Temperament influences the relationship between symptom severity and adaptive functioning in children with autism spectrum disorder

Abstract

Temperament is a construct that is relatively stable over time but varies between individuals. Research suggests that children with autism spectrum disorder have a ‘reactive’ temperament profile when compared to peers with or without disabilities. However, our understanding of how temperament varies within children with autism and how it relates to child symptoms and outcomes is limited. This study aimed to (a) explore the variation of individual temperament traits within a sample of school-aged children with autism to determine whether subgroups of children with similar trait profiles emerge and (b) examine whether temperament influences the relationship between autism symptoms and adaptive functioning outcomes. Results revealed that children with autism can be classified empirically into two distinct profiles – ‘Even’ and ‘Reactive’ temperaments. Correlational and hierarchical regression analyses indicated that both temperament profiles and baseline symptom severity predicted adaptive functioning outcomes 1 year later. There was a significant interaction between temperament and symptom severity, suggesting temperament can influence the impact of increasing symptom severity on adaptive functioning skills in children with autism. Study findings highlight the importance of considering temperament in understanding the individual differences that influence the development of daily functioning and developmental outcomes in children with autism.

Lay Abstract

Temperament is often thought of as behavioural traits that are relatively stable over time but can vary between individuals. Children diagnosed with autism spectrum disorder are often characterized as having ‘reactive’ and ‘negative’ temperaments when compared to same-aged peers with or without disabilities, which can negatively impact the development of adaptive functioning skills but little is known about variations of temperament between individual children diagnosed with autism spectrum disorder. This study aimed to (a) explore the variation of individual temperament traits within a sample of school-aged children with autism spectrum disorder to determine whether subgroups with similar trait profiles emerge and (b) examine whether temperament influences the relationship between autism symptoms and adaptive functioning outcomes. Results from our dataset suggest that children diagnosed with autism spectrum disorder fit under two profiles: ‘even’ and ‘reactive’. Furthermore, our analysis shows that temperament can influence the impact of increasing symptom severity on adaptive functioning skills in children with autism spectrum disorder. Study findings highlight the importance of considering temperament when trying to understand the individual differences that influence the development of functioning and developmental outcomes in children with autism spectrum disorder.

Keywords

adaptive functioning autism temperament

Autism spectrum disorder (ASD or autism) is a complex neurodevelopmental disorder characterized by symptoms related to impaired social communication and the presence of restricted and repetitive behavioural patterns (American Psychiatric Association [APA], 2013). Although general impairment often persists throughout the lifespan, the clinical presentation of ASD is heterogeneous and changes as children develop (Georgiades et al., 2013, 2017; Gotham et al., 2012; ManrNue-Vallier, 2014; Szatmari et al., 2015).

Researchers have begun to recognize the importance of expanding the focus of investigation beyond autism symptoms and deficits to include variables potentially associated with individual outcomes (Georgiades & Kasari, 2018; Henninger & Taylor, 2012). One outcome of particular interest and relevance to individuals with ASD and their families is the development of day-to-day adaptive functioning skills (Bal et al., 2017; Szatmari et al., 2015). Studies suggest that adaptive functioning in children with ASD – which includes communication, daily living and socialization skills – are influenced by factors such as access to early diagnosis and intervention (MacDonald et al., 2014; Reichow et al., 2012; Smith, 1999), cognitive functioning (Farley et al., 2009), language development (Bennett et al., 2015) and symptom severity (Gotham et al., 2012; Szatmari et al., 2015).

In a large inception cohort study, Szatmari et al. (2015) demonstrated that the trajectories of autism symptom severity and adaptive functioning skills are only moderately ‘yoked’. Specifically, it is possible for some children to show improvement in adaptive functioning skills even if their autism symptoms remain severe over that same period. It is also possible for other children to exhibit a reduction in autism symptom severity, without noticeable improvement in their adaptive functioning skills. These findings suggest that there may be other factors, at the individual child and/or contextual (i.e. environmental) level, that influence the association between autism symptomology and adaptive functioning. Understanding the role of these factors may help inform clinical decision-making regarding interventions, tailored to the needs of specific subgroups of children within the heterogeneous autism spectrum.

Temperament is defined as a set of individual behavioural tendencies (Goldsmith & Alansky, 1987) that are shaped by genetic predispositions and environmental input (Rothbart & Derryberry, 1981; Rothbart et al., 1995). Temperament is a construct that is potentially relevant to a child’s functioning and has been gaining attention within the autism research field (Burrows et al., 2016; Chetcuti et al., 2019; De Pauw & Mervield, 2010; Garon et al., 2009, 2016; Zwaigenbaum et al., 2005). According to Rothbart’s theoretical framework, temperament reflects an individual’s reactivity and self-regulation and is manifested in the ways an individual adjusts or responds to changes in the environment (Rothbart et al., 1995). For example, variability in underlying physiological systems, such as excitability, heart rate and other biological arousal responses, influences an individual’s emotional reactions, which in turn guide behavioural tendencies. Conversely, self-regulation, which includes both effortful control and attention regulation, governs and modulates reactivity (Rothbart & Bates, 2006; Rothbart et al., 1995).

Temperament is variable between individuals but appears to remain relatively stable across the lifespan within individuals (Kagan, 1971; Rothbart, 1986; Rothbart et al., 1995). In the general population, the connection between temperament traits and a range of important life outcomes – including relationships, academics, health, psychopathology and adaptive functioning – has been well-documented (Zentner & Shiner, 2015). Across the lifespan, temperament influences the physical and emotional development of an individual and shapes how individuals interpret and navigate the social world (Rothbart & Derryberry, 1981). This makes temperament a construct worth exploring when trying to understand the individual differences that influence the development of day-to-day functioning and developmental outcomes in children with autism.

Within the same theoretical framework, temperament is measured as variation among dimensions that relate to reactivity and self-regulation. These dimensions include activity level, frustration, assertiveness/dominance, attention focusing, inhibitory control, perceptual sensitivity, sadness, shyness and soothability (Fox & Henderson, 1999; Rothbart & Derryberry, 1981). Furthermore, covariation among these dimensions forms three higher order factors: surgency (i.e. activity level, sociability and pleasure expressed in anticipation of reward or during high-intensity activities), negative affectivity (i.e. anger, sadness, fear, physical discomfort and recovery from distress) and effortful control (i.e. an ability to focus attention, demonstrate satisfaction with low-intensity activities and exercise inhibitory control; Derryberry & Rothbart, 1997; Putnam et al., 2001, 2008). Although several models/frameworks of temperament have been suggested (see Chess & Thomas, 1987; Goldsmith & Alansky, 1987), agreement is seen on the underlying temperament constructs in children (Gosling et al., 2003; Shiner et al., 2012). Rothbart’s framework is often used in autism studies because it incorporates the role of attention and captures the relationship between temperament traits and the development of functional outcomes in children with autism.

Temperament in autism

Although temperament profiles have been well documented in typically developing children, this line of work has only recently appeared in the field of ASD (Hepburn & Stone, 2006; Konstantareas & Stewart, 2006). Over the past decade, most of the temperament research in children diagnosed with autism has focused on parent-reported measures collected during middle childhood. However, most of the studies compared temperament between children with or at risk of having a diagnosis of ASD in comparison to control groups such as typically developing peers (Bryson et al., 2007; Clifford et al., 2013) and children with other developmental disabilities (Bradley et al., 2004; Bradley & Isaacs, 2006), rather than investigating temperament variations within a sample of children with ASD. In these studies, children with autism, the infant siblings of children with autism and children who are at risk for autism are typically characterized as having more ‘Reactive’ temperament traits. Such traits include relatively low surgency (e.g. low social motivation, less social orienting, fewer anticipatory gestures and less smiling and laughing), low effortful control (e.g. difficulty disengaging attention, lower attentional focus, lower attention shifting and lower adaptability) and high negative affectivity (e.g. more dysregulation, higher reactivity, more distress and less soothability) when compared to same-aged neurotypical controls and same-aged peers with other developmental disorders (Bradley & Isaacs, 2006; Burrows et al., 2016; De Pauw et al., 2011; Schwartz et al., 2009). To date, only a few studies have noted temperament as a multidimensional construct and examined ways to better compare variation of temperament within middle childhood and youth with autism (Burrows et al., 2016; De Pauw et al., 2011; Kasari & Sigman, 1997; Konstantareas & Stewart, 2006).

Importantly, variability in temperament traits is often strong predictors of deviations from typical childhood development. Studies have suggested that certain temperament traits relating to effortful control are highly correlated to the development trajectories of executive functioning skills and self-regulation. Interestingly, research suggests that extremes in temperament including increases in negative affect, reduction in traits relating to surgency and effortful control and increases in perceptual sensitivity are among the earliest emerging signs of risk for autism and often detected by parents well before the age of diagnosis (Clifford et al., 2013). Studies of high-risk infant siblings have shown that having a specific profile of temperament traits (e.g. where patterns of traits become ‘temperament profiles’) characterized by high negative affect, low surgency and low effortful control may predict a later autism diagnosis (Clifford et al., 2013; Garon et al., 2009). In the study by Garon et al. (2009), infants were assessed at 24 months and followed until 36 months of age. Of these infants, some were (a) high-risk infants with an older sibling with autism who did not meet criteria for an autism diagnosis at 36 months (Non-ASD Siblings), (b) high-risk infants with an older sibling with autism who received a diagnosis of autism at 36 months (ASD Siblings) and (c) low-risk infants with no known first- or second-degree relatives with autism (Controls). Using parental reports of temperament using Rothbart’s model, this study reported three distinct temperament profiles at 24 months. Infants in the ASD Siblings group had lower positive anticipation, higher activity level and lower attentional shifting compared to Non-ASD Siblings and Controls. Temperament profiles of the ASD Siblings group at 24 months also predicted an ASD diagnosis at 36 months. The two sibling groups (ASD Siblings and Non-ASD Siblings) were characterized as having lower positive affect, poorer regulation of negative emotions, and more difficulty with attentional control when compared to the Controls.

Taken together, research suggests that certain temperament profiles can reflect phenotypic differences between controls and high-risk siblings as well as differences within high-risk siblings. Further research is needed to examine the association and potential interaction between the construct of temperament and autism symptoms, and their combined impact on functional outcomes in children diagnosed with autism (see Szatmari et al., 2015). Some dimensions of temperament – including attentional focusing and shifting difficulties, poor inhibitory and executive control, difficulties with soothing and high reactivity – have direct conceptual relevance to the core features of autism symptomatology. However, this does not necessarily mean that variability in diagnostic symptoms captures the variability of temperament traits in children with autism. Emerging evidence suggests that temperament is a transdiagnostic factor that contributes to the heterogeneity in individual outcomes (Chetcuti et al., 2019). The identification of temperament profiles can be useful in helping clinicians select individualized interventions for reducing symptom severity and increasing functioning (Muris & Ollendick, 2005).

Temperament and adaptive functioning

Temperament has been identified as an important factor in predicting the development of adaptive functioning skills including communication skills, daily-living skills and social skills (Blair et al., 2004; Fabes et al., 1999). A strong foundation of adaptive functioning skills is essential for successful navigation of an individuals’ environment, and therefore, these skills are typically an area of interest for parents and an area of focus for interventions with children on the spectrum. For example, studies have highlighted the correlation between temperament traits and their relationship with high effortful control (e.g. executive functioning skills including self-regulation, inhibiting inappropriate behaviour, ability to focus and shift attention and ability to follow through with tasks and actions) to the development of social communication skills (Spinrad et al., 2004) and the development of social competence (Eisenberg & Spinrad, 2004; Masten & Coatsworth, 1998; Riggs et al., 2006). Traits associated with negative affect and surgency (e.g. approach and avoidance behaviour, impulsivity and difficulty to sooth) have been associated with high externalizing behaviours, which relate to behaviours associated with challenges with rule following, low frustration tolerance and aggression, as well as difficulties with understanding and following social norms (Checa et al., 2008; Clasen & Brown, 1985; Nelson et al., 1999; Schwartz et al., 2006). Thus, even in neurotypical children and youth, variation in temperament profiles impacts the development of adaptive functioning skills. The development of good adaptive functioning skills can later on contribute to successful school readiness and adjustment, as well as academic success and better quality of life (Blair & Razza, 2007; Carey, 1998; Checa & Rueda, 2011; Ellis et al., 2004). However, little is known about how variation in temperament traits may influence individual differences in the development of basic adaptive functioning skills among children and youth with autism. One study by Schwartz et al. (2009) suggests that a relationship exists between temperament, the presence of autism symptomology and adaptive functioning outcomes. In this study, they demonstrated using self-reported measures of temperament and parental reports of adaptive functioning skills that individual variations in temperament characteristics predicted individual differences in autism symptomology, social skills and social–emotional outcomes in high-functioning youth diagnosed with autism. Given the relationship between autistic symptomology and adaptive functioning skills, this finding suggests that it would be important to consider how temperament may impact this relationship. For example, it is possible that autistic symptomology (i.e. difficulties with responsiveness and self-regulation) may be exacerbated by temperament difficulties (i.e. low effortful control, high negative affect and surgency) that then may be related to challenges in adaptive skill development. These relationships, however, have yet to be explored.

Current study

In the current study, we explored the variation of individual temperament traits within a sample of school-aged children with autism to determine whether groups of children with similar trait profiles emerge. Previous studies have noted variation in temperament traits within samples of children with autism and aged-matched peers (Kasari & Sigman, 1997; Konstantareas & Stewart, 2006) but few have investigated individual differences in temperament within a sample of school-aged children with autism without making comparisons to a neurotypical control group (see Burrows et al., 2016; De Pauw et al., 2011). Given the growing literature suggesting that temperament traits and associated behaviours have an impact on the development of adaptive functioning skills, the second aim of this study was to explore the influence temperament may have on the association between autism symptom severity and adaptive functioning outcomes, while controlling variables highly correlated with adaptive functioning outcomes including cognitive ability, age, sex and age of diagnosis. Given previous findings that the relationship between symptom severity (i.e. first level of identification and categorization when children with autism enter the service system) and adaptive functioning skills (usually an outcome variable of interest) is not always linear (see Szatmari et al., 2015), we were interested in exploring other variables that might influence this relationship, including temperament. The systematic investigation of temperament has the potential to inform our understanding of heterogeneity and individual differences within the autism spectrum.

Methods

Participants

Study participants (N = 421, 84.3% boys) came from the Pathways in ASD study, a Canadian longitudinal study of an inception cohort of newly diagnosed children with ASD (see Szatmari et al., 2015). From this sample, 185 children included in the cluster analysis (n = 11) had missing data and were excluded, and 150 children had data from all relevant measures to be included in the final hierarchical regression. Children’s diagnoses were determined using clinical judgement by a team with diagnostic expertise according to the criteria set out by the Diagnostic and Statistical Manual of Mental Disorders (Text Revised, 5th ed; DSM-TR-5; APA, 2013) using the Autism Diagnostic Observation Scale (ADOS, Lord et al., 2000) and the Autism Diagnostic Interview–Revised (ADI-R; Rutter et al., 2003). Children were excluded from the study if they had a known genetic syndrome or neurological disorder, severe visual or hearing impairments and/or had parents who did not speak English or French sufficiently for completion of data collection for all the study measures. Informed consent was obtained from parents and assent from children who were deemed capable. Ethics approval was obtained from each institution’s research ethics board.

For the purposes of this study, data from assessments at four time points during the preschool and early school years were used: diagnosis/baseline (T1), age 6 years (T2), age 7 years (T3) and age 8 years (T4). Demographic information at baseline was used to describe the sample including sex distribution, age of diagnosis, autism severity and maternal level of education (Table 1). Data on ASD symptom severity at T2, child temperament at T3 and adaptive functioning level at T4 were used to address the main study objectives.

Table 1.

Descriptive statistics and comparison between completed TMCQ data and missing TMCQ data group at baseline (T1).

Variable	Complete TMCQ data (n = 185)	Missing TMCQ data (n = 236)	Effect size of difference
Variable	M (SD)	M (SD)	(Cohen’s d)
Sex (%)
Male	83	85
Female	17	15
Age at diagnosis (months)	38.52 (8.92)	37.93 (8.46)	0.07
ADOS severity score T1	7.72 (1.61)	7.44 (1.77)	0.16
Maternal education (years)	6.24 (2.46)	5.82 (2.67)	0.16

ADOS: Autism Diagnostic Schedule; TMCQ Total Score: Temperament in Middle Childhood Questionnaire; Highest Maternal Education Level (1 = some high school, 2 = completed high school, 3 = some trade, tech or vocational training, 4 = some community college, 5 = some university, 6 = diploma or certificate from a community college, CEGEP or nursing school or university).

Measures

Autism symptom severity

The Autism Diagnostic Observation Schedule (ADOS; Lord et al., 2000, 2002) is a semi-structured play-based assessment that provides standardized contexts in which to elicit specific behaviour. The ADOS consists of four modules that are administered depending on an individual’s expressive language level. The activities in the modules are designed to allow an examiner to observe symptoms related to the DSM-5 diagnostic criteria for ASD (APA, 2013). Items are scored from 0 (no autism-specific abnormality present) to 3 (extreme abnormality indicative of autism). An overall severity metric, calculated from converted ADOS raw scores, ranges from 0 (no impairment) to 10 (most severe; Gotham et al., 2012).

Adaptive functioning

The Vineland Adaptive Behaviour Scales, Second edition (VABS-II; Sparrow et al., 2005) is a semi-structured interview administered to the parent or caregiver. The VABS-II assesses adaptive behaviour skills that include communication, socialization, daily-living skills and motor functioning, resulting in an overall index of daily functioning called the Adaptive Behaviour Composite (ABC). The VABS-II has high interrater reliability across domains (0.71–0.81; Sparrow et al., 2005). Previous studies have demonstrated that the VABS-II is a strong measure of overall adaptive functioning level in children with ASD (Ashwood et al., 2015). The ABC was used in the current analyses.

Cognitive ability

The Merrill-Palmer-Revised Scales of Development (M-P-R: Roid & Sampers, 2004) is an individually administered measure of intellectual ability for children aged 2 to 78 months. For the current analysis, we used the Developmental Index (DI), comprising Cognitive, Receptive Language and Fine Motor scales. The M-P-R is a standardized measure that is ideal for screening infants and children for early identification of developmental delays and learning difficulties. The measure has been validated for use in children diagnosed with autism and has demonstrated good concurrent validity (Dempsey et al., 2018) with other standardized cognitive (Wechsler Intelligence Scale; 5th ed.) and language (Preschool Language Scale–4th ed.) measures.

Temperament

The Temperament in Middle Childhood Questionnaire (TMCQ; Simonds & Rothbart, 2009) is a 157-item parent-report questionnaire that assesses temperament in children aged 7 to 10 years. The measure was adapted from and has convergent validity with the Children’s Behaviour Questionnaire (CBQ; Rothbart et al., 2001) and the Childhood Temperament and Personality Questionnaire (CTPQ: Victor et al., 2003). The questionnaire uses a 5-point scale, ranging from ‘Almost always untrue’ to ‘Almost always true’ with ‘Does not apply’ as an additional option. The TMCQ assesses 17 dimensions of temperament comprising activity level, affiliation, anger/frustration, assertiveness/dominance, attentional focusing, discomfort, fantasy/openness, fear, high-intensity pleasure, impulsivity, inhibitory control, low-intensity pleasure, perceptual sensitivity, sadness, shyness, soothability/falling reactivity and activation control. Internal validity was reported as adequate (Cronbach’s α > 0.70) for all temperament subscales for our study (see Nystrom & Bengtsson, 2017, for psychometric analysis of the TMCQ).

Challenging behaviours

The Child Behaviour Checklist (Achenbach & Edelbrock, 1983) for ages 1.5 to 5 (CBCL: 1.5–5 years) is a standardized parent-report measure of externalizing and internalizing behaviour problems in preschool children. The questionnaire consists of 99 items that use a 3-point scale (0 = not true, 1 = somewhat/sometimes true, 2 = very/often true). The externalizing scale consists of items that load onto a measure of hyperactivity, attention problems and aggressive behaviours. The internalizing scale consists of items related to emotionally reactive behaviours, anxious/depressed and withdrawal behaviours. The CBCL has been validated for use to measure problem behaviours in children diagnosed with ASD (Pandolfi et al., 2009).

Statistical analysis

Cluster analysis

Data analysis was conducted using IBM SPSS Version 25 (IBM Corp., 2019). To identify whether distinct temperament profiles emerge within our ASD sample (Aim 1), we conducted a hierarchical cluster analysis using the scores for the 17 dimensions of the TMCQ at T3, the only time point at which temperament data were collected in our sample. A cluster analysis was deemed appropriate for the current study because we were interested in using the temperament dimensions to identify empirically homogeneous groups of children with similar temperamental profiles. Recent studies have emphasized that appropriateness of using cluster analysis or other ‘person-centred’ methods to study temperament because ‘unlike variable-centred regression analysis, they take into account the nonorthogonal nature of temperament traits . . . which may be critical in seeking to draw conclusions about predictive associations’ (Chetcuti et al., 2019, p. 227). Furthermore, cluster analysis was used because it does not assume that underlying latent variables account for any associations between observed variables; rather this approach aims to find similarities between variables that are not known in advance and is also capable of handling a larger number of variables compared to other similar analytic approaches (e.g. latent class analysis (LCA)). Given the exploratory nature of our analysis and the absence of a priori expectations in group characteristics within and between children, a sequential cluster analysis was selected to address the first study objective (see Bartholomew & Knott, 1999, for more information on cluster analysis vs LCA).

The cluster analysis used the Ward minimum variance method to establish temperament profiles; the dissimilarity measure selected for the interval data was squared Euclidean distance. The optimal number of clusters was selected using a scree plot (or elbow criterion) and the agglomeration schedule. Discriminant function analysis (DFA) was used to determine the degree of association of the 17 dimensions to cluster membership.

Differences between the clusters were characterized using other variables of interest including age at diagnosis, autism symptom severity at T1, cognitive ability at T1, adaptive functioning at both T1 and T4 and total behavioural problems at T1.

Hierarchical regression analysis

Preliminary analysis

The second aim was to investigate the effect of temperament profile on the relationship between autism symptom severity and adaptive functioning. Bivariate correlation was used to examine the associations between five predictor variables – age of diagnosis, sex, cognitive ability, autism symptom severity and temperament profile – and the outcome variable, adaptive functioning. A hierarchical regression analysis was conducted by initially entering autism symptom severity followed by temperament profile, controlling for age of diagnosis, sex and cognitive ability. An interaction term of temperament profile and autism symptom severity was calculated after symptom severity scores were mean-centred to overcome issues with multicollinearity (Aiken et al., 1991). Simple slopes were then calculated to demonstrate the relations between autism symptom severity and adaptive functioning via temperament profiles.

Results

Of the 421 children enrolled in the Pathways study at baseline (T1), an assessment of temperament using the TMCQ was completed by the parents of 196 children (46%) at T3. Of the 196 children, 11 were excluded because one or more relevant scores were missing from their TMCQ questionnaire, leaving the final sample of 185 (83.7% boys) with complete data on all variables of interest. The children with missing TMCQ data at T3 did not significantly differ from the children with complete data on age of diagnosis (p = 0.49), ASD symptom severity at T1 (p = 0.09) or maternal education level (p = 0.11), providing evidence that missing data were unlikely to bias the results (see Table 1).

Cluster analyses

The agglomeration schedule and the scree plot suggested that a two-cluster solution provided the best fit to the data. Results of the DFA based on the significant Box M test (F(153, 73,927.55) = 277.30, p < 0.001) and the small Wilk’s λ value (0.391, chi-square (17) = 163.917, p < 0.001) indicated that within-group variation was small but a statistically significant difference in mean scores was found between the two clusters, which satisfies the nature of the cluster solution. Compared to children in Cluster 1, children in Cluster 2 (n = 73, 39.4% of sample) had higher scores on activation control, affiliation, assertiveness, attentional focusing, fantasy/openness and soothability and lower scores on anger/frustration, impulsivity, discomfort, fear, sadness and shyness. Based on these descriptions, Cluster 1 was labelled ‘Reactive’ Temperament to reflect its similarities with the components of Negative Affect and Cluster 2 was labelled ‘Even’ Temperament because of its similarities to components of Effortful Control (see Figure 1). Children with ‘Even’ Temperament also scored significantly higher on the TMCQ scores related to activation control, affiliation, assertiveness and dominance, attention focusing, fantasy and openness, high-intensity pleasure seeking, inhibitory control, perceptual sensitivities and soothability and reactivity compared to children with ‘Reactive’ Temperament. In contrast, children with ‘Reactive’ Temperament (n = 112, 60.5% of the sample) had significantly higher scores on anger and frustration, discomfort, fear, impulsivity, sadness and shyness (see Table 5 in Appendix 1).

Figure 1.

Plot depicting the temperament profiles of ‘Even’ Temperament and ‘Reactive’ Temperament profile along the 17 domains within the TMCQ. 1. Activation control; 2. Activity level; 3. Affiliation; 4. Assertiveness/dominance; 5. Attention focusing; 6. Fantasy/openness; 7. High-intensity pleasure; 8. Inhibitory control; 9. Low-intensity pleasure; 10. Perceptual sensitivity; 11. Soothability/falling reactivity; 12. Anger/frustration; 13. Discomfort; 14. Fear; 15. Impulsivity; 16. Sadness; 17. Shyness.

Children in the two clusters were identified as belonging to one of the two temperament profiles: ‘Even’ or ‘Reactive’. Children in the two temperament profiles did not differ significantly on age of diagnosis (F(1, 184) = 0.50, p = 0.81) or autism symptom severity at T1 (F(1, 184) = 0.21, p = 0.64). However, the two profiles differed in their cognitive ability, emotional and behavioural problems and adaptive functioning at T1. Specifically, children in the ‘Even’ temperament profile scored significantly higher on cognitive ability at T1 (F(1, 167) = 6.54, p = 0.01) and had significantly higher adaptive functioning composite scores at T1 (F(1, 181) = 18.23, p = 0.001) and at T4 (F(1, 136) = 38.34, p = 0.001). Children with a ‘Reactive’ temperament profile scored higher on the overall CBCL Externalizing Behavioural Total Score (F(1, 133) = 6.44, p = 0.01), compared to children with an ‘Even’ temperament profile (see Table 2).

Table 2.

Descriptive characteristics of ‘Even’ temperament profile and ‘Reactive’ temperament profile at baseline (T1) and Total Behavioural Problems Score at T4.

Variables	Even temperamentM (SD)(n = 73)	Reactive temperamentM (SD)(n = 112)	Effect size of difference (Cohen’s d)
Age at diagnosis	38.20 (7.88)	38.49 (8.68)	0.02
Symptom Severity (ADOS) T1	7.75 (1.59)	7.72 (1.59)	0.08
Cognitive Ability (M-P-R) T1	64.30 (23.34)	55.02 (23.08)*	0.31
Adaptive Functioning (VABS-II) at T1	76.38 (8.89)	70.02 (10.47)**	0.65
Adaptive Functioning (VABS-II) at T4	85.45 (10.68)	71.47 (14.15)**	1.12
Total Behavioural Problems T1 (CBCL)	51.07 (9.93)	61.05 (9.71)***	1.01

ADOS: Autism Diagnostic Observation Schedule; M-P-R: Merrill-Palmer-Revised Scales of Development; TMCQ Total Score: Temperament in Middle Childhood Questionnaire; VABS-II Composite Score: Vineland Adaptive Behavioural Scales–2nd ed.; CBCL: Childhood Behavioural Checklist.

p < 0.05. **p < 0.01. ***p < 0.001.

Hierarchical regression analysis

Of the 185 children used in the cluster analysis, 35 children had to be excluded from the regression analysis because they did not have a complete Merrill-Palmer (Cognitive Abilities) at T1, ADOS-II at T3 and VABS-II at T4. The group of 35 children did not differ from the group of 150 children used in the regression analysis by age of diagnosis or sex at baseline. The distribution of children from the two temperament profiles was not different between the group of children included in the regression analysis and the group that had missing data (see Table 6 Appendix 1).

We initially examined the bivariate correlations among child variables (i.e. age of diagnosis, sex and cognitive ability), temperament profile membership, autism symptom severity and adaptive functioning outcomes, using Spearman’s rho. As shown in Table 3, adaptive functioning was positively correlated with cognitive ability (r = 0.49, p < 0.001) but negatively correlated with temperament profile (r = −0.27, p < 0.001) and ASD symptom severity (r = −0.46, p < 0.01). Temperament profile was negatively correlated with cognitive ability (r = −0.19, p < 0.01) and ASD symptom severity (r = −0.23, p < 0.01).

Table 3.

Correlation measures between age of diagnosis at T1, sex at T1, cognitive ability at T2, autism severity score at T2, temperament profile at T3 and adaptive functioning at T4.

Measure	1	2	3	4	5	6
1. Age at diagnosis (months)	–
2. Sex	−0.08	–
3. Cognitive ability (M-P-R)	0.24**	−0.01	–
4. Autism symptom severity (ADOS)	−0.00	−0.09	−0.14	–
5. Temperament profile (TMCQ)	0.03	−0.04	−0.19*	−0.23**	–
6. Adaptive functioning (VABS-II)	0.09	−0.00	0.49**	−0.46**	−0.27**	–

M-P-R: Merrill-Palmer-Revised Scales of Development; ADOS: Autism Diagnostic Observation Schedule; TMCQ: Temperament in Middle Childhood Questionnaire; VABS-II: Vineland Adaptive Behavioural Scales–2nd ed.

An Even Temperament profile was coded as 2 and a Reactive Temperament was coded as 1.

p < 0.05; **p < 0.01 (two-tailed).

Association between temperament, ASD symptom severity and adaptive functioning

A hierarchical multiple variable regression analysis was conducted to confirm that symptom severity and cognitive ability would predict adaptive functioning outcomes in the sample. As shown in Table 4, child age at diagnosis, sex, cognitive functioning and autism symptom severity were entered in the first step, with an overall model that was significant, F(4, 145) = 14.21, p < 0.001 and accounted for 28% of the variance in adaptive functioning outcomes. Temperament profile was entered in the second step. The overall model was significant, F(5, 144) = 19.29, p < 0.001 and accounted for an additional 12% of the variance. Temperament profile membership emerged as a significant predictor of adaptive functioning outcome (Table 4), controlling for ASD symptom severity and other potentially confounding variables.

Table 4.

Hierarchical multiple regression of predictors and moderator relations with adaptive functioning.

Variable	Model 1					Model 2					Model 3
Variable	B	SE	95% CI	t	p value	B	SE	95% CI	t	p value	B	SE	95% CI	T	p value
Constant	73.46	6.29	61.91 to 85.90	11.66	0.00	84.63	6.13	72.51 to 96.76	13.79	0.00	100.85	9.77	81.53 to 120.16	10.3	0.00
Age at diagnosis	−0.35	0.10	−0.26 to 0.18	−0.30	0.75	−0.02	0.10	−0.22 to 0.17	−0.24	0.80	−0.03	0.99	−0.22 to 0.16	−0.27	0.79
Sex	−0.65	22.27	−5.16 to 3.84	−0.288	0.77	0.07	2.09	−4.96 to 4.20	0.34	0.97	0.52	2.08	−3.58 to 4.63	0.25	0.80
Cognitive ability	0.23	0.04	0.16 to 0.31	6.44	0.00	0.21	0.03	0.14 to 0.27	6.10	0.00	0.21	0.03	0.14 to 0.86	6.36	0.00
ASD symptom severity	−1.15	0.42	−1.98 to –0.31	−2.72	0.01	−0.67	0.40	−1.45 to 0.11	−1.69	0.09	3.18	1.24	−5.64 to –0.71	2.54	.000
Temperament profile						8.81	1.64	5.68 to 12.29	5.36	0.00	−20.43	5.73	9.11 to 31.73	3.56	0.00
ASD symptom × temperament profile											1.66	0.78	0.11 to 321	2.11	0.05

R ²	0.28					0.40					0.41
F for change in R²	14.21***					28.74***					4.47*

p < 0.05. **p < 0.01. ***p < 0.001.

In the presence of main effects (steps 1 and 2), the third step in the hierarchical multiple variable regression analysis tested the hypothesis that temperament would act as a moderator and would influence the association autism symptom severity on adaptive functioning outcomes. As shown in Table 4, the product of autism symptom severity and temperament emerged as a significant predictor, t(4) = 2.11, p = 0.05, accounting for an additional 1% of variance beyond that explained by other variables, F(6, 143) = 17.21, p < 0.001. A plot of the interaction is shown in Figure 2 and shows that children with the ‘Even’ temperament profile tended to have decreasing adaptive functioning scores as their autism symptom severity increased (B = –1.51, SE = 0.53, t(6) = −281, p < 0.001) compared to children with the ‘Reactive’ temperament profile had relatively stable but lower adaptive functioning skills across all autism symptom severity scores.

Figure 2.

Plot illustrating effect of interaction between ASD symptom severity and temperament profile on adaptive functioning (with confidence intervals).

Discussion

This study adds to the literature highlighting the importance of considering individual traits such as temperament in the study of children diagnosed with autism. Our results demonstrate that there is variation in temperament profiles within a large sample of children diagnosed with autism. In subsequent analysis, two temperament profiles derived from clusters analysis were shown to have an influence on the association between autism symptom severity and adaptive functioning outcome, after controlling for age of diagnosis, sex and cognitive ability. Importantly, there was a main effect of temperament profile on adaptive functioning outcomes. As with most models including an interaction term, the inclusion of the term added little because most of the variance was already accounted for by the main effects (McLelland & Judd, 1993).

Results of the cluster analysis demonstrated that our sample of school-aged children with autism were best categorized into two temperament profiles: ‘Even’ and ‘Reactive’. Although previous research has reported that children with autism have more traits related to a ‘Reactive’ temperament profile compared to controls (Bryson et al., 2007; Clifford et al., 2013), our results suggest that temperament variation exists within our dataset and can be described using two profiles. For example, 39.4% of children in our sample exhibited traits associated with an ‘Even’ temperament, including increased regulatory control and reactivity, and children in this group scored higher on dimensions related to activation control, high attention focusing, inhibitory control, perceptual sensitivity and soothability. In contrast, children classified as having a ‘Reactive’ temperament exhibited lower levels of behavioural and emotional regulation and higher reactivity, including traits such as increased anger and frustration, discomfort and impulsivity. These findings are consistent with previous research that also identified dichotomous temperament profiles within an autism sample (Kasari & Sigman, 1997; Konstantareas & Stewart, 2006), but differ from other established models of temperament which group dimensions into higher order factors (Surgency, Effortful Control and Negative Affect) derived from variable-centred analyses (Derryberry & Rothbart, 1997), or Low Behavioural Approach and Effortful Emotion Regulation in infant siblings of children diagnosed with autism using DFA (Garon et al., 2009). It should be noted that the statistical approach to creating our clusters differed from these models, but there are similarities in the dimensions underlying our profiles with those most associated with two of the higher order factors, Effortful Control/Effortful Emotion Regulation and Negative Affect. For example, parents of children with a ‘Even’ temperament seem to report more behaviours associated with an ability to regulate emotions and behaviours, particularly having behaviours associated with better control of attention and executive functioning skills. In contrast, children within a ‘Reactive’ temperament profile had more behaviours associated with Negative Affect or Low Behavioural Approach (e.g. proneness to negativity, fear, anxiety, sadness and anger).

The cluster analysis in this study was a simple and efficient exploratory statistical approach to categorizing temperament profiles within our sample, but the consequences of which may limit our ability to identify the heterogeneity of temperament within the spectrum and our ability to interpret the full impact temperament has on relationship between autism symptoms and adaptive functioning outcome. Specifically, the dichotomous profiles that emerged from our analysis does not fully reflect the multidimensional nature of temperament. Although variable-centred analyses are beyond the scope of the current study, given the limitations stated above it would be imperative for future directions to focus on how our temperament profiles and their associated dimensions may align statistically with the three higher order factors (Derryberry & Rothbart, 1997).

Results indicated that temperament profiles at between 7 and 9 years of age predicted adaptive functioning outcomes, even after controlling for cognitive ability, age, sex, age of diagnosis, and autism symptom severity at the age of 6 years. This specific finding may help in the interpretation of our previous findings. Using data from the same inception cohort, we found that a child’s trajectory of adaptive functioning skills is not exclusively determined by the severity of the child’s autism symptoms at diagnosis (Szatmari et al., 2015). Similarly, in the present study, even though there were no differences on autism symptom severity at baseline (2–4 years of age) between temperament profiles, children with an ‘Even’ temperament profile had significantly higher adaptive functioning outcome scores at baseline and at the age of 6 years when compared to children with a ‘Reactive’ temperament profile. Moreover, temperament explained an additional 8% of the variance, over and above autism symptom severity, after accounting for child age of diagnosis, sex and cognitive ability. This suggests some unique contribution of temperament in explaining adaptive functioning outcomes in this sample of children with autism.

The interaction term created as a product of temperament and autism symptom severity was also a significant predictor of adaptive functioning outcome. The pattern of results indicate that children with an ‘Even’ temperament profile tended to have decreasing adaptive functioning scores as their autism symptom severity increased. This was in contrast to children with a ‘Reactive’ temperament profile whom had relatively stable but lower adaptive functioning skills across all autism symptom severity scores. This may mean that ‘reactive’ temperament traits in children with autism may interfere with adaptive skills development; however, the exact mechanism is unclear. This finding also suggests that special consideration to temperament may be warranted when designing intervention programmes for children with autism with focus on general supports which foster the development of adaptive functioning skills, especially for children with a more ‘reactive’ temperament.

Evidence suggests that child temperament and parenting behaviour interact overtime by shaping each other and changing interpersonal dynamics and may predict child outcomes (Hirschler-Guttenberg et al., 2015; Lee & Bates, 1995). Most notably, it has been suggested that temperament influences the quality and stability of parent–child interactions and attachment (van den Boom, 1994), although these relations have yet to be explored within autism. Evidence from this study suggest that the quality of the caregiver–child relationships may interact with temperament traits and change the quality and quantity of positive interactions between children with autism and adult caregivers, including parents, daycare workers, teachers and therapists. Although the mechanism of these interactions is not clear, one might hypothesize that children who have a profile that fits more with an ‘Even’ temperament (i.e. those who are more engaging, focused and easily soothed) experience a higher frequency and duration of positive engagements with their caregivers and thus benefit from such interactions. In contrast, children with profiles that fit a more ‘Reactive’ temperament may experience a lower frequency and duration of social interactions with others and thus miss out on some of the benefits from these interactions. Some evidence for this is suggested by Kasari and Sigman (1997) where results demonstrated that children with autism who fit within the ‘reactive’ temperament profile were less responsive and engaged during interactions when compared to children with easy temperaments. Although the directionality of these predictors is difficult to determine, such that it is yet unclear whether parenting practices may attenuate or intensify the effects of child temperament or vice versa (Chetcuti et al., 2019), further longitudinal research on changes in temperament and parenting practices is warranted.

Overall, study findings demonstrate the importance of temperament as an informative child-level variable in predicting outcome in school-aged children with autism. As noted previously, temperament profiles may affect the nature and quality of an individual’s social interactions. Although at this point we can only speculate about the mechanisms by which temperament influences the relationship between autism symptom severity and adaptive functioning skills, the results of this study demonstrate differential levels of adaptive functioning skill acquisition based on specific temperament profiles.

Limitations and future directions

Study findings need to be interpreted within the context of several limitations. First, as noted in other studies of temperament (Clifford et al., 2013), parents were the critical informants of temperament and relying solely on their reports can lead to bias including overestimation and underestimations across the dimensions of temperament (Chetcuti et al., 2019; Schwartz et al., 2009). Importantly, parents provided their reports on a number of other variables of interest including adaptive functioning skills which may have biased estimates of associations between temperament and other constructs as well. Future research should improve upon this design and use multi-method and multi-modal assessments of temperament, including self-reported measures, teacher measures and if possible, direct observations or assessments of temperament.

Second, our measurement of parent-reported temperament profiles occurred at a single time point and the assumption was made that temperament traits remain relatively stable within children with autism, as they do for same-aged neurotypical peers. Although little research has investigated the stability of temperament profiles in children with autism, our measure of adaptive functioning outcomes was collected 12 months after parental reports of temperament. Future research could consider including longitudinal assessments of temperament in children with autism.

Third, the association between temperament and emotional–behavioural problems and high correlation between shared characteristics, descriptors and traits add to the difficulties of disentangling the two constructs and their shared role in influencing the development of adaptive function in children with and without autism (Blair et al., 2004; Fabes et al., 1999). For example, children diagnosed with Attention Deficit Hyperactivity Disorder (ADHD; Eaves et al., 1994; Wing, 1997) are often characterized as having ‘reactive’ temperament (e.g. difficulty focusing, hyperactivity) and at the same time are reported to have more emotional–behavioural problems (e.g. inability to shift, increased temper tantrums and aggression towards others). Furthermore, the presence of these challenging behaviours including emotional–behavioural problems can lead to difficulties in developing adaptive social skills and daily-living skills and may exacerbate problems in developing meaningful social relationships with others (Blair et al., 2004). Although the present study found a significant effect of temperament on the association between autism symptom severity and adaptive functioning, this association could be influenced by the presence of challenging behaviours, especially in children with a ‘Reactive’ temperament profile. Although, it may be possible to have low challenging behaviours even with an ‘Reactive’ temperament profile. Furthermore, these discussions lend themselves to the growing literature on the relationship between certain temperament traits and the co-development and/or vulnerabilities (susceptibility) to psychopathologies, including children and youth diagnosed with ASD (Rettew, 2013). Measurements of emotional–behavioural problems and temperament will be difficult to tease apart, including their shared influence on the relationship between autism symptom severity and adaptive function, but future research would benefit from a more fine-grained analysis of the two constructs in order to identify their unique influences on the development of adaptive functioning.

Finally, it is important to consider that our measure of cognitive ability was a significant predictor of adaptive functioning outcome, even after entering the interact term between temperament and autism symptom severity as a moderator. This is not surprising given the strong relationship between cognitive ability and adaptive functioning, such that children who, on average, score higher on cognitive ability measures tend to develop more adaptive functioning skills (Perry et al., 2009). Interestingly, children in the ‘Even’ temperament profile scored slightly higher on the cognitive ability measure at T1 compared to children in the ‘Reactive’ temperament profile. This may be due to a number of factors including the relative ease of testing children with an ‘Even’ temperament (e.g. it may be easier for an assessor to establish rapport with ‘Even’ temperament children). Finally, these results speak to the heterogeneity, complexity and intertwined relationships between variables that influence adaptive functioning outcomes in children with autism and the need to consider these factors in future research.

In conclusion, study findings highlight the importance of considering how temperament – a relatively stable, child-level characteristic – may play a role in influencing outcome in children with autism. The finding that temperament has an influence on the association between autism symptom severity and adaptive functioning illustrates the importance of considering temperament profiles in the development and implementation of individualized intervention plans for children with autism. The field of autism intervention science may benefit from future studies investigating in more depth the role of temperament traits on individual responsiveness to treatment.

Footnotes

Appendix 1

Table 6.

Descriptive statistics and comparison between completed regression model data and missing data group on age at diagnosis at T1, sex distribution, and temperament profile membership.

Variable	Complete data set (n = 150)	Missing data (n = 35)	Effect size of difference
Variable	M (SD)	M (SD)	(Cohen’s d)
Sex (%)
Male	83.3	85.7
Female	17	14.3
Age at diagnosis (months)	38.73 (8.00)	36.88 (9.64)	0.02
Temperament group (%)
Even	40.7	34.3
Reactive	59.3	65.7

Funding

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

ORCID iDs

Vivian Lee

Connor Kerns

Isabel M Smith

Anat Zaidman-Zait

References

Achenbach

T. M.

Edelbrock

C. S.

(1983). Manual for the child behavior checklist and revised child behavior profile. Chicago.

Aiken

L. S.

West

S. G.

Reno

R. R.

(1991). Multiple regression: Testing and interpreting interactions. SAGE.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.).

Ashwood

K. L.

Tye

Azadi

Cartwright

Asherson

Bolton

(2015). Brief report: Adaptive functioning in children with ASD, ADHD and ASD+ ADHD. Journal of Autism and Developmental Disorders, 45(7), 2235–2242.

Bal

V. H.

Farmer

Thurm

(2017). Describing function in ASD: Using the DSM-5 and other methods to improve precision. Journal of Autism and Developmental Disorders, 47(9), 2939–2941.

Bartholomew

D. J.

Knott

(1999). Latent variable models and factor analysis (Vol. 7). Arnold.

Bennett

T. A.

Szatmari

Georgiades

Hanna

Janus

Georgiades

. . .Mirenda

(2015). Do reciprocal associations exist between social and language pathways in preschoolers with autism spectrum disorders? Journal of Child Psychology and Psychiatry, 56(8), 874–883.

Blair

Razza

R. P.

(2007). Relating effortful control, executive function, and false belief understanding to emerging math and literacy ability in kindergarten. Child Development, 78, 647–663. https://doi.org/10.1111/j.1467-8624.2007.01019.x

Blair

K. A.

Denham

S. A.

Kochanoff

Whipple

(2004). Playing it cool: Temperament, emotion regulation, and social behavior in preschoolers. Journal of School Psychology, 42(6), 419–443.

10.

Bradley

E. A.

Isaacs

B. J.

(2006). Inattention, hyperactivity, and impulsivity in teenagers with intellectual disabilities with and without autism. Canadian Journal of Psychiatry, 51, 598–599.

11.

Bradley

E. A.

Summers

Wood

Bryson

(2004). Comparing rates of psychiatric and behaviour disorders in adolescents and young adults with severe intellectual disability with and without autism. Journal of Autism and Developmental Disorders, 34, 151–161.

12.

Bryson

Zwaigenbaum

Brian

Roberts

Szatmari

Rombough

McDermott

(2007). A prospective case series of high-risk infants who developed autism. Journal of Autism and Developmental Disorders, 37, 12–24.

13.

Burrows

C. A.

Usher

L. V.

Schwartz

C. B.

Mundy

P. C.

Henderson

H. A.

(2016). Supporting the spectrum hypothesis: Self-reported temperament in children and adolescents with high functioning autism. Journal of Autism and Developmental Disorders, 46(4), 1184–1195.

14.

Carey

W. B.

(1998). Temperament and behavior problems in the classroom. School Psychology Review, 27, 522–533.

15.

Checa

Rodríguez-Bailón

Rueda

M. R.

(2008). Neurocognitive and temperamental systems of self-regulation and early adolescents’ social and academic outcomes. Mind, Brain, and Education, 2(4), 177–187.

16.

Checa

Rueda

M. R.

(2011). Behavioral and brain measures of executive attention and school competence in late childhood. Developmental Neuropsychology, 36, 1018–1032. https://doi.org/10.1080/87565641.2011.591857

17.

Chess

Thomas

(1987). Origins and evolution of behavior disorders: From infancy to early adult life. Harvard University Press.

18.

Chetcuti

Uljarevic

Hudry

(2019). Editorial perspective: Furthering research on temperament in autism spectrum disorder. The Journal of Child Psychology and Psychiatry, 60(2), 225–228.

19.

Clasen

D. R.

Brown

B. B.

(1985). The multidimensionality of peer pressure in adolescence. Journal of Youth and Adolescence, 14, 451–468. https://doi.org/10.1007/BF02139520

20.

Clifford

S. M.

Hudry

Elsabbagh

Charman

Johnson

, & The BASIS Team. (2013). Temperament in the first 2 years of life in infants at high-risk for autism spectrum disorders. Journal of Autism and Developmental Disorders, 43, 673–686.

21.

Dempsey

E. E.

Smith

I. M.

Flanagan

H. E.

Duku

Lawrence

M. A.

Szatmari

. . . Bennett

(2018). Psychometric properties of the Merrill–Palmer–Revised Scales of development in preschool children with autism spectrum disorder. Assessment. Advance online publication. https://doi.org/10.1177/1073191118818754

22.

De Pauw

S. S. W.

Mervield

(2010). Temperament, personality and developmental psychopathology: A review based on the conceptual dimensions underlying childhood traits. Child Psychiatry and Human Development, 41, 313–329.

23.

De Pauw

S. S. W.

Mervielde

Van Leeuwen

K. G.

De Clercq

B. J.

(2011). How temperament and personality contribute to the maladjustment of children with autism. Journal of Autism and Developmental Disorders, 41(2), 196–212.

24.

Derryberry

Rothbart

M. K.

(1997). Reactive and effortful processes in the organization of temperament. Developmental Psychopathology, 9, 633–652.

25.

Eaves

L. C.

H. H.

Eaves

D. M.

(1994). Subtypes of autism by cluster-analysis. Journal of Autism and Developmental Disorders, 24, 3–22.

26.

Eisenberg

Spinrad

T. L.

(2004). Emotion-related regulation: Sharpening the definition. Child Development, 75(2), 334–339.

27.

Ellis

L. K.

Rothbart

M. K.

Posner

M. I.

(2004). Individual differences in executive attention predict self-regulation and adolescent psychosocial behaviors. Annals of the New York Academy of Sciences, 1021, 337–340. https://doi.org/10.1196/annals.1308.041

28.

Fabes

R. A.

Eisenberg

Jones

Smith

Guthrie

Poulin

Shepard

Friedman

(1999). Regulation, emotionality, and preschoolers’ socially competent peer interactions. Child Development, 70(2), 432–442.

29.

Farley

M. A.

McMahon

W. M.

Fombonne

Jenson

W. R.

Miller

Gardner

. . .Coon

(2009). Twenty-year outcome for individuals with autism and average or near-average cognitive abilities. Autism Research, 2(2), 109–118.

30.

Fox

N. A.

Henderson

H. A.

(1999). Does infancy matter? Predicting social behavior from infant temperament. Infant Behavior and Development, 22(4), 445–455.

31.

Garon

Bryson

S. E.

Zwaigenbaum

Smith

I. M.

Brian

Roberts

Szatmari

(2009). Temperament and its relationship to autistic symptoms in a high-risk infant sib cohort. Journal of Abnormal Child Psychology, 37, 59–78.

32.

Garon

Zwaigenbaum

Bryson

Smith

I. M.

Brian

Roncadin

. . .Roberts

(2016). Temperament and its association with autism symptoms in a high-risk population. Journal of Abnormal Child Psychology, 44(4), 757–769.

33.

Georgiades

Kasari

(2018). Reframing optimal outcomes in autism. JAMA Pediatrics, 172, 716–717. https://doi.org/10.1001/jamapediatrics.2018.1016

34.

Georgiades

Bishop

S. L.

Frazier

(2017). Editorial perspective: Longitudinal research in autism-introducing the concept of ‘chronogeneity’. Journal of Child Psychology and Psychiatry, 58(5), 634–636.

35.

Georgiades

Szatmari

Boyle

Hanna

Duku

Zwaigenbaum

. . .Smith

(2013). Investigating phenotypic heterogeneity in children with autism spectrum disorder: A factor mixture modeling approach. Journal of Child Psychology and Psychiatry, 54(2), 206–215.

36.

Goldsmith

H. H.

Alansky

J. A.

(1987). Maternal and infant temperamental predictors of attachment: A meta-analytic review. Journal of Consulting and Clinical Psychology, 55(6), 805.

37.

Gosling

S. D.

Kwan

V. S.

John

O. P.

(2003). A dog’s got personality: A cross-species comparative approach to personality judgments in dogs and humans. Journal of Personality and Social Psychology, 85(6), 1161–1169.

38.

Gotham

Pickles

Lord

(2012). Trajectories of autism severity in children using standardized ADOS scores. Pediatrics, 130, e1278–e1284.

39.

Henninger

N. A.

Taylor

J. L.

(2012). Outcomes in adults with autism spectrum disorders: A historical perspective. Autism, 17(1), 103–116.

40.

Hepburn

S. L.

Stone

W. L.

(2006). Using Carey Temperament Scales to assess behavioral style in children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 36(5), 637–642.

41.

Hirschler-Guttenberg

Golan

Ostfeld-Etzion

Feldman

(2015). Mother, fathering, and the regulation of negative and positive emotions in high-functioning preschools with autism spectrum disorders. Journal of Child Psychology and Psychiatry, 56, 530–539.

42.

IBM Corp. (2017). IBM SPSS statistics for windows, version 25.0.

43.

Kagan

(1971). Change and continuity in infancy. Wiley.

44.

Kasari

Sigman

(1997). Linking parental perceptions to interactions in young children with autism. Journal of Autism and Developmental Disabilities, 27, 39–57.

45.

Konstantareas

M. M.

Stewart

(2006). Affect regulation and temperament in children with autism spectrum disorder. Journal of Autism and Developmental Disabilities, 36(2), 143–154.

46.

Lee

C. L.

Bates

J. E.

(1995). Mother-child interactions at age two years and perceived difficult temperament. Child Development, 56(5), 1314–1325.

47.

Lord

Risi

Lambrecht

Cook

E. H.

Leventhal

B. L.

DiLavore

P. C.

. . .Rutter

(2000). 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, 30(3), 205–223.

48.

MacDonald

Parry-Cruwys

Dupere

Ahearn

(2014). Assessing progress and outcome of early intensive behavioral intervention for toddlers with autism. Research in Developmental Disabilities, 35(12), 3632–3644.

49.

ManrNue-Vallier

(2014). Longitudinal mixed membership trajectory models for disability survey data. The Annals of Applied Statistics, 8, 2268–2291.

50.

Masten

A. S.

Coatsworth

J. D.

(1998). The development of competence in favorable and unfavorable environments: Lessons from research on successful children. American Psychologist, 53(2), 205–220.

51.

McLelland

G. H.

Judd

C. M.

(1993). Statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114(2), 376–390.

52.

Muris

Ollendick

T. H.

(2005). The role of temperament in the etiology of child psychopathology. Clinical Child and Family Psychology Review, 8(4), 271–289.

53.

Nelson

Martin

R. P.

Hodge

Havill

Kamphaus

(1999). Modeling the prediction of elementary school adjustment from preschool temperament. Personality and Individual Differences, 26, 687–700. https://doi.org/10.1016/S0191-8869(98)00174-3

54.

Nystrom

Bengtsson

(2017). A psychometric evaluation of the Temperament in Middle Childhood Questionnaire (TMCQ) in Swedish sample. Scandinavian Journal of Psychology, 58, 477–484.

55.

Pandolfi

Magyar

C. I.

Dill

C. A.

(2009). Confirmatory factor analysis of the child behavior checklist 1.5–5 in a sample of children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 39, 986–995.

56.

Perry

Flanagan

H. E.

Geier

J. D.

Freeman

N. L.

(2009). Brief report: The Vineland Adaptive Behavior Scales in young children with autism spectrum disorders at different cognitive levels. Journal of Autism and Developmental Disorders, 39(7), 1066–1078.

57.

Putnam

S. P.

Ellis

L. K.

Rothbart

M. K.

(2001). The structure of temperament from infancy through adolescence. In Eliasz

Angleitner

(Eds.), Advances/proceedings in research on temperament (pp. 165–182). Pabst Scientist.

58.

Putnam

S. P.

Rothbart

M. K.

Gartstein

M. A.

(2008). Homotypic and heterotypic continuity of fine-grained temperament during infancy, toddlerhood, and early childhood. Infant and Child Development: An International Journal of Research and Practice, 17(4), 387–405.

59.

Reichow

Barton

E. E.

Boyd

B. A.

Hume

(2012). Early intensive behavioral intervention (EIBI) for young children with autism spectrum disorders (ASD). Cochrane Database of Systematic Reviews, 10, Article CD009260.

60.

Rettew

(2013). Child temperament: New thinking about the boundary between traits and illness. WW Norton & Company.

61.

Riggs

N. R.

Jahromi

L. B.

Razza

R. P.

Dillworth-Bart

J. E.

Mueller

(2006). Executive function and the promotion of social-emotional competence. Journal of Applied Developmental Psychology, 27, 300–309. https://doi.org/10.1016/j.appdev.2006.04.002

62.

Roid

Sampers

(2004). Merrill-Palmer-revised scales of development. Stoelting Co.

63.

Rothbart

M. K.

(1986). Longitudinal observation of infant temperament. Developmental Psychology, 22(3), 356–365.

64.

Rothbart

M. K.

Ahadi

S. A.

Hershey

K. L.

Fisher

(2001). Investigations of temperament at three to seven years: The Children’s Behavior Questionnaire. Child Development, 72(5), 1394–1408.

65.

Rothbart

M. K.

Bates

J. E.

(2006). Temperament. In Damon

Lerner

R. M.

(Series Eds.) & Eisenberg

(Vol. Ed.), Handbook of child psychology: Vol. 3. Social, emotional and, personality development (6th ed., pp. 99–166). Wiley.

66.

Rothbart

M. K.

Derryberry

(1981). Development of individual differences in temperament. In Lamb

Brown

(Eds.), Advances in developmental psychology (Vol. 1, pp. 37–86). Erlbaum.

67.

Rothbart

M. K.

Posner

M. I.

Hershey

K. L.

(1995). Temperament, attention, and developmental psychopathology. In Cicchetti

(Ed.), Developmental psychopathology: Theory and methods (Vol. 1, pp. 315–340). Wiley.

68.

Rutter

Le Couteur

Lord

(2003). Autism diagnostic interview-revised. Western Psychological Services.

69.

Schwartz

C. B.

Henderson

H. A.

Inge

A. P.

Zahka

N. E.

Coman

D. C.

Kojkowski

N. M.

Hileman

C. M.

Mundy

P. C.

(2009). Temperament as a predictor of symptomology and adaptive functioning in adolescents with high-functioning autism. Journal of Autism and Developmental Disabilities, 39, 842–855.

70.

Schwartz

Gorman

A. H.

Nakamoto

McKay

(2006). Popularity, social acceptance, and aggression in adolescent peer groups: Links with academic performance and school attendance. Developmental Psychology, 42, 1116–1127. https://doi.org/10.1037/0012-1649.42.6.1116

71.

Shiner

R. L.

Buss

K. A.

McClowry

S. G.

Putnam

S. P.

Saudino

K. J.

Zentner

(2012). What is temperament now? Assessing progress in temperament research on the twenty-fifth anniversary of Goldsmith et al. Child Development Perspectives, 6(4), 436–444.

72.

Simonds

Rothbart

M. K.

(2009). Scoring procedure, Temperament in Middle Childhood Questionnaire (TMCQ), version 3.0. Acquired online after permission via Bowdoin University Webpage.

73.

Smith

(1999). Outcome of early intervention for children with autism. Clinical Psychology: Science and Practice, 6, 33–49. https://doi.org/10.1093/clipsy.6.1.33

74.

Sparrow

S. S.

Balla

D. A.

Cicchetti

D. V.

(2005). Vineland adaptive behavior scales: Survey forms manual. American Guidance Service.

75.

Spinrad

T. L.

Eisenberg

Harris

Hanish

Fabes

R. A.

Kupanoff

Ringwald

Holmes

(2004). The relation of children’s everyday nonsocial peer play behavior to their emotionality, regulation, and social functioning. Developmental Psychology, 40(1), 67–80.

76.

Szatmari

Georgiades

Duku

Bennett

T. A.

Bryson

Fombonne

. . .Volden

(2015). Developmental trajectories of symptom severity and adaptive functioning in an inception cohort of preschool children with autism spectrum disorder. JAMA Psychiatry, 72(3), 276–283.

77.

van den Boom

D. C.

(1994). The influence of temperament and mothering on attachment and exploration: An experimental manipulation of sensitive responsiveness among lower-class mothers with irritable infants. Child Development, 65(5), 1457–1477.

78.

Victor

J. B.

Rothbart

M. K.

Baker

S. R.

(2003). Manual for the Child Temperament and Personality Questionnaire (CTPQ). Hampton University.

79.

Wing

(1997). The autistic spectrum. The Lancet, 350, 1761–1766.

80.

Zentner

Shiner

R. L.

(Eds.). (2015). Handbook of temperament. Guilford Press.

81.

Zwaigenbaum

Bryson

Rogers

Roberts

Brian

Szatmari

(2005). Behavioral manifestations of autism in the first year of life. International Journal of Developmental Neuroscience, 23, 143–152.