Predictors and growth in receptive vocabulary from 4 to 8 years in children with and without autism spectrum disorder: A population-based study

Abstract

Few studies have examined growth and predictors of receptive vocabulary in children with autism spectrum disorder. Here we aimed to compare receptive vocabulary from 4 to 8 years and identify predictors of receptive vocabulary, at 8 years, in children with and without autism spectrum disorder. Participants were drawn from a nationally representative population-based study with two cohorts recruited at birth (N = 4983) and kindergarten (N = 5107). Receptive vocabulary growth was compared for children with and without autism spectrum disorder at 4 (n = 188, n = 7136), 6 (n = 215, n = 7297) and 8 (n = 216, n = 7408) years. Predictors of receptive vocabulary were analysed. Estimated mean receptive vocabulary scores for children without autism spectrum disorder were 2.3 units higher than the autism spectrum disorder group across three time points. This difference was significant (p = 0.004; 95% confidence interval 0.769–3.927). Children with and without autism spectrum disorder progressed at a similar pace. There was no significant difference between the proportions of children with and without autism spectrum disorder who had stable, improving and declining trajectories. Age was the only significant predictor of greater receptive vocabulary growth in children with autism spectrum disorder. Baseline receptive language and nonverbal IQ were significant predictors of receptive vocabulary ability at 8 years. These findings inform prognostic advice given to families on language outcomes.

Keywords

autism spectrum disorder communication development,language longitudinal trajectory

Introduction

Autism spectrum disorder (ASD) is characterised by difficulties with social communication and repetitive and restricted behaviours and interests (American Psychiatric Association, 2013). Difficulty with pragmatic language (the way we use language to interact with others) is a universal feature of ASD; yet there is substantial variability in children’s use of semantics, grammar and syntax (Tager-Flusberg, 2006). Some children with ASD have average or advanced vocabulary, grammar and sentence structure but around 15%–25% are nonverbal or minimally verbal and use less than five words (Norrelgen et al., 2014; Rose et al., 2016). There has been ongoing debate about whether there is a subgroup of children with ASD who have a phenotype that closely resembles those with specific language impairment (Rapin et al., 2009; Tager-Flusberg, 2006) and whether children with ASD and specific language impairment have shared or separate aetiological underpinnings (Kjelgaard and Tager-Flusberg, 2001; Whitehouse et al., 2008). Receptive and expressive language difficulties are one of the most commonly co-occurring features of ASD (Carlsson et al., 2013), with just over 60% estimated to have some degree of language impairment (Carlsson et al., 2013; Levy et al., 2010). The ability to use language by 5 years of age is an important predictor of later outcomes in children with ASD (Pickett et al., 2009) and difficulties using language in children with ASD are associated with several adverse outcomes including behaviour difficulties (McClintock et al., 2003; Sigafoos, 2000), poor adaptive functioning and social skills (Anderson et al., 2007; Baghdadli et al., 2007; Hudry et al., 2010) and self-injurious behaviours (Baghdadli et al., 2003).

Receptive and expressive language growth in ASD

A number of studies have investigated language development in ASD longitudinally. With few exceptions (e.g. Magiati et al., 2011; Pickles et al., 2014; Sigman and McGovern, 2005), most have focused on children with ASD. A variety of measures have been used to assess language development in children with ASD, including parent questionnaires (MacArthur-Bates Communicative Development Inventories; Bopp and Mirenda, 2011; Charman et al., 2003b; Luyster et al., 2007; Tek et al., 2013), standardised assessments of receptive and expressive language (Preschool Language Scales; Bopp and Mirenda, 2011; Ellis Weismer and Kover, 2015 and Test of Oral Language Development; Bennett et al., 2008), language subtests of developmental assessment tools (Mullen Scales of Early Learning; Paul et al., 2008; Tek et al., 2013; Thurm et al., 2015; Toth et al., 2006) and adaptive measures of communication (Vineland Adaptive Behaviour Scales; for example, Pickles et al., 2014; Szatmari et al., 2000; Toth et al., 2006). Findings from the afore-mentioned studies have shown substantial variation in language ability and trajectory over time with some reporting declines in adaptive communication (e.g. Magiati et al., 2011) and others reporting a proportion of children with language acceleration who no longer meet criteria for language delay (e.g. Szatmari et al., 2000). Most of these studies reported that children with ASD had lower mean language ability, but on average, tracked at a similar pace to reference norms during the preschool and early school years. That is, the gap between children with and without ASD did not grow larger.

Change in mean scores over time can provide valuable information about language growth, but our understanding of how individual children track over time and whether there are different trajectories for different levels of language ability remains limited. This is because few studies have sub-grouped children into different language ability types, such as age appropriate or delayed. One study of 17 children with ASD (mean age 32 months) identified two distinct expressive language profiles of growth over time (Tek et al., 2013). In this study, those children with ASD and higher language ability (n = 8) had comparable trajectories to typically developing children (n = 18) on most language measures and those with lower verbal ability (n = 9) had substantially flatter language trajectories.

Receptive vocabulary growth in ASD

Six studies (n > 30) have reported change in receptive vocabulary over time in children with ASD using standard scores or age-equivalent scores, with inconsistent findings. Pellicano administered the Peabody Picture Vocabulary Test – Third Edition (PPVT-III) with 37 children with IQ >80 at 5 and 8 years (Pellicano, 2010). Children demonstrated a decrease in standard scores from 97.1 (95% confidence interval (CI) 93.7–100.5) to 93.9 (CI 88.7–99.1) over the 2.7-year period (average range for PPVT-III = 85–115; standard scores should remain stable over time (similar values at each time point) if a child is making age appropriate progress). Receptive vocabulary was measured in another study using the British Picture Vocabulary Scales (BPVS) with children who had a range of intellectual abilities assessed at 3, 5 and 10 years (Magiati et al., 2011). Here mean scores were substantially below average (average range for BPVS = 85–115), but children had an increase in standard scores from 48.6 (95% CI 43.1–54.1) to 55.6 (95% CI 48.4–62.6) over 6.9 years, indicating some mean ‘catch up’ to reference norms over time. Similarly, Thomas (2009) reported an increase in mean PPVT-III standard scores for 69 children with ASD and mixed intellectual ability, from 70.9 (95% CI 64.4–77.3) at baseline (aged 4.5 years) to 77.5 (95% CI 71.3–83.7) 5 years later. None of these three studies demonstrated a statistically significant mean change (increase or decrease) in receptive vocabulary standard scores at a group level over time, although there may have been clinically meaningful change in some individuals.

Further to the use of standard scores as described above, three studies (n > 30) have reported change in receptive vocabulary (raw scores were converted to age-equivalents with reference to the PPVT-III test manual) over time using the PPVT-III (Bopp et al., 2009; Freeman et al., 1985; Steele et al., 2003). In all three studies, mean age-equivalents increased over time. In two studies, the gap between mean age-equivalent and chronological age grew larger (Bopp et al., 2009; Freeman et al., 1985). In the other study, children demonstrated some ‘catch up’ towards chronological age (Steele et al., 2003). The clinical significance of these growing or narrowing gaps between chronological and receptive vocabulary age is difficult to interpret given natural variation in trajectories, lack of control groups for comparison, and the way the data are presented by the studies (e.g. most studies present mean scores at two time points but have not used specific analyses to compare language change over time and most have not factored in repeated measures over time).

Given the heterogeneity of language development in ASD and the contrasting findings of longitudinal studies of receptive vocabulary, more information about how individual children track over time using large, representative samples are critical if we are to fully understand the prognosis of language difficulties in ASD.

Predictors of language growth and outcome in children with and without ASD

A substantial number of studies have examined putative demographic and behavioural predictors of language outcomes in children with ASD. Interest in this topic has been fuelled by the need to identify key ingredients for interventions that may facilitate language development. The types of predictors, ages at which predictors have been assessed and outcomes measured have varied widely and very few studies have replicated the methods of earlier studies so that findings can be compared. Two of the most consistent demographic risk factors for later language impairment are being male and a family history of speech and language delays (Bishop et al., 2003; Lindgren et al., 2009; McKean et al., 2015; Zubrick et al., 2007). Socio-economic disadvantage has also been identified as a risk factor for later language delay in studies of children with and without ASD (Christensen et al., 2014; Ellis Weismer and Kover, 2015; McKean et al., 2015; Rescorla, 2011; Taylor et al., 2013).

For abilities and behaviours, early language and IQ are probably the most consistent predictors of language outcomes in children with and without ASD (Anderson et al., 2007; Botting et al., 2001; Christensen et al., 2014; Ellis and Thal, 2008; Ellis Weismer and Kover, 2015; Thurm et al., 2007; Wodka et al., 2013). There is also some evidence that motor ability between 2 and 3 years of age (Bedford et al., 2015; Sandbank et al., 2017; Zubrick et al., 2007) and use of gesture between 10 months to 3 years of age (Ellis and Thal, 2008; Iverson and Wozniak, 2007; Mitchell et al., 2006) may predict rate of language development and initiation of communication. Joint attention and severity of autism symptoms prior to 3 years of age have also been identified as important predictors of language outcomes in younger children with ASD (Anderson et al., 2007; Charman et al., 2003a; Ellis Weismer and Kover, 2015), although these findings are not consistent across all studies (e.g. Thurm et al., 2007) when co-varying for cognition or other relevant factors.

Gaps in the literature

None of the afore-mentioned studies investigating receptive vocabulary contained samples of children without ASD for comparison, making it difficult to place receptive vocabulary growth in ASD within a developmental context. Furthermore, mean scores were reported, so it is unclear how individual children tracked over time and whether there are important subgroups of children with ASD with and without language difficulties that have different outcomes. Importantly, we understand little about the proportions of children who demonstrated improvement, stability or decline relative to what is expected for age. With one exception (Bopp et al., 2009), studies of receptive vocabulary development have not specifically assessed potential predictors of language outcomes and have had relatively small sample sizes (n range = 36–69) of children with ASD. Larger sample sizes studies are needed to generate more representative study findings. All studies used clinical/intervention rather than population-derived samples, so it is unclear whether the findings generalise to the broader population of children with ASD (Ellis Weismer and Kover, 2015).

Study aims

The aims of this study were to use a large population cohort to (1) compare mean receptive vocabulary between 4 and 8 years of age for children with and without ASD, (2) compare the proportions of children who demonstrate decreasing, stable or accelerating receptive vocabulary from 4 to 8 years of age for children with and without ASD, (3) examine predictors of receptive vocabulary growth from 4 to 8 years in children with ASD and (4) examine predictors of receptive vocabulary ability at 8 years in children with ASD. The sample size in this study is significantly larger than most studies to date.

Method

Study design

Participants were drawn from the Longitudinal Study of Australian Children (LSAC). A two-stage cluster sampling design was used (Gray and Sanson, 2005; Soloff et al., 2006) to recruit the participants. Stratification was used to ensure that the numbers of children invited to participate in the study were roughly proportionate to the numbers of children within each state or territory, the capital city statistical districts and the rest of each state. There is slight under-representation by single-parent families, families living in rental properties and non-English speaking families but LSAC is generally considered representative of Australian children. Weighting was used to adjust for non-response with characteristics of responders and non-responders identifying two factors associated with the probability of participating which were whether the mother speaks a language other than English at home and whether the mother has completed year 12. These factors were incorporated to offset the bias introduced by differential non-response patterns. Further detail of this including longitudinal and cross-sectional weighting is available (Cusack and Defina, 2013).

Two cohorts of children were recruited in 2004: one at birth (n = 4983) and the other in kindergarten (n = 5107). Here we combined both cohorts. At each LSAC wave (bi-yearly), trained staff completed face-to-face interviews with caregivers and direct assessments with children in the home. Primary caregivers also completed questionnaires about the child’s health and development. There have been seven waves of data collection in the LSAC (from 2004 to 2016). We use data from children when they were aged 4, 6 and 8 years (Waves 3–5 for the cohort recruited at birth and Waves 1–3 for the cohort recruited in kindergarten). Figure 1 shows the number of participants at each age and those who had data required for each analysis.

Figure 1.

Number of participants at each age enrolled in the study and the number of participants who had data included in each analysis.

In this study, we only include data for those children who completed the PPVT-III (Dunn et al., 1997) at the required time points (see Table 1 for further details).

Table 1.

Number of children who participated in each wave and completed the PPVT-III for birth and kinder cohort.

		Wave 1N (age)	Wave 2N (age)	Wave 3N (age)	Wave 4N (age)	Wave 5N (age)
Birth cohort	Total	4983 (<1 year)		4280(4 years)	4140(6 years)	3986(8 years)
	Completed PPVT-III			4151(4 years)	4098(6 years)	3946(8 years)
Kinder cohort	Total	5107(4 years)	4576(6 years)	4387(8 years)
	Completed PPVT-III	4494(4 years)	4530(6 years)	4343(8 years)

PPVT-III: Peabody Picture Vocabulary Test – Third Edition.

The control group consisted of all children whose parents did not report a diagnosis of ASD. We did not exclude children from the control group if they had other conditions (e.g. intellectual disability or developmental language disorder) because we wanted to provide an accurate representation of the distribution of risk in all children and for the children to be representative of the general population. If we compare to only children with typical development, we may introduce bias and artificially inflate any identified differences between the groups (Fombonne, 2016).

Measures

Demographic information

Demographic information was collected from the primary caregiver when children were 10 years of age (Table 2). Neighbourhood disadvantage was measured using the Socio-Economic Indexes for Areas Disadvantage Index (SEIFA). SEIFA scores summarise socio-economic advantage and disadvantage of areas using the Census of Population and Housing and includes income, education and unemployment (Australian Bureau of Statistics, 2013).

Table 2.

Sample characteristics at 10 years of age.

Demographic measure	ASD (N = 214–237)^a	Non-ASD (N = 7496–8853)^a	p value
Child age in months, mean (SD)	130.4 (3.8)N = 237	130.4 (3.8)N = 7696	0.892
Male (%)	79.8%N = 237	50.3%N = 7696	<0.001*
English main language spoken at home (%)	92.0%N = 229	86.1%N = 7496	<0.047*
Indigenous (%)	3.0%N = 237	2.8%N = 7477	0.903
Remote geographic location (%)	1.7%N = 236	3.8%N = 7620	0.106
Single-parent family (%)	23.2%N = 237	15.1%N = 7686	0.001*
Primary caregiver did not complete high school (%)	433.3%N = 231	33.6%N = 7593	0.307
Neighbourhood disadvantage, mean (SD)^b	998.0 (64.3)N = 237	1014.6 (68.6)N = 7695	<0.001*
NVIQ at 6 years, mean (SD)	9.92 (3.5)N = 214	10.6 (3.0)N = 8379	0.03
SDQ prosocial at 10 years, mean (SD)	6.25 (4.1)N = 236	7.9 (6.3)N = 7644	<0.001

ASD: autism spectrum disorder; SD: standard deviation; NVIQ: nonverbal IQ; SDQ: Strengths and Difficulties Questionnaire.

Sample size range.

The neighbourhood disadvantage index has a mean of 1000 and a standard deviation of 100, with higher scores indicating greater advantage.

p < 0.05.

ASD diagnosis

The primary caregiver was asked during interview ‘Does your child have any of the following ongoing conditions?’ They were then asked to specify ‘Autism, Aspergers or other autism spectrum’. Parents responded yes/no which was coded into a binary variable. The age of diagnosis and whether the diagnosis was mild/moderate/severe was also asked. This information was collected at 10 years.

Receptive vocabulary

An adapted (shortened) version of the PPVT-III (Dunn et al., 1997) was used to assess receptive vocabulary. This test involves children identifying pictures in response to a spoken word. The adapted PPVT-III was developed and validated for Australian children at 4, 6 and 8 years of age (Rothman, 2003). Reliability of the adapted PPVT-III was 0.76 and the Pearson product-moment correlation between the full and adapted PPVT-III was 0.93 for the whole sample (Rothman, 2003). Here we report PPVT-III raw scores at 4, 6 and 8 years.

Social functioning

Social functioning at 4 years was measured using both the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997) and the Paediatric Quality of Life Inventory (PedsQL 4.0 – parent report form; Varni et al., 2001). The PedsQL 4.0 is a 23-item questionnaire that asks caregivers to rate on a 5-point scale how much of a problem a particular area has been over the past month. A rating of 0 indicates ‘never a problem’ and 5 indicates ‘almost always a problem’. There are four subscales: physical, emotional, social and school functioning. Here we report on two questions from the ‘social functioning’ subscale: ‘Child often/almost always has difficulty getting along with other children’ and ‘Children often/almost always don’t want to be friends’.

The SDQ is a 25-item parent questionnaire. It has five subscale scores (range 0–10) and a total score (range 0–40), with higher scores indicating greater difficulties. Past studies using the SDQ have reported that children with ASD have greater difficulties in prosocial behaviour and peer relationships than controls or children with other developmental conditions (e.g. Iizuka et al., 2010; May et al., 2017; Randall et al., 2015; Russell et al., 2013). Therefore, we elected to use the ‘prosocial’ subscale of the SDQ in this study.

Nonverbal intelligence

An estimate of nonverbal IQ (NVIQ) was obtained using the Matrix Reasoning subtest from the Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV; Wechsler, 2003) when the children were 6 years old.

Ethical considerations

The LSAC study is approved by the Australian Institute of Family Studies Ethics Committee, and parents provide written informed consent. Permission was granted to use the LSAC data for addressing the current study aims.

Statistical analyses

Analyses were conducted in Stata version 14.0. Survey methods were used to account for the unequal probability of participant selection into the sample, non-response and sample attrition, and the multi-stage, clustered sampling design (Soloff et al., 2006). Summary statistics were used to describe the demographic characteristics of children with and without ASD. Variance inflation factors (VIFs) were around 1 for language and NVIQ scores which was well below the generally used cut-off of above 4.

Comparison of receptive vocabulary between groups (mean differences and trajectory types)

Receptive vocabulary raw scores for children with ASD were compared to those for children without ASD at 4 years (n = 188, n = 7136, respectively), 6 years (n = 215, n = 7297) and 8 years (n = 216, n = 7408). Mean receptive vocabulary scores were plotted at 4, 6 and 8 years using the generalised estimating equations (GEE) method for fitting marginal models. An exchangeable correlation structure was used and robust standard errors. This enabled us to take the dependence of the multiple responses from each participant into account. The mean trajectories from 4 to 6 and 6 to 8 years for ASD and non-ASD were plotted for these models and compared between the two groups. Included in the model were possible confounding variables: child gender, single-parent family, English as a second language and neighbourhood disadvantage.

The numbers of children in each group who had declining, stable and improving trajectories between 4 and 8 years of age were compared using the chi-square statistic. We did not include the 6–year-old time point data in this analysis. We used one standard deviation change as the cut point for defining these three different trajectory types. One standard deviation was used as this would indicate a substantial change clinically. This cut point has also been used in previous studies of language development (e.g. Snowling et al., 2016).

Predictors of receptive vocabulary growth from 4 to 8 years

Linear regression analyses were used to examine putative predictors of receptive vocabulary growth. Predictors included nonverbal intelligence, social functioning (PedsQL 4.0 and SDQ measures), primary caregiver education, child age, gender, single-parent family status, English as second language and neighbourhood disadvantage. Receptive vocabulary growth was measured by the difference in PPVT-III raw score from 4 to 8 years.

Predictors of receptive vocabulary outcome at 8 years

Linear regression analyses were used to examine putative predictors of receptive vocabulary outcome on the PPVT-III at 8 years. We used the same predictors that we used to examine receptive vocabulary growth (described in the section above) but added receptive vocabulary raw scores on the PPVT-III at 4 years to the model.

Results

Of the combined cohort, 237 (n = 145 birth cohort; n = 92 kinder cohort) children were reported by parents to have received a diagnosis of ASD by 10 years (see Table 2). A higher proportion of children with ASD were males compared to those without ASD. A significantly higher proportion of children from single-parent families and homes where English was the main language were reported to have ASD. Children with ASD also had significantly greater neighbourhood disadvantage and significantly lower scores on the NVIQ measure and in social functioning on the SDQ. We compared the characteristics of those who did and did not have data included in the declining, stable and improving trajectories analysis. There were no differences in sex or the proportions of children without data for those with ASD (24%) and without ASD (24%). When we combined the two groups (ASD and non-ASD), we found children missing from the analysis were younger (z = 3.62, p < 0.001), had higher socio-economic disadvantage (z = 6.99, p < 0.001), were more likely to be of Aboriginal or Torres Strait Islander descent (z = –11.2, p < 0.001), were more likely to be living in remote areas (z = –4.67, p < 0.001) and were more likely to be from homes where English was not the first language spoken (z = 46.5, p < 0.001). They were also more likely to be from single-parent families (z = 5.33, p < 0.001) and were more likely to have a primary caregiver who did not complete high school (z = 13.25, p < 0.001).

All descriptive variables were measured at 10 years except NVIQ which was measured at 6 years using the Matrix Reasoning subtest of the WISC-IV. Social communication was measured using the prosocial subscale from the SDQ. The N in this table relates to the number of children with ASD at age 4 who were then classified as declining, improving or stable based on the receptive vocabulary change between 4 and 8 years.

Comparison of receptive vocabulary between groups (mean difference and trajectory type)

There was substantial variability in individual receptive vocabulary scores for children with and without ASD. Mean receptive vocabulary raw scores for children with and without ASD at 4, 6 and 8 years are provided in Supplementary Table 1. The GEE analysis, controlling for covariates, found estimated mean receptive vocabulary scores for children without ASD were 2.3 units higher than the ASD group across the three waves of data collection. The difference between the two groups was significant (p = 0.004; 95% CI 0.769–3.927). The following covariates were also significant in this model: single-parent family (B = 0.84, p =< 0.001; CI 0.5429–1.1371), parent education (B = 1.60; p =< 0.001; CI 1.4124–1.7931), SEIFA (B = 0.01, p < 0.001; CI 0.0103–0.1334) and having English as the main language at home (B = 3.09, p < 0.001; CI 2.781–3.398).

Despite children with ASD having lower mean scores at each time point based on the GEE estimates (solid lines), on average, the children with ASD tracked at a similar pace, which was broadly in parallel to children without ASD over time (Figure 2). We completed further analysis using a linear mixed model to assess convergence of trajectories between the children with and without ASD and found very similar estimates between the two analyses indicating the convergence between the groups was not adequate to violate the constant assumption.

Figure 2.

Mean receptive vocabulary (raw scores and generalised estimating equation estimates) for children with and without ASD at 4, 6 and 8 years.

There was no significant difference between the proportion of children who had declining, stable and improving change (relative to rest of cohort) between the two groups (χ² = 4.7, p = 0.094). Children with ASD in the declining group did not have lower mean scores at baseline than the other two groups. Rather, the declining group had significantly higher mean scores at baseline than both the stable and improving trajectories (t = 32, p = 0.0001, 95% CI = 5.066–5.734 and t = 36.9, p = 0.0001, 95% CI = 11.726–13.074, respectively). Likewise, for children without ASD, the declining group had significantly higher mean raw scores than the stable and improving groups (t = 1920.2346, p = 0.0001 and t = 1770.6443, p = 0.0001, respectively) (Table 3).

Table 3.

Proportions of children with and without ASD identified at age 4 with improving, stable and declining change over time and mean raw scores at baseline (4 years) for each group.

Change type	ASD (n = 181)		Non-ASD (n = 7474)
Change type	n (%)	Mean (SD), 95% CI	n (%)	Mean (SD), 95% CI
Declining	18 (10)	69.5 (1.4), 66.5–72.5	1085 (15)	70.9 (0.1), 70.6–71.2
Stable	129 (71)	64.1 (0.5), 63.1–65.0	5301 (71)	65.0 (0.1), 64.9–65.2
Improving	34 (19)	57.1 (1.0), 55.0–59.1	1088 (15)	58.9 (0.18), 58.6–59.3

ASD: autism spectrum disorder; CI: confidence interval; SD: standard deviation.

Predictors of receptive vocabulary change over time from 4 to 8 years in children with ASD

Only one of 11 predictors examined using multivariate analysis made a statistically significant independent contribution to growth in receptive vocabulary in the children with ASD. That is, only younger age predicted greater positive change in raw sores from 4 to 8 years. The model explained 13% of the variance in receptive vocabulary change over time (p = 0.037) (Table 4).

Table 4.

Predictors of receptive vocabulary change over time from 4 to 8 years in children with ASD (n = 156).

Variable	Coefficient	95% CI	p value
Age at year 4 assessment	–0.60	–0.94 to −0.25	0.001*
Male gender	–0.76	–3.15 to 1.62	0.528
Single-parent family	–0.58	–3.44 to 2.28	0.689
Number of children in the home	0.17	–0.82 to 1.16	0.730
English main language spoken at home	–4.31	–11.56 to 2.94	0.242
Primary caregiver finished high school	1.08	–1.19 to 3.35	0.350
SEIFA disadvantage	–0.02	–0.03 to 0.001	0.071
SDQ prosocial	–0.31	–0.85 to 0.24	0.270
Peds QL 4.0 Q1 (getting along with others)	–0.51	–3.72 to 2.71	0.755
Peds QL 4.0 Q2 (peers wanting to be friends)	–0.07	–3.81 to 3.67	0.971
NVIQ (matrix reasoning) at 6 years	–0.11	–0.42 to 0.20	0.491

ASD: autism spectrum disorder; CI: confidence interval; SEIFA: Socio-Economic Indexes for Areas; SDQ: Strengths and Difficulties Questionnaire; NVIQ: nonverbal IQ; PedsQL 4.0: Paediatric Quality of Life Inventory; WISC-IV: Wechsler Intelligence Scale for Children.

All predictor variables were collected at 4 years. Social functioning was measured at 4 years using the prosocial subscale from the SDQ and two social questions from the PedsQL 4.0. NVIQ was measured at 6 years using the Matrix Reasoning subtest of the WISC-IV. The neighbourhood disadvantage index has a mean of 1000 and a standard deviation of 100, with higher scores indicating greater advantage.

p < 0.05.

Predictors of receptive vocabulary outcomes at 8 years in children with ASD

Linear regression analyses found scores on the PPVT-III at 4 years and NVIQ at 6 years significantly predicted receptive vocabulary outcome at 8 years (p < 0.001 and p = 0.012, respectively). None of the other variables in the model (demographic factors and social functioning) were significant predictors. The model explained 31% of the variance in receptive vocabulary at 8 years (p < 0.001) (Table 5).

Table 5.

Predictors of receptive vocabulary at 8 years in children with ASD (n = 156).

Variable	Coefficient	95% CI	p value
Age at year 4 assessment	–0.09	–0.37 to 0.2	0.543
Male gender	–1.33	–3.17 to 0.51	0.154
Single-parent family	–0.46	–2.66 to 1.74	0.679
Number of children in the home	0.07	–0.70 to 0.83	0.866
English main language spoken at home	–3.05	–8.63 to 2.52	0.281
Primary caregiver finished high school	1.50	–0.25 to 3.25	0.091
SEIFA disadvantage	–0.002	–0.02 to 0.01	0.765
SDQ prosocial	–0.08	–0.50 to 0.34	0.704
Peds QL 4.0 Q1 (getting along with others)	–1.27	–3.75 to 1.21	0.314
Peds QL 4.0 Q2 (peer wanting to be friends)	0.29	–2.58 to 3.17	0.840
PPVT-III score at 4 years	0.32	0.19 to 0.46	<0.001*
NVIQ (matrix reasoning) at 6 years	0.33	0.07 to 0.58	0.012*

ASD: autism spectrum disorder; CI: confidence interval; SEIFA: Socio-Economic Indexes for Areas; SDQ: Strengths and Difficulties Questionnaire; PPVT-III: Peabody Picture Vocabulary Test – Third Edition; NVIQ: nonverbal IQ; PedsQL 4.0: Paediatric Quality of Life Inventory; WISC-IV: Wechsler Intelligence Scale for Children - 4th ed.

All predictor variables were collected at 4 years. Social communication was measured at 4 years using the prosocial subscale from the SDQ and two social questions from the PedsQL 4.0. NVIQ was measured at 6 years using the Matrix Reasoning subtest of the WISC-IV.

p < 0.05.

Discussion

In this study, we compare growth and predictors of receptive vocabulary in children with and without ASD ascertained through a population-based study. There was receptive vocabulary growth over time for both the ASD and non-ASD groups. This finding was unsurprising based on previous studies of language growth in children with and without ASD (Anderson et al., 2007; Ellis Weismer and Kover, 2015; McKean et al., 2015; Tek et al., 2013; Ukoumunne et al., 2012). At a group level, mean receptive vocabulary raw scores were lower for children with ASD; yet on average, the children progressed at a similar pace to children without ASD between 4, 6 and 8 years of age. Comparable proportions of children with and without ASD had declining, stable and improving receptive vocabulary change over time relative to the rest of the cohort’s development from 4 to 8 years. In general, the children in our sample with ASD were no more likely than children without ASD to fall further behind their peers with around 90% of children with ASD in our study demonstrating stable (less than 1 standard deviation change) or improving growth relative to 86% in the non-ASD group.

Previous studies, with sample sizes of n > 30, and using different methods to the current study have reported increases (Magiati et al., 2011; Thomas, 2009) and decreases (Pellicano, 2010) in mean receptive vocabulary standard scores over time. Mean follow-up scores did, however remain within one standard deviation of baseline scores across all three studies. While the different methodology and types of scores used in the afore-mentioned studies limits direct comparison with the current study, the findings are generally consistent. That is, on average, children with ASD have lower scores than those without ASD; yet most maintain progress at a rate that is comparable to children without ASD or reference norms.

Children with ASD who had declining or stable trajectories in our study did not have lower mean baseline PPVT-III scores at 4 years compared to those in the improving trajectory group. This is in contrast to other studies that have found children with the most delayed language (especially those described as minimally verbal) tend to have flatter trajectories and make more limited progress (Anderson et al., 2007; Ellis Weismer and Kover, 2015; Tek et al., 2013). It is also important to note however, that in these studies there was a group of children with severely delayed language at baseline who made rapid progress over time and ‘caught up’ to peers (Anderson et al., 2007; Ellis Weismer and Kover, 2015). Our study may not have included the same types of children reported in previous studies because inclusion required the ability to complete the PPVT-III and we recruited from a population sample rather than a clinical sample.

Apart from younger age at assessment, none of the variables included in our regression model were significant predictors of greater receptive vocabulary growth from 4 to 8 years in the children with ASD. The finding that age was a significant predictor of greater growth in receptive vocabulary may be explained because children in the LSAC were not all assessed at the same time point relative to birth date within the year. It is possible some children had larger gaps between testing and therefore appeared to make more progress (larger increase in raw scores).

Another study of language development in preschool children with ASD found maternal education in combination with response to joint attention and cognition correctly predicted 80% of children into the high or low language groups at 5.5 years. These same factors did not predict language growth with ASD symptom severity and cognition having the strongest influence on growth (Ellis Weismer and Kover, 2015). An explanation for discrepancies across our study and others could be the types of language measures used, different age periods studied and the environmental changes that occur at these different ages, for example, school attendance. Furthermore, some authors have suggested factors such as joint attention, and ASD symptom severity may be more important for children who have more severely delayed language or who are younger relative to older children with less severe language impairments (Ellis Weismer and Kover, 2015; Paul et al., 2013; Thurm et al., 2015).

It was of interest that social functioning (peer relations and prosocial behaviour) did not impact amount of change in receptive vocabulary in children with ASD in this study. While no other longitudinal studies have investigated predictors of receptive vocabulary in ASD during the same time periods examined here, our finding is consistent with another longitudinal study that used a more comprehensive measure of language as the outcome (Ellis Weismer and Kover, 2015). Specifically, Ellis Weismer and colleagues found that while ASD symptom severity was a significant predictor of receptive and expressive language growth from 2.5 to 5.5 years, social ability (measured by the socialisation domain of the Vineland Adaptive Behaviour Scales) was not a predictor (Ellis Weismer and Kover, 2015). Our work extends Ellis Weismer and colleagues’ findings to older children, suggesting social functioning, including peer relationships, may not be as important as other factors in promoting or impeding receptive vocabulary growth. The regression model used here explained only 13% of the variance in receptive vocabulary growth indicating factors that reliably predict receptive language growth remain largely elusive. It has been suggested that child characteristics may be more important than environmental influences on language growth by some authors (Ellis Weismer and Kover, 2015) and it is also likely genetic and neural factors play roles that are yet to be fully understood. More recently, studies have started to look beyond demographic and behavioural characteristics by combining neuroimaging findings with behavioural markers to predict language outcomes in preschoolers (e.g. Lombardo et al., 2015). These factors, along with new knowledge about the genetic aetiology of language difficulties in ASD, should assist in providing more individualised prognosis information for families.

Receptive vocabulary and NVIQ at baseline (4 and 6 years, respectively) were predictive of receptive vocabulary outcomes in children with ASD and these two factors explained a moderate amount (31%) of the variance. Children with better receptive vocabulary at 4 years, and better IQ at 6 years, were more likely to have better receptive vocabulary at 8 years. This finding supports prior research that has identified early IQ/cognition and language ability to be important predictors of later language ability (e.g. Anderson et al., 2007; Charman et al., 2005; Kover et al., 2016; Luyster et al., 2007; Thurm et al., 2007, 2015; Wodka et al., 2013). These findings are also consistent with another study that used a standardised, clinician-administered language assessment (Preschool Language Scales – Fourth Edition) to measure language outcomes in preschool children with ASD (Ellis Weismer and Kover, 2015). Ellis Weismer and colleagues included a clinical sample and slightly different methods to the current study however, despite the different predictor variables and language outcome tools used in the regression models, both studies identified IQ/cognition and baseline verbal ability to be important predictors of receptive language outcome.

The strengths of this study are that it is the first to our knowledge to use a large, representative population-based sample to examine receptive vocabulary development in children with ASD. Findings from population-based studies are typically more generalisable than those from selected clinical samples and are more likely to contain children with milder phenotypes relative to clinical samples previously reported (Kim et al., 2011). These milder phenotypes are important if we are to understand the full picture of language development in ASD. Here we included a large sample of children without ASD from the same cohort for comparison. Most studies to date include clinical samples of children with ASD and compare these children to reference norms in test manuals. Comparison within the same cohort is needed to establish the specificity of language differences to ASD relative to other children of the same age and generation (Tager-Flusberg, 2016). Furthermore, we investigated subgroups of children with ASD by examining different trajectory types. This had rarely been done in previous research. The longitudinal design of this study allows us to assess predictors of both receptive vocabulary outcome and change over time. Predictors were selected at 4 years of age which is the age most children receive a diagnosis of ASD (Centers for Disease Control and Prevention, 2014). This is also a time when parents are seeking answers to their questions about prognosis and considering key factors to prioritise for intervention.

A number of limitations in this study should be acknowledged. While, as outlined earlier, the population-based study design brought a number of strengths, the nature of the design meant it was not possible to provide uniform ASD diagnoses. Rather, children were given a diagnosis in the community, and this diagnosis was reported to the study by the child’s parents. Furthermore, the LSAC is a broad study on child development with limitations regarding burden to families who participate, so we did not have comprehensive measure of receptive language and there was no associated measure of expressive language. The PPVT-III is a widely used, standardised tool with good measurement properties; however, it only measures one semantic-based aspect of receptive language. It has been suggested by some authors that children with ASD may perform better on vocabulary measures relative to omnibus measures of language (Kjelgaard and Tager-Flusberg, 2001; Kover et al., 2013). Our control sample may have included children with other diagnoses (e.g. intellectual disability) and this may have contributed to higher proportions of children with declining trajectories in the control group relative to if we had compared to a ‘typical’ group. We would, however, expect the numbers of these children to be small.

The study may be broadly representative of the population of children with ASD but the findings will only be applicable to children who have the ability to complete the PPVT-III. That is, they are at a developmental level where they can listen and point to pictures in a manual. We used raw scores (rather than standard scores) so we could include as many children as possible with low language ability. This also facilitated greater differentiation of scores and a more nuanced examination of change over time relative to standard scores. Nevertheless, there will be some children with very low language ability that our findings do not apply to, and there is some evidence these children may have slower, flatter language trajectories than children with ASD who have stronger verbal skills (Anderson et al., 2007; Ellis Weismer and Kover, 2015; Tek et al., 2013; Thurm et al., 2015). By excluding these children, it is possible our findings artificially inflated the rate of language growth and outcomes. Longitudinal dropout of participants and missing data resulted in only around 75% of the original sample being included in these analyses. While sample weights were used, the possibility of bias arising from non-response at recruitment, item-missing data and selective dropout cannot be excluded.

There are several clinical implications that arise from the findings of this study. In general, we provide prognostic information that can be considered when giving general advice to parents about how receptive vocabulary development may track over time in children with ASD and how this may differ to children without ASD. We also provide information about the factors that may be important in influencing receptive vocabulary growth and outcomes at 8 years. Children with lower cognitive and receptive vocabulary skills at 4 years appear to be more likely to have ongoing receptive vocabulary challenges at 8 years and this is important when considering the focus of interventions. For example, if the aim is to improve language skills, developing and implementing evidence-based intervention programmes that focus on specific cognition and language skills may be important. The findings from this study can also inform policy makers, service providers and parents around anticipated future support and resource needs for children with ASD.

Supplemental Material

AUT801617_Lay_Abstract – Supplemental material for Predictors and growth in receptive vocabulary from 4 to 8 years in children with and without autism spectrum disorder: A population-based study

Supplemental material, AUT801617_Lay_Abstract for Predictors and growth in receptive vocabulary from 4 to 8 years in children with and without autism spectrum disorder: A population-based study by Amanda Brignell, Tamara May, Angela T Morgan and Katrina Williams in Autism

Supplemental Material

AUT801617_Supplementary_material – Supplemental material for Predictors and growth in receptive vocabulary from 4 to 8 years in children with and without autism spectrum disorder: A population-based study

Supplemental material, AUT801617_Supplementary_material for Predictors and growth in receptive vocabulary from 4 to 8 years in children with and without autism spectrum disorder: A population-based study by Amanda Brignell, Tamara May, Angela T Morgan and Katrina Williams in Autism

Footnotes

Acknowledgements

The researchers acknowledge the Australian NHMRC for salary support through a Practitioner Fellowship #1105008 (A.T.M.) and NHMRC Centre of Research Excellence in Speech and Language Neurobiology #1116976 (A.T.M., A.B.). The authors also acknowledge the Apex Foundation for Research into Intellectual Disabilities and The RCH Foundation for support of the APEX Australia Chair of Developmental Medicine (K.W.). Infrastructure support was provided by the Victorian Government’s Operational Infrastructure Support Programme. The authors wish to thank Professor Cate Taylor, Telethon Kids Institute, for her review of the manuscript and insightful comments.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Amanda Brignell

References

American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Association.

Anderson

Lord

Risi

et al . (2007) Patterns of growth in verbal abilities among children with autism spectrum disorder. Journal of Consulting and Clinical Psychology 75: 594–604.

Australian Bureau of Statistics (2013) Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA). Canberra, ACT, Australia: Australian Bureau of Statistics.

Baghdadli

Pascal

Grisi

et al . (2003) Risk factors for self-injurious behaviours among 222 young children with autistic disorders. Journal of Intellectual Disability Research 47: 622–627.

Baghdadli

Picot

M-C

Michelon Bodet

et al . (2007) What happens to children with PDD when they grow up? Prospective follow-up of 219 children from preschool age to mid-childhood. Acta Psychiatrica Scandinavica 115: 403–412.

Bedford

Pickles

Lord

(2015) Early gross motor skills predict the subsequent development of language in children with autism spectrum disorder. Autism Research 9: 993–1001.

Bennett

Szatmari

Bryson

et al . (2008) Differentiating autism and Asperger syndrome on the basis of language delay or impairment. Journal of Autism and Developmental Disorders 38: 616–625.

Bishop

DVM

Price

Dale

et al . (2003) Outcomes of early language delay: II. Etiology of transient and persistent language difficulties. Journal of Speech, Language, and Hearing Research 46: 561–575.

Bopp

Mirenda

(2011) Prelinguistic predictors of language development in children with autism spectrum disorders over four-five years. Journal of Child Language 38: 485–503.

10.

Bopp

Mirenda

Zumbo

(2009) Behavior predictors of language development over 2 years in children with autism spectrum disorders. Journal of Speech, Language, and Hearing Research 52: 1106–1120.

11.

Botting

Faragher

Simkin

et al . (2001) Predicting pathways of specific language impairment: what differentiates good and poor outcome? Journal of Child Psychology and Psychiatry 42: 1013–1020.

12.

Carlsson

Norrelgen

Kjellmer

et al . (2013) Coexisting disorders and problems in preschool children with autism spectrum disorders. Scientific World Journal 2013: 213979.

13.

Centers for Disease Control and Prevention (2014) Prevalence of autism spectrum disorder among children aged 8 years – Autism and developmental disabilities monitoring network, 11 sites, United States, 2010. Morbidity and Mortality Weekly Report 63: 1–21.

14.

Charman

Baron-Cohen

Swettenham

et al . (2003a) Predicting language outcome in infants with autism and pervasive developmental disorder. International Journal of Language & Communication Disorders 38: 265–285.

15.

Charman

Drew

Baird

et al . (2003b) Measuring early language development in preschool children with autism spectrum disorder using the MacArthur Communicative Development Inventory (infant form). Journal of Child Language 30: 213–236.

16.

Charman

Taylor

Drew

et al . (2005) Outcome at 7 years of children diagnosed with autism at age 2: predictive validity of assessments conducted at 2 and 3 years of age and pattern of symptom change over time. Journal of Child Psychology and Psychiatry 46: 500–513.

17.

Christensen

Zubrick

Lawrence

et al . (2014) Risk factors for low receptive vocabulary abilities in the preschool and early school years in the longitudinal study of Australian children. PLoS ONE 9: e101476.

18.

Cusack

Defina

(2013) Growing up in Australia: the longitudinal study of Australian children. Technical paper No. 10 Wave 5 weighting & non response. Australian Institute of Family Studies. Available at: http://data.growingupinaustralia.gov.au/pubs/technical/tp10.pdf (Accessed 26 February 2018).

19.

Dunn

Williams

(1997) Peabody Picture Vocabulary Test-III. Circle Pines, MN: American Guidance Service.

20.

Ellis

Thal

(2008) Early language delay and risk for language impairment. Perspectives on Language Learning and Education 15: 93–100.

21.

Ellis Weismer

Kover

(2015) Preschool language variation, growth, and predictors in children on the autism spectrum. Journal of Child Psychology and Psychiatry 56: 1327–1337.

22.

Fombonne

(2016) Editorial: isolating the essential difference – The importance of choosing the right type and sufficient numbers of controls in research on neurodevelopmental disorders and mental health conditions. Journal of Child Psychology and Psychiatry 57(11): 1203–1204.

23.

Freeman

Ritvo

Needleman

et al . (1985) The stability of cognitive and linguistic parameters in autism: a five-year prospective study. Journal of the American Academy of Child and Adolescent Psychiatry 24: 459–464.

24.

Goodman

(1997) The Strengths and Difficulties Questionnaire: a research note. Journal of Child Psychology and Psychiatry 38: 581–586.

25.

Gray

Sanson

(2005) Growing up in Australia: The longitudinal study of Australian children. Family Matters 72: 4–9.

26.

Hudry

Leadbitter

Temple

et al . (2010) Preschoolers with autism show greater impairment in receptive compared with expressive language abilities. International Journal of Language & Communication Disorders 45: 681–690.

27.

Iizuka

Yamashita

Nagamitsu

et al . (2010) Comparison of the Strengths and Difficulties Questionnaire (SDQ) scores between children with High-Functioning Autism Spectrum Disorder (HFASD) and Attention-Deficit/Hyperactivity Disorder (AD/HD). Brain & Development 32: 609–612.

28.

Iverson

Wozniak

(2007) Variation in vocal-motor development in infant siblings of children with autism. Journal of Autism and Developmental Disorders 37: 158–170.

29.

Kim

Leventhal

Koh

et al . (2011) Prevalence of autism spectrum disorders in a total population sample. American Journal of Psychiatry 168: 904–912.

30.

Kjelgaard

Tager-Flusberg

(2001) An investigation of language impairment in autism: implications for genetic subgroups. Language and Cognitive Processes 16: 287–308.

31.

Kover

Edmunds

Ellis Weismer

(2016) Brief report: ages of language milestones as predictors of developmental trajectories in young children with autism spectrum disorder. Journal of Autism and Childhood Schizophrenia 46: 2501–2507.

32.

Kover

McDuffie

Hagerman

et al . (2013) Receptive vocabulary in boys with autism spectrum disorder: cross-sectional developmental trajectories. Journal of Autism and Developmental Disorders 43: 2696–2709.

33.

Levy

Giarelli

Lee

L-C

et al . (2010) Autism spectrum disorder and co-occurring developmental, psychiatric, and medical conditions among children in multiple populations of the United States. Journal of Developmental and Behavioral Pediatrics 31: 267–275.

34.

Lindgren

Folstein

Tomblin

et al . (2009) Language and reading abilities of children with autism spectrum disorders and specific language impairment and their first-degree relatives. Autism Research 2: 22–38.

35.

Lombardo

Pierce

Eyler

et al . (2015) Different functional neural substrates for good and poor language outcome in autism. Neuron 86: 567–577.

36.

Luyster

Qiu

Lopez

et al . (2007) Predicting outcomes of children referred for autism using the MacArthur-Bates Communicative Development Inventory. Journal of Speech, Language, and Hearing Research 50: 667–681.

37.

McClintock

Hall

Oliver

(2003) Risk markers associated with challenging behaviours in people with developmental disabilities: a meta-analytic study. Journal of Intellectual Disability Research 47: 405–416.

38.

McKean

Mensah

Eadie

et al . (2015) Levers for language growth: characteristics and predictors of language trajectories between 4 and 7 years. PLoS ONE 10: e0134251.

39.

Magiati

Moss

Charman

et al . (2011) Patterns of change in children with autism spectrum disorders who received community based comprehensive interventions in their pre-school years: a seven year follow-up study. Research in Autism Spectrum Disorders 5: 1016–1027.

40.

May

Sciberras

Brignell

et al . (2017) Autism spectrum disorder: Updated prevalence and comparison of two birth cohorts in a nationally representative Australian sample. BMJ Open 7: e015549.

41.

Mitchell

Brian

Zwaigenbaum

et al . (2006) Early language and communication development of infants later diagnosed with autism spectrum disorder. Journal of Developmental and Behavioral Pediatrics 27: S69–S78.

42.

Norrelgen

Fernell

Eriksson

et al . (2014) Children with autism spectrum disorders who do not develop phrase speech in the preschool years. Autism 19: 934–943.

43.

Paul

Campbell

Gilbert

et al . (2013) Comparing spoken language treatments for minimally verbal preschoolers with autism spectrum disorders. Journal of Autism and Developmental Disorders 43: 418–431.

44.

Paul

Chawarska

Cicchetti

et al . (2008) Language outcomes of toddlers with autism spectrum disorders: a two year follow-up. Autism Research 1: 97–107.

45.

Pellicano

(2010) The development of core cognitive skills in autism: a 3-year prospective study. Child Development 81: 1400–1416.

46.

Pickett

Pullara

O’Grady

et al . (2009) Speech acquisition in older nonverbal individuals with autism: a review of features, methods, and prognosis. Cognitive and Behavioral Neurology 22: 1–21.

47.

Pickles

Anderson

Lord

(2014) Heterogeneity and plasticity in the development of language: a 17-year follow-up of children referred early for possible autism. Journal of Child Psychology and Psychiatry, and Allied Disciplines 55: 1354–1362.

48.

Randall

Sciberras

Brignell

et al . (2016) Autism spectrum disorder: presentation and prevalence in a nationally representative Australian sample. Australian and New Zealand Journal of Psychiatry 50: 243–253.

49.

Rapin

Dunn

Allen

et al . (2009) Subtypes of language disorders in school-age children with autism. Developmental Neuropsychology 34(1): 66–84.

50.

Rescorla

(2011) Late talkers: do good predictors of outcome exist? Developmental Disabilities Research Reviews 17: 141–150.

51.

Rose

Trembath

Keen

et al . (2016) The proportion of minimally verbal children with autism spectrum disorder in a community-based early intervention programme. Journal of Intellectual Disability Research 60: 464–477.

52.

Rothman

(2003) An Australian Version of the Adapted PPVT-III for Use in Research. Melbourne, VIC, Australia: Australian Council for Educational Research.

53.

Russell

Rodgers

Ford

(2013) The strengths and difficulties questionnaire as a predictor of parent-reported diagnosis of autism spectrum disorder and attention deficit hyperactivity disorder. PLoS ONE 8: e80247.

54.

Sandbank

Woynaroski

Watson

et al . (2017) Predicting intentional communication in preverbal preschoolers with autism spectrum disorder. Journal of Autism and Childhood Schizophrenia 47: 1581–1594.

55.

Sigafoos

(2000) Communication development and aberrant behavior in children with developmental disabilities. Education and Training in Mental Retardation and Developmental Disabilities 35: 168–176.

56.

Sigman

McGovern

(2005) Improvement in cognitive and language skills from preschool to adolescence in autism. Journal of Autism and Developmental Disorders 35: 15–23.

57.

Snowling

Duff

Nash

et al . (2016) Language profiles and literacy outcomes of children with resolving, emerging, or persisting language impairments. Journal of Child Psychology and Psychiatry 57: 1360–1369.

58.

Soloff

Lawrence

Misson

et al . (2006) Wave 1 Weighting and Non-Response. Melbourne, VIC, Australia: Australian Institute of Family Studies.

59.

Steele

Joseph

Tager-Flusberg

(2003) Brief report: developmental change in theory of mind abilities in children with autism. Journal of Autism and Developmental Disorders 33: 461–467.

60.

Szatmari

Bryson

Streiner

et al . (2000) Two-year outcome of preschool children with autism or Asperger’s syndrome. American Journal of Psychiatry 157: 1980–1987.

61.

Tager-Flusberg

(2006) Defining language phenotypes in autism. Clinical Neuroscience Research 6: 219–224.

62.

Tager-Flusberg

(2016) Risk factors associated with language in autism spectrum disorder: clues to underlying mechanisms. Journal of Speech, Language, and Hearing Research 59: 143–154.

63.

Taylor

Christensen

Lawrence

et al . (2013) Risk factors for children’s receptive vocabulary development from four to eight years in the longitudinal study of Australian children. PLoS ONE 8: e73046.

64.

Tek

Mesite

Fein

et al . (2013) Longitudinal analyses of expressive language development reveal two distinct language profiles among young children with autism spectrum disorders. Journal of Autism and Developmental Disorders 44: 75–89.

65.

Thomas

(2009) Analyses of Five Years of Results from Project Data (Developmentally Appropriate Treatment for Autism), a High-Quality Early Childhood Special Education Program for Students with Autism Spectrum Disorders. Seattle, WA: University of Washington.

66.

Thurm

Lord

Lee

et al . (2007) Predictors of language acquisition in preschool children with autism spectrum disorders. Journal of Autism and Developmental Disorders 37: 1721–1734.

67.

Thurm

Manwaring

Swineford

et al . (2015) Longitudinal study of symptom severity and language in minimally verbal children with autism. Journal of Child Psychology and Psychiatry, and Allied Disciplines 56: 97–104.

68.

Toth

Munson

Meltzoff

et al . (2006) Early predictors of communication development in young children with autism spectrum disorder: joint attention, imitation, and toy play. Journal of Autism and Developmental Disorders 36: 993–1005.

69.

Ukoumunne

Wake

Carlin

et al . (2012) Profiles of language development in pre-school children: a longitudinal latent class analysis of data from the Early Language in Victoria Study. Child: Care, Health and Development 38: 341–349.

70.

Varni

Seid

Kurtin

(2001) PedsQL™ 4.0: reliability and validity of the Pediatric Quality of Life Inventory™ Version 4.0 Generic Core Scales in healthy and patient populations. Medical Care 39: 800–812.

71.

Wechsler

(2003) Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV). San Antonio, TX: Psychological Corporation.

72.

Whitehouse

Barry

Bishop

(2008) Further defining the language impairment of autism: is there a specific language impairment subtype? Journal of Communication Disorders 41(4): 319–336.

73.

Wodka

Mathy

Kalb

(2013) Predictors of phrase and fluent speech in children with autism and severe language delay. Pediatrics 131: e1128–e1134.

74.

Zubrick

Taylor

Rice

et al . (2007) Late language emergence at 24 months: an epidemiological study of prevalence, predictors, and covariates. Journal of Speech, Language, and Hearing Research 50: 1562–1592.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.44 MB

0.14 MB