GCA				.004	.003	.279
R²	.714			.741
F	17.474^**			11.442^**

Variable	Model 7c			Model 7d

	B	SE_B	β	B	SE_B	β

T2 VABS communication	.017	.004	.757^**	.013	.005	.575^*
T2 CARS total	−.009	.013	−.132	−.007	.013	−.102
T3 DAS-II GCA				.004	.002	.274
R²	.731			.761
F	23.142^**			15.890^**

Open in a separate window

Note. B = unstandardized coefficient, β = standardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). MSEL = Mullen Scales of Early Learning; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability; CARS = Childhood Autism Rating Scale; VABS = Vineland Adaptive Behavior Scales.

N_{Model 6a} = 18; N_{Model 6b} = 17; N_{Model 7a} = 16; N_{Model 7b} = 15; N_{Model 7c} = 19; N_{Model 7d} = 18

^*p < .05;

^**p < .01

Mathematics

We examined the impact of preschool fine motor skills and level of ASD severity on school-age numerical operations ability (see Table 9). The initial model predicting numerical operations from T2 MSEL variables (i.e., fine motor, visual reception, and verbal ability) revealed evidence of multicollinearity (VIF = 4.2). Independent examination of each predictor found that each was significantly predictive of numerical operations scores. Without the ability to evaluate the relative strength of these predictors and no other statistical rationale to guide selection, we decided, based on our a priori interest in fine motor skills, to retain this variable in subsequent models. It must be noted that the effect of this construct in predicting numerical operations ability may be due to some specific aspect of fine motor skills or to some general aspect it shares with other MSEL constructs.

Table 9

Summary of Preschool Multiple Regression Analyses Predicting Numerical Operations

Variable	Model 8
Variable	B		SE_B		β
T1 MSEL fine motor	.942		.541		.391
T1 CARS total	−1.9		2.0		−.213
R²	.234
F	2.450

Variable	Model 9a			Model 9b

	B	SE_B	β	B	SE_B	β

T2 MSEL fine motor	.829	.312	.636^*	−.069	.371	−.052
T2 CARS total	−.533	1.8	−.071	.135	1.4	.018
T3 DAS-II GCA				.981	.288	.895^**
R²	.469			.709
F	7.057^**			11.372^**

Open in a separate window

Note. B = unstandardized coefficient, β = standardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). MSEL = Mullen Scales of Early Learning; CARS = Childhood Autism Rating Scale; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability.

N_{Model 8} = 18; N_{Model 9a} = 18; N_{Model 9b} = 17

^*p < .05;

^**p < .01

As shown in Table 9, T1 predictors did not account for significant variance in T3 numerical operations ability. Better fine motor abilities at age four years (i.e., T2), though, predicted higher WIAT-II numerical operations scores between the ages of eight and 10 years (i.e., T3).

We then assessed the impact of preschool cognitive, adaptive, and ASD symptom constructs on school-age math reasoning ability. Results are summarized in Table 10. Better fine motor skills at age two years predicted higher WIAT-II math reasoning scores at school age. Taken together, T2 MSEL fine motor scores, T2 VABS communication scores, and T2 CARS total scores accounted for significant variance in math reasoning; however, no individual predictor significantly contributed to the regression model (i.e., Model 11a).

Table 10

Summary of Preschool Multiple Regression Analyses Predicting Math Reasoning

Variable	Model 10a		Model 10b

	B	SE_B	B	SE_B
T1 MSEL fine motor	1.431^*	.625	.100	.494
T1 CARS total	−1.425	2.389	.524	1.787
T3 DAS-II GCA			1.040^**	.247
R²	.336		.772
Wald Χ²	8.25^*		41.09^**

Variable	Model 11a		Model 11b

	B	SE_B	B	SE_B

T2 MSEL fine motor	.933	.556	−.173	.405
T2 VABS communication	−.028	1.227	−.867	1.159
T2 CARS total	−1.393	2.759	−2.263	2.665
T3 DAS-II GCA			1.261^**	.509
R²	.512		.785
Wald Χ²	22.33^**		18.10^**

Open in a separate window

Note. B = unstandardized coefficient, R² = coefficient of determination (i.e., proportion of variance in dependent variable predictable from independent variables). MSEL = Mullen Scales of Early Learning; CARS = Childhood Autism Rating Scale; DAS-II = Differential Ability Scales, Second Edition; GCA = General Conceptual Ability; VABS = Vineland Adaptive Behavior Scales.

N_{Model 10a} = 17; N_{Model 10b} = 16; N_{Model 11a} = 17; N_{Model 11b} = 16

^*p < .05;

^**p < .01

For each of the models described above, we ran additional multiple regression models adding T3 DAS-II GCA, which was the strongest individual predictor of school-age academic achievement in all cases except for reading comprehension ability, which, of note, was best explained by T2 VABS communication skills (i.e., Model 7d; see Table 8). That is, although much of the variance in school-age academic attainment initially explained by preschool constructs appeared to be subsumed by concurrent intellectual functioning, four-year-old functional communication skills predicted school-age reading comprehension ability above and beyond T3 IQ.

Supplementary Analyses

Due to concerns regarding the representativeness of our sample, we re-analyzed all data excluding the three participants who fell in the non-spectrum range on the ADOS at T3. There were no meaningful changes to the reported results.

Discussion

The purpose of the current study was threefold: first, to characterize achievement profiles in the domains of reading and mathematics; second, to replicate associations between cognitive, adaptive, and social functioning and concurrent academic abilities seen in the existing literature; and third, to examine potential preschool predictors of school-age academic achievement. Prior research has implicated intellectual capacity, social skills, and motor functioning in the prediction of academic ability, largely in hom*ogenous subgroups of children with ASD, including those with AS and HFA. Furthermore, the unique constellation of symptoms associated with ASD appears to put children at greater risk for academic weaknesses, particularly in domains requiring inferential processing and abstraction (e.g., reading comprehension, math problem solving), as hypothesized by Minshew and Goldstein (1998). ASD symptoms in the areas of cognition, communication, and socialization generally present in the preschool period and negatively impact the child’s ability to engage in learning opportunities, thus exacerbating achievement deficits. Knowing that early functioning predicts subsequent academic attainment would allow professionals who work with young children with ASD to both help families form appropriate expectations for their child’s educational future and, perhaps more significantly, target early intervention strategies to potentially maximize later academic success in this population.

As expected, the results of the current study showed that individuals with ASD, even within a heterogeneous sample (i.e., including both low- and high-functioning children), demonstrate greater impairment in reading comprehension compared to word reading ability. This is consistent with the existing body of literature, which suggests that deficits in reading comprehension shown in many individuals with ASD are not simply reflective of deficits in decoding, as this domain appears largely intact (i.e., within the broadly average range). Instead, impaired reading comprehension is likely a product of more general impairments in linguistic processing and social understanding (Jones et al., 2009; ; Nation et al., 2006). Skilled reading and reading comprehension requires skills in word recognition, phonetic decoding, receptive language (i.e., the ability to understand verbal material), and processing of more abstract and inferential material. Individuals with ASD often demonstrate an impaired ability to integrate meaningful information, focusing on details instead (; ; Kanner, 1943). This may underlie deficits in the ability to make inferences, as well as poor social cognition (Ricketts et al., 2013). Together, these deficits likely contribute to impaired skills necessary for skilled reading comprehension (Ricketts et al., 2013). These reading findings are consistent with the general language profile of ASD, in which individuals show difficulty with comprehension of complex verbal material, pragmatic language, verbal reasoning and abstraction, and non-literal language (Tager-Flusberg, 1996). They may thus have difficulty inferring intentions of characters and causal connections between story events, contributing to impaired reading comprehension (Troyb et al., 2014).

Although there are fewer studies on math skills than on reading ability in individuals with ASD, research suggests that academic domains requiring problem solving, such as math reasoning, are generally more impaired than rote numerical skills (Bae et al., 2015; Minshew et al., 1994; Troyb et al., 2014). Despite failing to replicate this finding in the current study, overall, children’s WIAT-II scores did trend in the direction of weaker performance on subtests requiring reasoning and abstraction. As seen in Table 1, WIAT-II word reading and numerical operations scores fell within the low average range, whereas reading comprehension and math reasoning scores fell within the borderline classification, indicating greater impairment. Math reasoning may be more impaired than numerical computation ability in individuals with ASD because of the greater need to use a ‘linguistic, varied, and conceptually demanding’ approach to problem solving (Jones et al., 2009, p. 726). As such, it is likely that math deficits in ASD reflect an underlying difficulty with comprehension of linguistic information and inferential processing.

With respect to the second aim of the current study, which sought to confirm associations between cognitive and social functioning and concurrent academic attainment, results were largely consistent with hypotheses, in that level of intellectual functioning and, to an extent, social communication skills, at ages eight to 10 years significantly predicted concurrent academic achievement, especially in word reading and reading comprehension. Participants with greater ASD severity tended to have lower academic abilities, particularly in reading comprehension, numerical operations, and math reasoning. However, overall, concurrent IQ accounted for significant variance in academic achievement (i.e., in word reading, numerical operations, and math reasoning abilities) above and beyond that explained by adaptive and ASD symptom constructs. This finding is largely consistent with previous literature suggesting that FSIQ is the best predictor of concurrent academic ability.

Overall, the results of preschool regression analyses suggested that early cognitive, adaptive, and social functioning accounts for a significant proportion of variance (i.e., 33 to 73 percent) in school-age academic ability. Despite this, exploration of regression weights indicated that it may be difficult to specifically predict school-age achievement, particularly within the domains of word reading, numerical operations, and reading comprehension, from functioning at ages two and four years. Moreover, it appears that the variance in school-age academic ability explained by preschool cognitive, adaptive, and ASD symptom constructs may be largely subsumed by school-age IQ. However, several interesting variables emerged as significant in preschool prediction analyses, with potentially important implications for early intervention in ASD. The fact that these variables were largely subsumed under school-age IQ is in line with Dennis et al.’s (2009) argument that FSIQ combines ‘genetic, biological, neural, cognitive, educational, and experiential’ constructs, thereby necessarily subsuming heterogeneous and often relevant variance (p. 341). To use IQ as a covariate, or to adjust for IQ in a regression model, then, may lead to an underestimation of the contributions of specific cognitive variables that go into both IQ and the outcome of interest (Dennis et al., 2009); as such, we believe it is valuable to interpret notable preschool variables, regardless of the impact of school-age IQ on the statistical significance of findings.

Several important themes, particularly relating to early language and motor functioning, emerged from preschool prediction analyses. Preschool verbal abilities significantly predicted reading comprehension ability at T3. Of particular note, functional communication skills at age four years accounted for significant variance in school-age reading comprehension, even after controlling for T3 cognitive level. Contrary to our hypotheses, though, social skills and level of ASD symptomatology seen at ages two and four years did not significantly predict later academic achievement. Although the observed associations diminished once T3 IQ was added to regression models, results also supported the reported link between preschool motor functioning and math reasoning ability at school age. Four-year-old fine motor abilities predicted school-age numerical operations scores, although multicollinearity problems limited the ability to confirm this association as specific to fine motor skills rather than to global cognitive functioning.

One contributor to the association between motor and academic functioning might be the fact that many math-based academic tasks involve the use of fine motor skills, especially written computations (i.e., WIAT-II numerical operations), which require graphom*otor skills as well as hand-eye coordination to grasp a writing utensil and form letters and numbers correctly and in sequence. Mathematical problem solving tasks (i.e., WIAT-II math reasoning), too, generally involve tracking of visual information, which requires fine motor control of eye movements. However, the motor-achievement linkage appears to be much more complex, with both skills potentially relying on a shared manipulation of internal neural representations (; Ito, 2005). Ito (2005) suggests that individuals use a model of the body in the external environment to predict and execute desired motor actions; similarly, for certain academic skills, particularly those relating to mathematical ability, the internal representation may be abstract symbols which are then manipulated toward a solution. As infants and toddlers motorically explore their environment and develop early motor skills, it is likely that they build a neural infrastructure that later allows them to execute simple and complex cognitive and academic tasks (Grissmer et al., 2010). Thus, foundational motor skills acquired early in development appear to be necessary for successful attainment of mathematical abilities by middle childhood.

Taken together, the results of the current study have implications for early intervention and educational strategies in children with ASD. Individuals with ASD may benefit from rote learning, repetition, visual representations, and hands-on instruction to build basic academic skills (i.e., decoding, vocabulary, arithmetic), whereas more specialized instruction in inferential and pragmatic language use might encourage development of weaker areas (i.e., reading comprehension). Furthermore, results of preschool predictor analyses suggest that it may be possible to meaningfully impact later academic functioning by intervening on language and motor skills in the preschool period.

Limitations and Future Directions

Due to several limitations, the results of the current study should be generalized with some caution, pending further replication. Most significantly, the sample on which these findings were based was small, and a power analysis revealed that only large effects would likely be detected. Additionally, due to the small sample size, individual cases may have had a greater impact on study results, contributing to potentially biased findings. A related concern is that the number of comparisons that were explored in these analyses raised the possibility of a Type I error, or falsely identifying as significant a predictor that was not truly significant. In many analyses, it would be advisable to make a correction for multiple comparisons to reduce this possibility. However, doing so raises the possibility of making a Type II error, or failing to identify a true relationship when one is present. Given that our intention was to explore potentially important predictors rather than to confirm any established candidates, we have opted not to correct for multiple comparisons. It is thus important to keep in mind the possibility that significant predictors here may be spurious.

To address these concerns, although it would be difficult to collect a large sample of children who were diagnosed with ASD early in development and subsequently followed and re-evaluated throughout preschool and middle childhood, doing so would allow for more definitive and specific conclusions about the ability to predict school-age academic functioning from cognitive, adaptive, and social constructs seen in the preschool period. Furthermore, it is possible that a larger sample would allow the math reasoning-numerical operations difference to reach statistical significance.

Because of the recruitment area of the current study, this sample was predominately White (85 percent). Therefore, the results of the present study may not generalize to a broader ASD population. Future studies should attempt to replicate these findings in a more racially and geographically diverse sample. The presence of only four female children in the current sample also limits generalization to females with ASD, who may have academic achievement profiles and patterns of skill acquisition different from those of male children. Furthermore, although the current study contributes to the literature on academic achievement in ASD by including a functionally diverse sample of children with ASD, this heterogeneity, particularly within a small sample size, limited our ability to form definitive conclusions about within-group differences across academic skills and prevented exploration of potential patterns of impairment and prediction unique to specific subgroups, such as high- and low-functioning ASD.

Further, despite contributing to our understanding of the development of academic ability in ASD, predictors in the current study were limited to variables measuring cognitive, language, and motor skills and social functioning. Other domains, particularly executive functioning (EF), have been implicated in academic achievement in TD children (), as well as those with ASD (Pellicano, 2012). Because EF is difficult to measure in the preschool period, this domain was left out of the current study. Future research might explore the predictive value of EF, particularly working memory, in explaining academic attainment in individuals with ASD.

Finally, because of the original goal of the larger study (i.e., to validate an ASD-specific screening tool) of which the current study was one component, there was no readily available comparison group of TD individuals. Thus, the results of the present study may not be unique to individuals with ASD. Future studies should also investigate the development of academic abilities in other groups, including TD children and children with other neurodevelopmental disorders, to determine the extent to which specific patterns of early prediction observed in the present study are unique to ASD.

To our knowledge, this is the first study to examine preschool predictors of academic achievement in a well-characterized sample of children with a history of ASD. Future research should build on current study findings by assessing the impact of early intervention, particularly services targeting language and motor skills, on later academic ability in the domains of reading and mathematics. As the educational climate for children with ASD continues to shift, with more individuals being placed in mainstream, or general education, classrooms with their TD peers, it is increasingly necessary to improve our understanding of achievement in individuals with ASD and, more importantly, to develop and refine strategies to encourage academic success in this population.

Acknowledgments

This work was supported by the Maternal and Child Health Bureau under Grant R40MC00270-04-00 and by the Eunice Kennedy Shriver National Institute of Child Health and Human Development under Grant 5 R01 HD039961-05.

Footnotes

Disclosure Statement: Deborah Fein is part owner of the M-CHAT-R LLC. The M-CHAT-R is available free of charge for providers’ use but provides royalties from companies that incorporate it into commercial products and charge for its use. No other authors report conflicts of interest.

References

American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4. Washington, DC: Author; 2000. text rev. [Google Scholar]
American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5. Washington, DC: Author; 2013. [Google Scholar]
Asberg J, Kopp S, Berg-Kelly K, Gillberg C. Reading comprehension, word decoding and spelling in girls with autism spectrum disorders (ASD) or attention-deficit/hyperactivity disorder (AD/HD): Performance and predictors. International Journal of Language & Communication Disorders. 2010;45(1):61–71. http://doi.org/10.3109/13682820902745438. [PubMed] [Google Scholar]
Assouline SG, Nicpon MF, Dockery L. Predicting the academic achievement of gifted students with autism spectrum disorder. Journal of Autism and Developmental Disorders. 2012;42:1781–1789. http://doi.org/10.1007/s10803-011-1403-x. [PubMed] [Google Scholar]
Bae YS, Chiang HM, Hickson L. Mathematical word problem solving ability of children with autism spectrum disorder and their typically developing peers. Journal of Autism and Developmental Disorders. 2015;45:2200–2208. http://doi.org/10.1007/s10803-015-2387-8. [PubMed] [Google Scholar]
Best JR, Miller PH, Naglieri JA. Relations between executive function and academic achievement from ages 5 to 17 in a large, representative national sample. Learning and Individual Differences. 2011;21:327–336. [PMC free article] [PubMed] [Google Scholar]
Brown HM, Oram-Cardy J, Johnson A. A meta-analysis of the reading comprehension skills of individuals on the autism spectrum. Journal of Autism and Developmental Disorders. 2013;43:932–955. http://doi.org/10.1007/s10803-012-1638-1. [PubMed] [Google Scholar]
Centers for Disease Control and Prevention. March 28th, Morbidity and Mortality Weekly Report Surveillance Summaries. 2014. Prevalence of autism spectrum disorder among children aged 8 years - Autism and developmental disabilities monitoring network, 11 sites, United States, 2010. [PubMed] [Google Scholar]
Chiang HM, Lin YH. Mathematical ability of students with asperger syndrome and high-functioning autism. Autism. 2007;11(6):547–556. http://doi.org/10.1177/1362361307083259. [PubMed] [Google Scholar]
Chlebowski C, Green JA, Barton ML, Fein D. Using the Childhood Autism Rating Scale to diagnose autism spectrum disorders. Journal of Autism and Developmental Disorders. 2010;40(7):787–799. [PMC free article] [PubMed] [Google Scholar]
Dennis M, Francis DJ, Cirino PT, Schachar R, Barnes MA, Fletcher JM. Why IQ is not a covariate in cognitive studies of neurodevelopmental disorders. Journal of the International Neuropsychological Society. 2009;15:331–343. [PMC free article] [PubMed] [Google Scholar]
Elliott CD. Differential ability scales. 2. San Antonio, TX: Pearson Education Inc; 2007. [Google Scholar]
Estes A, Rivera V, Bryan M, Cali P, Dawson G. Discrepancies between academic achievement and intellectual ability in higher-functioning school-aged children with autism spectrum disorder. Journal of Autism and Developmental Disorders. 2011;41:1044–1052. http://doi.org/10.1007/s10803-010-1127-3. [PubMed] [Google Scholar]
Fein D, Barton M, Eigsti IM, Kelley E, Naigles L, Schultz RT, … Tyson K. Optimal outcome in individuals with a history of autism. Journal of Child Psychology and Psychiatry. 2013;54(2):195–205. http://doi.org/10.1111/jcpp.12037. [PMC free article] [PubMed] [Google Scholar]
Fitch A, Fein DA, Eigsti IM. Detail and gestalt focus in individuals with optimal outcomes from autism spectrum disorders. Journal of Autism and Developmental Disorders. 2015;45:1887–1896. http://doi.org/10.1007/s10803-014-2347-8. [PMC free article] [PubMed] [Google Scholar]
Fuentes CT, Mostofsky SH, Bastian AJ. Children with autism show specific handwriting impairments. Neurology. 2009;73:1532–1537. [PMC free article] [PubMed] [Google Scholar]
Ghaziuddin M, Butler F. Clumsiness in autism and Asperger syndrome: A further report. Journal of Intellectual Disability Research. 1998;42:43–48. [PubMed] [Google Scholar]
Gotham K, Pickles A, Lord C. Standardizing ADOS scores for a measure of severity in autism spectrum disorders. Journal of Autism and Developmental Disorders. 2009;39(5):693–705. http://doi.org/10.1007/s10803-008-0674-3. [PMC free article] [PubMed] [Google Scholar]
Grissmer D, Grimm KJ, Aiyer SM, Murrah WM, Steele JS. Fine motor skills and early comprehension of the world: Two new school readiness indicators. Developmental Psychology. 2010;46(5):1008–1017. http://doi.org/10.1037/a0020104. [PubMed] [Google Scholar]
Griswold DE, Barnhill GP, Myles BS, Hagiwara T, Simpson RL. Asperger syndrome and academic achievement. Focus on Autism and Other Developmental Disabilities. 2002;17(2):94–102. [Google Scholar]
Happe F, Frith U. The weak coherence account: Detail-focused cognitive style in autism spectrum disorders. Journal of Autism and Developmental Disorders. 2006;36(1):5–25. http://doi.org/10.1007/s10803-005-0039-0. [PubMed] [Google Scholar]
IBM Corporation. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: Author; 2013. [Google Scholar]
Ito M. Bases and implications of learning in the cerebellum: Adaptive control and internal model mechanism. In: Zeeuw C, Cicarita F, editors. Progress in brain research: Creating coordination in the cerebellum. Amsterdam, the Netherlands: Elsevier; 2005. pp. 95–109. [PubMed] [Google Scholar]
Jones CRG, Happe F, Golden H, Marsden AJS, Tregay J, Simonoff E, … Charman T. Reading and arithmetic in adolescents with autism spectrum disorders: Peaks and dips in attainment. Neuropsychology. 2009;23(6):718–728. http://doi.org/10.1037/a0016360. [PubMed] [Google Scholar]
Kanner L. Autistic disturbances of affective content. Nervous Child. 1943;2:217–250. [Google Scholar]
Knoch K. Unpublished doctoral dissertation. University of Connecticut; Storrs, CT: 2014. Early predictors of executive function abilities in school-aged children with autism spectrum disorders. [Google Scholar]
Lord C, Risi S, Lambrecht L. The autism diagnostic observation schedule - generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders. 2000;30(3):205–223. [PubMed] [Google Scholar]
May T, Rinehart N, Wilding J, Cornish K. The role of attention in the academic attainment of children with autism spectrum disorder. Journal of Autism and Developmental Disorders. 2013;43:2147–2158. http://doi.org/10.1007/s10803-013-1766-2. [PubMed] [Google Scholar]
Mayes SD, Calhoun SL. Ability profiles in children with autism: Influence of age and IQ. Autism. 2003a;6(4):65–80. [PubMed] [Google Scholar]
Mayes SD, Calhoun SL. Analysis of WISC-III, Stanford-Binet:IV, and academic achievement test scores in children with autism. Journal of Autism and Developmental Disorders. 2003b;33(3):329–341. [PubMed] [Google Scholar]
Mayes SD, Calhoun SL. WISC-IV and WIAT-II profiles in children with high-functioning autism. Journal of Autism and Developmental Disorders. 2008;38:428–439. http://doi.org/10.1007/s10803-007-0410-4. [PubMed] [Google Scholar]
Menard S. Sage University Paper series on Quantitative Applications in the Social Sciences, series no. 07-106. Thousand Oaks, CA: Sage; 1995. Applied logistic regression analysis. [Google Scholar]
Minshew NJ, Goldstein G. Autism as a disorder of complex information processing. Mental Retardation and Developmental Disabilities Research Reviews. 1998;4(2):129–136. [Google Scholar]
Minshew NJ, Goldstein G, Siegel DJ. Speech and language in high-functioning autistic individuals. Neuropsychology. 1995;9(2):255–261. [PubMed] [Google Scholar]
Minshew NJ, Goldstein G, Taylor HG, Siegel DJ. Academic achievement in high functioning autistic individuals. Journal of Clinical and Experimental Neuropsychology. 1994;16(2):261–270. [PubMed] [Google Scholar]
Mullen EM. Mullen Scales of Early Learning. Circle Pines, MN: American Guidance Services; 1995. [Google Scholar]
Nation K, Clarke P, Wright B, Williams C. Patterns of reading ability in children with autism spectrum disorder. Journal of Autism and Developmental Disorders. 2006;36:911–919. http://doi.org/10.1007/s10803-006-0130-1. [PubMed] [Google Scholar]
Pellicano E. The development of executive function in autism. Autism Research and Treatment. 2012;2012:1–8. http://dx.doi.org/10.1155/2012/146132. [PMC free article] [PubMed] [Google Scholar]
Reikeras E, Moser T, Tonnessen FE. Mathematical skills and motor life skills in toddlers: Do differences in mathematical skills reflect differences in motor skills? European Early Childhood Education Research Journal. 2015:1–17. http://doi.org/10.1080/1350293X.2015.1062664.
Reitzel J, Summers J, Lorv B, Szatmari P, Zwaigenbaum L, Georgiades S, Duku E. Pilot randomized controlled trial of a functional behavior skills training program for young children with autism spectrum disorder who have significant early learning skill impairments and their families. Research in Autism Spectrum Disorders. 2013;7(11):1418–1432. [Google Scholar]
Ricketts J, Jones CRG, Happe F, Charman T. Reading comprehension in autism spectrum disorders: The role of oral language and social functioning. Journal of Autism and Developmental Disorders. 2013;43:807–816. http://doi.org/10.1007/s10803-012-1619-4. [PubMed] [Google Scholar]
Robins DL, Fein D, Barton ML, Green JA. The modified checklist for autism in toddlers: An initial study investigating the early detection of autism and pervasive developmental disorders. Journal of Autism and Developmental Disorders. 2001;31(2):131–144. [PubMed] [Google Scholar]
Schopler E, Reichler RJ, Renner BR. Childhood autism rating scale. Los Angeles, CA: Western Psychological Services; 1988. [Google Scholar]
Sparrow SS, Balla DA, Cicchetti DV. Vineland adaptive behavior scales. Circle Pines, MN: American Guidance Services; 1984. [Google Scholar]
Sparrow SS, Cicchetti DV, Balla DA. Vineland adaptive behavior scales. 2. Circle Pines, MN: American Guidance Services; 2005. [Google Scholar]
StataCorp. Stata Statistical Software: Release 14. College Station, TX: Author; 2015. [Google Scholar]
Tager-Flusberg H. Brief report: Current theory and research on language and communication in autism. Journal of Autism and Developmental Disorders. 1996;26(2):169–172. [PubMed] [Google Scholar]
Towgood KJ, Meuwese JDI, Gilbert SJ, Turner MS, Burgess PW. Advantages of the multiple case series approach to the study of cognitive deficits in autism spectrum disorder. Neuropsychologia. 2009;47(13):2981–2988. [PMC free article] [PubMed] [Google Scholar]
Troyb E, Orinstein A, Tyson K, Helt M, Eigsti IM, Stevens M, Fein D. Academic abilities in children and adolescents with a history of autism spectrum disorders who have achieved optimal outcomes. Autism. 2014;18(3):233–243. http://doi.org/10.1177/1362361312473519. [PMC free article] [PubMed] [Google Scholar]
Troyb E, Knoch K, Herlihy L, Stevens MC, Chen CM, Barton M, … Fein D. Restricted and repetitive behaviors as predictors of outcome in autism spectrum disorders. Journal of Autism and Developmental Disorders. 2016;46:1282–1296. http://doi.org/10.1007/s10803-015-2668-2. [PMC free article] [PubMed] [Google Scholar]
Wechsler D. Wechsler individual achievement test. 2. San Antonio, TX: Psychological Corporation; 2001. [Google Scholar]
Welsh M, Parke RD, Widaman K, O’Neil R. Linkages between children’s social and academic competence: A longitudinal analysis. Journal of School Psychology. 2001;39(6):463–481. [Google Scholar]
World Health Organization. International statistical classification of diseases and related health problems. Geneva, Switzerland: Author; 2016. 10th rev. [Google Scholar]