Child and setting characteristics affecting the adult talk directed at preschoolers with autism spectrum disorder in the inclusive classroom

Abstract

Difficulty with social competence is a core deficit of autism spectrum disorder. Research on typically developing children and children with disabilities, in general, suggests the adult talk received in the classroom is related to their social development. The aims of this study were to examine (1) the types and amounts of adult talk children with autism spectrum disorder are exposed to in the preschool classroom and (2) the associations between child characteristics (e.g. language), activity area, and adult talk. Kontos’ Teacher Talk classification was used to code videos approximately 30 min in length of 73 children with autism spectrum disorder (ages 3–5) in inclusive classrooms (n = 33) during center time. The results indicated practical/personal assistance was the most common type of adult talk coded, and behavior management talk least often coded. Child characteristics (i.e. age and autism severity) and activity area were found to be related to specific types of adult talk. Given the findings, implications for future research are discussed.

Keywords

adult talk autism child characteristics preschoolers setting features

Introduction

The increasing prevalence of autism spectrum disorder (ASD) has resulted in more children with the disorder entering into mainstreamed preschool classrooms (Data Accountability Center, 2007). Such an inclusive education approach is thought to provide children with ASD opportunities to observe and interact with socially competent peers, which should promote their social interaction and play skills (Bailey et al., 1998; Brown et al., 2008). However, to realize the benefits of inclusion, children with disabilities must receive the appropriate support from adults in their environment (Bricker, 1995). Accommodations, modifications, and teaching strategies for young children with ASD can sometimes be challenging for classroom professionals to implement (Kohler and Strain, 1999; Stahmer et al., 2005). Simple strategies are therefore needed to support the social competence of young children with this disorder.

Adult talk, an expanded variant of teacher talk, is a phenomenon that can be conceptualized as the type and frequency of verbalizations directed at students (see Kontos and Wilcox-Herzog, 1997). This talk students receive from classroom adults, similar to language children receive in the home from parents (e.g. Hart and Risley, 1995, 1999), is likely to be essential for their development. A better understanding of the adult talk directed at children with ASD could result in an easy-to-use strategy for classroom practitioners. For example, early childhood educators could be taught how to provide more specific and targeted kinds of talk to children to promote their development and classroom engagement (e.g. using language that encourages children to stay in proximity to promote play skills). That said, key tenets of the bio-ecological model of development posit that the types of adult talk children receive may be related to both proximal child characteristics and classroom features, such as activity areas (Bronfenbrenner and Morris, 1998). Thus, in order to accurately determine the adult talk directed at young children with ASD and augment current and develop future intervention strategies, child and classroom characteristics that may affect this talk must be considered.

There is evidence that adult talk is related to the socially competent behavior of typically developing young children (e.g. Phillips et al., 1987) and those with disabilities (e.g. Sontag, 1997). Although these previous studies suggest adult talk affects children’s social competence, there is evidence that the talk teachers provide young children with and without disabilities is more centered on learning-related skills than peer-related skills (File, 1994; File and Kontos, 1993; Sontag, 1997). For example, Kontos (1999) examined the types of teacher talk thought to be related to typically developing preschooler’s classroom-based social competence. This teacher talk scheme included supporting peer relations, supporting object play (e.g. “You want to add one more track, so you can race your truck?”), positive social contacts (e.g. “Did you go to the fair over the weekend?”), behavior management, practical/personal assistance, talking to adults, and reading to the child. She found that teacher verbalizations around supporting object play (39%) and practical/personal assistance (e.g. helping child obtain classroom materials) (19%) were more common compared to supporting peer relations (3%). In a more recent study, Reszka (2010) examined adult behavior (i.e. verbal and non-verbal) directed at preschoolers with ASD. She reported that adult support, such as “Do you need help getting down the puzzle?,” was the most common behavior directed at preschoolers with ASD, with adult comments and approval (e.g. “Good job finishing your painting”) occurring less frequently. This adult support category, however, does not distinguish between talk supporting engagement in classroom activities and routines and those supporting peer relations. To our knowledge, the specific types and amounts of adult talk preschoolers with ASD experience in the classroom is unknown. Further, with low levels of supporting peer relations talk taking place in previous studies and appropriate interactions with peers being a primarily goal of inclusion, it is important to identify how much or little of this type of talk preschoolers with ASD are experiencing.

In addition to a need to better understand the types and amounts of adult talk children with ASD receive, knowledge about the child and setting characteristics that affect this talk is necessary. One particular ecological variable that may affect the adult talk children experience is activity area. For example, Kontos (1999) found that adult talk around supporting play with objects, positive social contacts, and practical/personal assistance was, on average, higher in constructive play areas (e.g. art activities) of the classroom compared to manipulatives (e.g. puzzles) and non-play areas. Girolametto et al. (2000) found book-reading activities elicited more behavior management talk, whereas during play-dough the talk adults provided was more conversational (e.g. open-ended questions). Thus, there is some evidence that the activity area a child is located in can affect the talk he or she receives from adults. It is likely that the types of talk classroom adults working with preschoolers with ASD use will also differ in specific activity areas. That said, identifying the types of adult talk that occur within specific activity areas could be important for informing future intervention efforts targeted at the level of classroom environmental modifications.

Child characteristics, such as age and amount of problem behavior, also have been identified as factors that influence the adult talk directed at young children. The research on the relationship between adult talk and age is mixed. For example, De Rivera et al. (2005) found that the number of open- or closed-ended questions teachers directed to typically developing toddlers versus preschoolers did not differ by age. Examining this same age range, Wilcox-Herzog and Kontos (1998) found that younger children received more elaborative (i.e. open questions, elaborative statements and suggestions) and non-elaborative talk (i.e. closed and declarative questions) combined than older children in the classroom. Finally, Girolametto and Weitzman (2002) reported that teachers’ mean length of utterance (MLU) was longer with preschoolers than toddlers. The relationship between children’s displays of challenging behavior and the types of talk adults direct to them is more consistent. Children who are perceived as displaying more challenging behavior often receive more command statements from teachers (Dobbs and Arnold, 2009), sometimes irrespective of whether misbehavior actually occurred (Dobbs et al., 2004).

Autism severity as well as the degree of language and cognitive impairments are additional child characteristics that may affect the adult talk children with ASD experience. Warren et al. (2010) examined the home language environment of young children with ASD in comparison to a normative sample of typically developing children. The results indicated that caregiver report measures of symptom severity were associated with the amount of words caregivers (i.e. adult language) directed to children. In other words, children with fewer autistic symptoms received more language from adults. In self-contained preschool classrooms, there is evidence that young children with ASD experience fewer words from classroom adults if they have lower cognitive (Dykstra et al., 2012; Irvin et al., 2013) and language abilities (Dykstra et al., 2012). Taken together, these studies suggest a link between adult language, which encompasses different types of talk, directed at young children with ASD and certain child characteristics.

Though informative, the prior studies that focused on preschoolers with ASD are limited in that the type of adult talk (e.g. supporting peer relations vs supporting object play) directed at these children was not fully considered. In addition, studies that elucidate the link between child and classroom characteristics and the adult talk directed at preschoolers with ASD are absent. Adult talk is an understudied but important phenomenon that is associated with the development of typical children as well as children with ASD. Understanding the relationship between type of adult talk, child (e.g. autism severity), and classroom characteristics (e.g. activity area) has important implications for school-based interventions, and ultimately the outcomes of children with ASD. Thus, the following research questions were addressed in this study: (1) what are the types and amounts of adult talk children with ASD receive in the preschool classroom; (2) how are child characteristics including autism severity, language, cognitive ability, age, and behavior associated with adult talk in classrooms serving preschoolers with ASD; and (3) how is activity area related to adult talk in classrooms serving preschoolers with ASD?

Method

Participants

Students

A total of 73 children who were part of a larger project evaluating the efficacy of preschool programs serving young children with ASD were included in this study. All participating children were enrolled in inclusive classrooms, and were between the ages of 3–5 years old (M = 48.64 months) at the time of study participation. Participants’ race and ethnicity were determined through parental report using a project-created demographic form. The racial/ethnic choices provided to participants were derived from the US Census guidelines; thus, participants were asked to report their racial group status and to separately report whether they were also Hispanic/Latino (ethnic group status). Children in the study had a prior educational or community diagnosis that placed them on the autism spectrum based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) criteria (i.e. autistic disorder, pervasive developmental disorder–not otherwise specified, or Asperger’s) or developmental delay, and had to meet diagnostic criteria for ASD on the most recent Autism Diagnostic Observation Schedule (ADOS) algorithms (American Psychiatric Association, 2000; Gotham et al., 2009). The ADOS is considered a gold-standard measure for ASD diagnosis. A research-certified ADOS assessor administered the majority of the ADOS assessments (72.5%); the remaining assessments were administered by assessors who had attained reliability with the research-certified assessors based on achieving a reliability criterion of 80% on the diagnostic algorithm. All measures were administered by trained reliable staff who possessed or were in the process of obtaining an MA or a PhD. Direct observations from the vantage point of the focal child with ASD took place on a single day, and assessments of students participating in this study took place within the first 3 months of the school year. The participants had varying levels of autistic symptoms and behavior problems as well as verbal and cognitive abilities (see Table 1).

Table 1.

Child assessment descriptives.

	N	M	Min.	Max.	SD
CARS total (raw score)	73	30.76	19.5	42.5	5.89
C-TRF total (raw score)	71	47.04	3	107	23.29
PLS-4 total (raw score)	72	69.68	19	129	29.25
PLS-4 total (standard score)	70	70.83	50	134	22.49
Mullen visual reception subscale (raw score)	73	34.81	7	50	9.85
Mullen (standard score)	71	67.03	49	113	18.44
Age (in months)	73	48.64	36	63	8.00

PLS-4: Preschool Language Scale; CARS: Childhood Autism Rating Scale; C-TRF: Caregiver-Teacher Rating Form.

Classrooms

A total of 33 classrooms participated in this study, with an average of two children recruited per classroom. The inclusive classrooms were comprised of two classroom types: Learning Experiences: Alternative Program for Preschoolers and Parents (LEAP) (Strain and Hoyson, 2000) and Business-as-usual (BAU) (i.e. eclectic programs). The prevailing theoretical/conceptual foundation for the LEAP model is applied behavior analysis. In LEAP, typically developing children learn peer-mediated methods aimed at improving the social interaction skills of children with ASD. The term “BAU” refers to classrooms in which teachers do not adhere to one theoretical/conceptual framework to guide their instructional practices (e.g. applied behavior analysis); rather, teachers in these classrooms use a variety of intervention approaches from multiple models. This eclectic model is the one typically used with children with ASD in the US public school system (Stahmer et al., 2005). LEAP and BAU classrooms included in the study had to be operating in a publically funded preschool and were screened for study inclusion by trained observers to ensure overall quality using a psychometrically validated tool, that is, the Professional Development in Autism (PDA) Program Assessment (Professional Development in Autism Center, 2008).Specifically, all classrooms had to meet an “average” rating (score of 3 out of 5) on four subscales of the PDA assessment during an initial classroom visit. These subscales included classroom structure, classroom environment, curriculum and instruction, and positive instructional climate. In addition, LEAP classrooms had to meet above-average ratings (3.5 out of 5) on model-specific subscales and items on its respective fidelity of implementation measures. If required average scores were not met on the initial visit, research staff conducted one additional classroom visit using the same criteria described above to make a final determination on study eligibility. Classrooms that did not meet study requirements were excluded. Trained and reliable research staff collected data on the fidelity of implementation of LEAP practices (M = 4.6, SD = 0.29) (i.e. LEAP fidelity measure) as well as overall classroom quality (i.e. PDA Program Assessment) for BAU classrooms (M = 4.3, SD = 0.34), four times across the school year. Both measures have adequate psychometric properties (see Hume et al., 2011) and were scored on a 1 (no/minimal implementation) to 5 (full implementation) scale. Two of the four visits included a reliability observer and the mean inter-rater reliability was 95% for the LEAP measure in LEAP classrooms (range 88%–100%), and 88 % for the PDA Quality measure in BAU schools (range 87%–99%). LEAP classrooms operate on a half-day schedule; thus the duration of classroom instructional time for students was 2–3 h per day (M = 2.05 h). The instructional time for BAU classrooms ranged from 2–5 h (M = 3 h). The number of children with ASD was slightly higher in LEAP (3.27) compared to BAU (2.18). Conversely, there were a higher number of children with other types of disabilities in BAU (M = 4.00) classrooms relative to LEAP (M = 1.05). The average number of typically developing preschoolers to children with ASD in LEAP (8.7 to 3.3) and BAU (7.4 to 2.2) classrooms was similar as were overall average adult-to-child ratios (LEAP = 4.19; BAU = 5.46). It was not the purpose of this study to examine group differences; therefore, classroom type (i.e. LEAP or BAU) was included as a covariate in data analyses.

Teachers

All teachers (n = 33) had to be certified to teach in their respective state and must have taught in a LEAP or BAU classroom for at least 2 years prior to study enrollment. Across both classroom types, teachers had an average of 10.6 years of experience in the classroom. Each LEAP and BAU classroom had 1–2 teachers as well as 1–2 teaching assistants. Teacher and student demographic information can be found in Table 2.

Table 2.

Teacher and child demographics.

Demographics		Frequency	%
Teacher demographics
Teacher race	Asian	0	0
	Black	1	3
	Multi-racial	0	0
	White	32	97
Teacher ethnicity	Hispanic	8	24.2
Teacher ethnicity	Non-Hispanic	25	75.8
Teacher gender	Male	1	3
Teacher gender	Female	32	97
Teacher degree	Associates	2	6.1
	Bachelors	11	33.3
	Masters	19	57.6
	Above masters	1	3.0
Student demographics
Child race	Asian	3	4.1
	Black	4	5.5
	Multi-racial	2	2.7
	White	64	87.7
Child ethnicity	Hispanic	32	43.8
Child ethnicity	Non-Hispanic	41	56.2
Child gender	Male	62	84.9
Child gender	Female	11	15.1

Measures

Childhood Autism Rating Scale (CARS; Schopler et al., 1988) is an instrument designed to assess children’s behaviors (e.g. relating to others, fear, non-verbal communication) in order to distinguish children with ASD (age 2 through adulthood) from those children who have a developmental delay. Scores on the measure range from 15 to 60 and are based on a 4-point scale, with higher scores indicating more symptom severity. The instrument has more than acceptable reliability (i.e. an alpha of 0.94). CARS administrators had to meet the following reliability criterion: be within ±0.5 on 12 of 15 items and within ±3 on the total score. Total scores were used in data analysis for this study.

Caregiver-Teacher Rating Form (C-TRF; Achenbach and Rescorla, 2000) is a teacher/caregiver report measure used to assess emotional and behavior problems and consists of internalizing and externalizing scales. The tool is appropriate for children aged 1.5–5 years old. There are 99 problem items that can be rated as 0 (not true) to 2 (very true or often true). The instrument has acceptable test–retest reliability (r = 0.81) and agreement between caregivers and teachers (r = 0.65). Total scores from the teacher-reported forms were used in data analysis.

Mullen Scales of Early Learning (Mullen, 1995) is a standardized, comprehensive assessment that measures gross motor, fine motor, visual reception, and language abilities of children from birth through 68 months. Akshoomoff (2006) established the validity of the measure for children with ASD. The calculated visual reception (VR) subscale raw score was used in this study as a proxy for the child’s non-verbal cognitive abilities, allowing for the separation of language from cognitive ability. For the current sample, the VR subscale was highly correlated with children’s total standard score (r = 0.81, p = < 0.0001) on the Mullen.

Preschool Language Scale (PLS-4; Zimmerman et al., 2002) is a standardized measure used to assess language development for children from birth through 6 years, 11 months. There is adequate test–retest reliability (0.90 to 0.97 for total language score). For this study, the raw total scores were used to estimate communication abilities. For the Mullen and PLS-4, raw scores were used because standard scores had floor effects and age equivalents scores were not available for all subscales or composite scores.

Procedure

For this study, the real-time activities and routines as well as the interactions participating children had with adults and peers were videotaped during 30 min of center time. Because it was a consistent activity across classrooms and occurred at a similar time of day, center time was a better choice for observing than outdoor time, for example, which was more variable across classroom types. No instructions were provided to teachers other than to continue with their normal classroom routines and activities. The behavior coding software, PROCODER, with its ability to capture the event and interval data, was used to code videos (Tapp and Walden, 2000). PROCODER was set up to divide the roughly 30-min videos into 10-s intervals, resulting in approximately 179 intervals per observation (range: 126–182 and M = 176). Partial interval sampling was used to code adult talk directed at the focal child for each 10-s interval. Activity area was based on where the child spent the majority of the time during each 10-s interval (see Powell et al., 2008 for an example of this approach).

Coding scheme

The coding scheme was comprised of categories and codes adapted from Kontos’ (1999) Teacher Talk scheme (see Appendix 1 for categories and examples) and the Code for Active Student Participation and Engagement Revised-III (CASPER- III; Tsao et al., 2001). Kontos’ coding scheme was deemed to be appropriate for this study because it was designed to capture the types of adult talk in the preschool setting that have been associated with children’s social competence. Her classification includes questions and statements that fall under the following categories: (a) supporting peer relations; (b) supporting object play; (c) positive social contacts; (d) behavior management; (e) practical/personal assistance; (f) talking to adults; and (g) teacher reads to child. For the current study, several modifications have been made to Kontos’ coding scheme. The individual codes were placed into the following hierarchy to establish a mutually exclusive coding system: supporting peer relations, positive social contacts, supporting object play, practical/personal assistance, and behavior management. Since social impairments are a core deficit of autism and children with this disorder appear to have the most difficulty engaging with peers followed by adults, these two categories come first in the hierarchy (Hauck et al., 1995). Although not a core deficit, supporting object play follows those two codes since appropriate engagement with objects is important for peer play, a skill that may be stilted or absent among children with ASD (Carter et al., 2005; Strain et al., 2008). Adaptive/independence competencies are another difficulty that can prevent opportunities to have meaningful exchanges with peers and adults (e.g., Hume et al., 2009); however, for our purposes, we were more interested in adult talk statements that are likely to directly influence children’s social competence. Behavior management came last in the hierarchy because we were more interested in pro-social than maladaptive exchanges between children with ASD and classroom adults and peers. Furthermore, preschoolers in this study had, on average, levels of problem behavior that fell below the clinical cut-off of 60 on the C-TRF measure (i.e. M = 47.04). Only the highest form of talk that occurred during an interval was coded. No talk and can’t tell (e.g. when it was unclear whether the adult was directing talk at the focal child) categories were added to the coding scheme so that codes were exhaustive. Finally, the talk of any classroom staff who directed a verbalization at the focal child with ASD was coded. Thus, adult talk reflects more than just teacher talk.

The following CASPER-III activity area categories were coded: transition, manipulatives, large motor, storytime/books, art, pretend/socio-dramatic play, large blocks, sensory, dance/music/recitation, food/snack, self-care/self-help, pre-academics, computer, and circle/large group time. However, activity areas that were related to each other conceptually were combined for data analysis purposes: (a) “large blocks/manipulatives” refers to large blocks and manipulative activities; (b) story/books, computer, and pre-academics were combined to form “pre-academic activities”; and (c) “sensory-motor” included sensory, art, and large motor activities. Finally, food/snack, dance/music/recitation, and self-care/self-help were not included in the analysis because children spent little time in these areas (i.e. less than 2%), and transition was not an area of interest for this study and therefore was excluded. Although not typically part of center time, teachers were allowed a maximum of 5 min for circle-time in order to limit disruption to regularly occurring activities; thus, it has been included in the analysis.

Inter-rater reliability

All videos were coded by two independent coders naïve to the purposes of this study. Of each coders’ videos, 20% was subsequently coded by a secondary rater (the first author (D.W.I.) of this study) for reliability purposes. Raters had to reach at least an average of 80% agreement per code as well as overall agreement with the author of this investigation across three consecutive non-study videos to begin coding videos used in the actual study. Coders had to maintain at least a mean 80% reliability criterion for each code as well as overall agreement throughout the coding period. Reliability values for both the overall adult talk (coder 1: 0.81 to 0.91, M = 0.83; coder 2: 0.80 to 0.94, M = 0.86) and activity area categories (coder 1: 0.73 to 0.98, M = 0.90; coder 2: 0.84 to 1.00, M = 0.94) were well within an acceptable range.

Results

Data analysis

The purpose of this study was to examine the adult talk that occurred in inclusive classrooms serving preschoolers with ASD as well as its relationship to setting (i.e. activity areas) and child characteristics (i.e. autism severity, language, cognitive ability, age, and problem behavior). For research question 2, a mixed multinomial logit model was relied on using PROC GLIMMIX (SAS Institute Inc., 2010). This type of model was necessary because of the clustering of observations within the child and children within classrooms as well as the categorical nature of the adult talk outcome (Raudenbush and Bryk, 2002). For this type of model, the dependent variable (i.e. adult talk) is based on the proportions of the different adult talk categories per child, which are transformed using log-odds to account for the intra-individual differences among children. In addition, this type of model requires a reference group so that comparisons can be made between categories of the outcome variable (Twisk, 2006). Although there is no standard procedure for choosing a reference group (Dr Chris Wiesen (statistician), personal communication, 13 February 2012), supporting peer relations was decided on because children with ASD appear to have greater difficulty socially engaging with peers compared to adults (Hauck et al., 1995). Covariates included in the model were race/ethnicity, gender, adult–child ratio as well as classroom treatment type (i.e. LEAP or BAU). Imputation using the expectation-maximization algorithm was chosen to address the small amount of missing data for the C-TRF (2.7%) and PLS-4 (1.4%). Finally, the following child characteristic predictors were grand mean centered: CARS, C-TRF, PLS-4, Mullen VR, and child age. For research question 3, a multinomial logistic regression was used to obtain marginal probabilities in order to determine the probability of preschoolers with ASD receiving a specific type of adult talk within a given activity area.

Research

question 1: What are the types and amounts of adult talk children with ASD receive in the preschool classroom?

Descriptives were used to identify the types and amounts of adult talk children with ASD received in the preschool classroom. We found the adult talk category that was most commonly coded was no talk (52.1%). Statements and questions around practical/personal assistance (14.6%) were the most frequently coded verbalization that children with ASD experienced. Behavior management (2.7%) was the type of talk least coded. Figure 1 provides an illustration for the percentages of each adult talk category.

Figure 1.

Percentages of adult talk coded.

Research

question 2: How are child characteristics including autism severity, language, cognitive ability, age, and problem behavior associated with adult talk in classrooms serving preschoolers with ASD?

The estimates for each child characteristic variable describe the likelihood of receiving a specific type of adult talk (e.g. supporting object play) relative to the reference group (i.e. supporting peer relations). Compared to supporting peer relations, coding of behavior management was more likely for preschoolers with ASD who had more severe autistic symptoms (β = 0.089, p = 0.026) than those with few autistic symptoms. Supporting object play was more likely to be coded for children with fewer cognitive difficulties than supporting peer relations (β = −0.104, p = 0.004). Positive social contact talk (β = 0.043, p = 0.021) was more likely to be coded for older children relative to younger children, who were more likely to receive supporting peer relations talk. Coding of positive social contacts (β = 2.139, p = 0.004) and practical/personal assistance (β = 1.953, p = 0.041) talk was more likely for children who were Black than supporting peer relations when compared to children who were White. For children who were Hispanic, practical/personal assistance (β = −1.089, p = 0.002) was more likely to be coded as opposed to supporting peer relations talk in comparison to preschoolers who were White and non-Hispanic. Coding of behavior management (β = 1.238, p = 0.013), positive social contacts (β = 0.844, p = 0.015), supporting object play (β = 1.779, p = 0.0003) was more likely for preschool girls with ASD over supporting peer relations compared to boys with ASD. In classrooms with lower adult–child ratios (i.e. fewer adults relative to children), behavior management (β = 0.226, p = 0.045), positive social contacts (β = 0.178, p = 0.015), and supporting object play talk (β = 0.318, p = 0.004) were more likely to be coded compared to supporting peer relations (Table 3).

Table 3.

Adult talk by child characteristics.

	Behavior management		Positive social contacts		Practical/personal assistance		Supporting object play
	β (Std err)	p	β (Std err)	p	β (Std err)	p	β (Std err)	p
Intercept	−2.309 (0.547)	<0.0001***	−0.321 (0.359)	0.372	0.799 (0.460)	0.084	−1.199 (0.538)	0.027*
CARS	0.089 (0.040)	0.026*	0.013 (0.025)	0.601	0.000 (0.033)	0.995	−0.006 (0.038)	0.869
C-TRF	0.003 (0.007)	0.681	−0.002 (0.005)	0.642	−0.008 (0.006)	0.18	0.005 (0.007)	0.48
PLS-4	0.013 (0.012)	0.279	−0.008 (0.008)	0.351	−0.002 (0.010)	0.818	0.004 (0.010)	0.699
Mullen VR	−0.067 (0.038)	0.08	−0.042 (0.026)	0.108	−0.052 (0.032)	0.104	−0.104 (0.036)	0.004**
Chronological age	−0.027 (0.028)	0.336	0.043 (0.019)	0.021*	0.032 (0.024)	0.184	0.030 (0.027)	0.27
Race-ethnicity: caregiver = Asian	1.239 (0.820)	0.131	−0.484 (0.603)	0.423	−0.270 (0.745)	0.717	0.102 (0.849)	0.905
Race-ethnicity: caregiver = Black	1.374 (1.067)	0.198	2.139 (0.732)	0.004**	1.781 (0.865)	0.04*	0.971 (1.100)	0.377
Race-ethnicity: caregiver = Hispanic	−0.194 (0.349)	0.579	−0.246 (0.227)	0.277	−1.089 (0.294)	0.0002**	0.110 (0.332)	0.74
Race-ethnicity: caregiver = Multi-racial	3.883 (2.907)	0.182	2.973 (2.771)	0.283	4.227 (2.847)	0.138	3.966 (2.918)	0.174
Gender: female	1.238 (0.499)	0.013*	0.844 (0.347)	0.015*	0.745 (0.436)	0.088	1.779 (0.491)	0.0003**
Classroom Type: BAU	0.568 (0.410)	0.165	−0.006 (0.272)	0.982	0.372 (0.348)	0.284	0.103 (0.398)	0.796
A-C Ratio	0.226 (0.112)	0.045*	0.178 (0.073)	0.015*	0.088 (0.094)	0.352	0.318 (0.109)	0.004**

PLS-4: Preschool Language Scale; CARS: Childhood Autism Rating Scale; C-TRF: Caregiver-Teacher Rating Form; LEAP: Learning Experiences: Alternative Program for Preschoolers and Parents; BAU: Business-as-Usual; VR: visual reception.

Reference groups include supporting peer relations (adult talk), race/ethnicity (White), gender (male), and classroom type (LEAP).

p < 0.05; **p< 0.01; ***p< 0.0001.

Research

question 3: How is activity area related to adult talk in classrooms serving preschoolers with ASD?

Children with ASD spent the most time in large blocks/manipulatives (39.80%) and the least amount of time in circle-time (5.40%). The time spent in pretend/socio-dramatic (23.79%), sensory-motor (20.08%), and pre-academics (10.93%) fell between large blocks/manipulatives and circle-time. When adults were talking to children with ASD, supporting peer relations talk was more likely to be coded in circle-time (0.227) and less often in large blocks/manipulatives (0.108) and sensory-motor (0.107) followed by pre-academics (0.106). Verbalizations related to positive social contacts were most likely to be coded in circle-time (0.317) and least likely to be coded in large blocks/manipulatives (0.162). Practical/personal assistance was coded more often in sensory-motor (0.506) relative to pretend/socio-dramatic (0.133). Supporting object play was coded most often in pretend/socio-dramatic (0.449) and was absent in circle-time (0). Finally, behavior management talk was coded more frequently in circle-time (0.076) compared to sensory-motor (0.043). The probability of each specific type of adult talk being coded within individual activity areas can be found in Table 4.

Table 4.

Probability of types of adult talk occurring within an activity area.

Adult talk	Pre-academics	Sensory/motor	Pretend/socio-dramatic	Circle-time	Large blocks/manipulatives
Behavior management	0.053	0.043	0.061	0.076	0.064
Positive social contacts	0.253	0.173	0.220	0.317	0.162
Practical/personal assistance	0.457	0.506	0.133	0.380	0.242
Supporting object play	0.131	0.171	0.449	0.000	0.424
Supporting peer relations	0.106	0.107	0.137	0.227	0.108

Discussion

The notion that features of children’s primary environments and the interactions within them affect children’s development is the crux of the bio-ecological model. This study provides new knowledge about the amount of specific types of adult talk directed at children with ASD that may, in turn, affect their socially competent behavior. Furthermore, our results support the view that certain child and setting characteristics may affect the interactions between classroom adults and preschoolers with this disorder. Our findings should be interpreted with caution since the coding hierarchy could have resulted in some instances of specific types of adult talk being missed.

First, our results indicate that adults were not talking to preschoolers with ASD a little over half the time observed (i.e. 52.1%). This amount of no talk resembled prior studies of adult behavior (i.e. verbal and non-verbal) directed at preschoolers with disabilities (58%) and children without disabilities (65%) (Brown et al., 1999) and is not surprising, because as Kontos (1999) points out, children typically outnumber adults in the classroom; thus, it may be difficult for adults to direct a lot of talk to each child individually. Still, it is important to note that adults in this study did spend a good deal of their time providing children with some form of talk (42.8%), when talk is aggregated across categories. Considering children with this disorder may require assistance from adults to navigate the preschool classroom and their potential interest in objects over humans, it is not surprising that practical/personal assistance and supporting object play were frequent types of talk children experienced. The low amounts of positive social contact and peer relations talk suggest that adults may view relationship-building with students with ASD as a difficult endeavor. Another possibility is that adults were focused on ensuring children learn to participate in activities and classroom routines at the beginning of the school year and their talk is a reflection of this. It was puzzling that the least used type of adult talk was behavior management, since behavioral problems are often attributed to children with ASD (Matson et al., 2009). However, it may be a product of preschoolers in this study having, on average, levels of problem behavior that fell below the clinical cut-off on the C-TRF measure or an artifact of the coding hierarchy (i.e. behavior management was the lowest code in the hierarchy and therefore the least likely to be coded when another form of talk was present in a given 10-s interval). It must be noted that these rates of adult talk resemble those reported in Kontos’ (1999) study of the teacher talk directed at typically developing preschoolers. Therefore, children’s autistic symptoms might not be the primary driver of the type of adult talk to which children with ASD are exposed. A comparison study of the adult talk directed at typically developing preschoolers and preschoolers with ASD is needed to better understand how young children’s autistic symptoms do or do not affect adult talk in the classroom.

Studies have found that a preschooler’s disability status may be related to the talk he or she receives from adults (e.g., Hestenes et al., 2004; Kontos et al., 1998). This study adds to these previous studies by identifying certain child characteristics (i.e. autism severity and degree of cognitive impairments) within disability status (i.e. ASD) that may affect specific types of adult talk. Warren et al. (2010) found that autism severity was negatively related to adult language in the home. Our findings suggest that in the preschool classroom, children with more severe autistic symptoms were more likely to hear behavior management talk from adults as opposed to supporting peer relations. One possibility for this finding is that adults were more focused on children with more autistic symptoms meeting classroom expectations than engaging with peers, particularly at the beginning of the school year. Previous studies have reported that the children with ASD who have more severe language (i.e., Dykstra et al., 2012) and cognitive impairments (Dykstra et al., 2012; Irvin et al., 2013) receive fewer adult verbalizations than those with intact language and cognitive abilities. Our study qualifies these studies by identifying the specific types of adult talk that are related to cognitive ability (i.e. more supporting peer relations talk directed at children with fewer cognitive difficulties). Unlike Dykstra et al.s’ study, we found language ability was not associated with the types of talk children were exposed to. This divergent finding could be related to the length of the classroom observations and/or different instruments used to capture and code adult verbalizations (i.e. Language ENvironment Analysis (LENA) audio recordings vs video recordings). It would be important to understand the relationship between adult talk and children’s developmental outcomes and learning to determine whether adults should modify the types of talk they provide based on children’s cognitive and language profiles.

Previous research also suggests that the relationship between children’s age and the talk they receive from adults is mixed (e.g. De Rivera et al., 2005; Wilcox-Herzog and Kontos, 1998). We found that younger children with ASD were more likely to hear supporting peer relations talk than older children, who were more likely to hear positive social contacts talk (i.e. positive social statements from adults). This may be related to teachers focusing more on social customs, such as greeting adults, with older children who are on the verge of transitioning to kindergarten. Another possibly is the fact that older children in this study had less severe autistic symptoms than younger children (r = −0.253 p = 0.031), which may have prompted classroom adults to provide more support to younger children around relationships with peers than older ones. Taken together, these findings suggest that certain child characteristics (i.e. autism severity, cognitive ability, and age) may affect the types of adult talk children receive in the inclusive classroom.

Previous studies suggest that the activity area a young child is located in is related to the types of talk heard from classroom adults (Girolametto et al., 2000; Kontos, 1999). We found that preschoolers with ASD spent the most amount of time in large blocks/manipulatives and the least in circle-time. However, supporting peer relations talk was more likely to be used in circle-time relative to large blocks/manipulatives. One possibility for this finding is that circle-time often provides a natural time for children to be grouped together, so teachers may use this time to work on peer interactions. With the difficulties children with this disorder have in positively interacting with peers, it is concerning not only that supporting peer relations was a low frequency type of talk in this study but also that it was most commonly used in an activity area where these children spent the least amount of time.

Although not a central focus of this study, adult–child ratio was a setting feature that was related to the adult talk children with ASD received. Specifically, as the number of children to adults increased in classrooms, children were more likely to experience behavior management, positive social contact, and supporting object play talk relative to supporting peer relations. The research is not definitive as to whether adult–child ratio impacts the quality of teacher–student interactions, with some studies suggesting a relationship (File and Kontos, 1993; Hestenes et al., 2008) and others suggesting minimal to no relationship (Mashburn et al., 2008; Pianta et al., 2005). It is important to note that this study only examined the types and amount of talk children received and not the quality of teacher–student interactions. Nevertheless, a greater number of children relative to classroom adults may result in adults spending less time encouraging peer relations.

Limitations

This study has several limitations that must be noted. First and foremost, the adult talk coding scheme was hierarchical, which may have resulted in some types of adult talk being missed during coding intervals—particularly categories that were infrequent and at the bottom of the hierarchy (e.g. behavior management). For example, if a focal child received verbalizations related to supporting object play and behavior management in the same 10-s interval, then only the higher level of talk (i.e. supporting object play) would have been captured. However, the adult talk hierarchy was necessary for conceptual behavior coding and data analysis purposes. The videotaping of children with ASD took place during center time for 30 min on a single day at the beginning of the school year, so whether similar types and amounts of adult talk are present at other times of the preschool day and year is unknown. This investigation lacks direct measures of children’s social competence; thus, the relationship between specific types of adult talk and the social outcomes of children with ASD remains tenuous. Only high-quality classrooms were included in this study; thus, it is uncertain whether these findings generalize to lower or modal quality inclusive preschool classrooms. The 3-month data collection period for the child measures could have introduced some maturation effects; however, this is of less concern since this study was not examining change over time. Finally, length of enrollment was not considered in this study, so whether the adult talk directed at children with ASD who had the same teacher for 2 years versus 1 year is different remains unknown.

Future research

Overall, this study builds on previous research by offering new knowledge about the adult talk children with ASD receive in the inclusive classroom. In addition, our findings qualify existing research and provide new knowledge on the link between adult verbalizations directed at children with ASD and child and setting characteristics. Future research should establish whether a relationship between children’s social competence and specific types of adult talk exists; if there is a link, professional development studies will be needed in order to determine the most effective way to train the various adults (e.g. lead teachers, paraprofessionals) in inclusive classrooms to better match their talk to children’s symptoms (e.g. more practical/personal assistance talk for children with cognitive difficulties). Investigations that use the existing coding system to examine whether the adult talk directed at children varies by disability type (e.g. ASD vs Down syndrome) also could provide a better understanding of the relationship between adult talk and disability status. Future investigations should incorporate characteristics of early childhood practitioners (e.g. burnout level, lead teacher vs paraprofessional, teacher race and ethnicity), other setting features (e.g. group composition variables such as the number and type of peers and adults within a given activity area), and child characteristics (e.g. socioeconomic status (SES)) that have been shown to affect adult talk but that were not able to be included in this study. Finally, studies are needed to determine thresholds of specific types of adult talk that result in changes in the social competence of children with ASD, given that this is the core deficit of the disorder.

Footnotes

Appendix

Appendix 1.

Categories and examples of adapted teacher talk classification from Kontos (1999).

Supporting Peer Relations: Statements: Adult describes, explains, encourages, gives a reason or information about a child’s interactions with another child (e.g. “When you told Mary she was mean, it upset her”). Questions: Adult asks questions about a child’s interactions with a peer or group of peers (e.g. “Did you ask Tim if he would like to play?”).

Non-examples: If adult tells the focal child and two other children, “Let’s each grab a funnel to play with, friends,” supporting object play should be coded because the statement is primarily concerned with the funnel.

Positive Social Contacts: Statements: Adult uses statements to praise or encourage the child or a group the child is part of to interact with the focal adult or another adult in order to support positive adult–student relations in the classroom (e.g. “Want to come keep me company while I put out snack?). Questions: Adult socializes with a child (e.g. “What did you do last night?”).

Non-examples: If an adult says, “You did a great job counting the cars,” it should be coded practical/personal assistance, even though it is positive statement coming from the adult—the talk is about counting, an academic activity.

Support Object Play: Statements: Adult praises, encourages, describes, explains, gives reason or information about a child’s object/fantasy play (e.g. “This fire helmet can turn you into a firefighter”). Questions: Adult asks questions about a child’s object play (e.g. “Are you going to make a train out of play-dough today?”).

Non-examples: If an adult and focal child are rolling a toy car back and forth, and the adult says, “Ready, set, go” to the child, positive social contacts should be coded, since the aim of the request is to encourage relations between the adult and focal child.

Practical/Personal Assistance: Statements: Adult provides reminders/encouragement/praise about self-help (e.g. clothing, toileting). Praise, encouragement, and information related to academic and pre-academic activities (e.g. counting, matching) should be coded practical/personal assistance. These statements also include those aimed at helping a child store or clean up materials after the activity is done (e.g. The blocks go on the second shelf). Questions: Adult helps child obtain materials needed for an activity (e.g. “Do you have the pizza cutter for our pizza party?”).

Non-examples: If adult says, “You don’t get to go to art until you clean up the puzzles” behavior management should be coded, since this statement is about following a classroom rule.

Behavior Management: Statements: These are statements where the child displays an undesirable behavior and an adult commands/requests/implies a different one be used. Examples include the adult redirecting a child who is not engaged toward an activity without explanation (e.g. “Tory, stop looking at the computer area, we are building a tower with blocks right now”). Questions: Adult asks focal child about classroom rules and routines (e.g. “Why aren’t you sitting on rug during circle-time?).

Non-examples: If after the adult tells the focal child not to take a toy, she goes on to say “I think apologizing to Becky for taking her toy would make her happy,” supporting peer relations should be coded, since its purpose is to mend strained peer relations.

Can’t Tell: Acknowledgments or responses that do not fit into one of the above-mentioned categories (e.g. “Okay,” “What?,” “Yeah”). You’re unable to tell if a verbalization is directed at the focal child or you are not able to hear what the adult is saying (e.g. high levels of video static). If the adult and child are outside of camera range and you are not able to determine if interactions between them are taking place (e.g. the adult and child enter the bathroom)—code can’t tell. Keep in mind, this code should be used sparingly.

No Talk: Code if during a 10-s interval, no talk is directed at the focal child. If an adult is assigning children to different activity areas during circle-time, no talk should be coded unless the verbalization is directed at the target child.

Acknowledgements

The first author (DWI) was a doctoral student at the University of North Carolina at Chapel Hill at the time the work for this article was completed. Special thanks to Chris Wiesen for all the help with data analysis.

Funding

The research reported here was supported in part by the Institute of Education Sciences, US Department of Education through Grant R324B070219 awarded to UNC-Chapel Hill. The opinions expressed represent those of the authors and do not necessarily represent views of the institute or the US Department of Education.

References

Achenbach

Rescorla

(2000) Caregiver-Teacher Rating Form. Burlington, VT: The University of Vermont.

Akshoomoff

(2006) Use of the Mullen scales of early learning for the assessment of young children with autism spectrum disorders. Child Neuropsychology 12: 269–277.

American Psychiatric Association (2000) Diagnostic and statistical manual of mental disorders: DSM-IV-TR. American Psychiatric Publications.

Bailey

McWilliam

Buysse

(1998) Inclusion in the context of competing values in early childhood education. Early Childhood Research Quarterly 13: 27–47.

Bricker

(1995) The challenge of inclusion. Journal of Early Intervention 19(3): 179–194.

Bronfenbrenner

Morris

(1998) The ecology of developmental processes. In: Damon

Lerner

(eds) Handbook of Child Psychology. 5th ed., vol. 1. New York: Wiley, pp.992–1028.

Brown

Odom

(1999) Ecobehavioral assessment in early childhood programs: a portrait of preschool inclusion. Journal of Special Education 33: 138–153.

Brown

Odom

McConnell

(2008) Peer interaction interventions for preschool children with developmental difficulties. In: Brown

Odom

McConnell

(eds) Social Competence of Young Children: Risk, Disability, and Intervention. Baltimore, MD: Paul H Brookes, pp.141–163.

Carter

Ornstein-Davis

Klin

(2005) Social development in autism. In: Ornstein-Davis

(ed.) Handbook of Autism and Pervasive Developmental Disorders. 3rd ed., vol. 1. Hoboken, NJ: John Wiley & Sons, Inc., pp.312–334.

10.

Data Accountability Center (2007) Children with disabilities receiving special education under part B of the individuals with disabilities act, 1997–2006. U.S. Department of Education, Office of Special Education Programs, Data Analysis (DANS). OMB #1820-0043. Available at: http://tadnet.public.tadnet.org/pages/712

11.

De Rivera

Girolametto

Greenberg

(2005) Children’s responses to educators’ questions in day care play groups. American Journal of Speech-Language Pathology 14: 14–26.

12.

Dobbs

Arnold

(2009) Relationship between preschool teachers’ reports of children’s behavior and their behavior toward children. School Psychology Quarterly 24(2): 95–105.

13.

Dobbs

Arnold

Doctoroff

(2004) Attention in the preschool classroom: the relationship among child gender, child misbehavior, and types of teacher attention. Early Child Development and Care 174: 281–295.

14.

Dykstra

Sabatos-DeVito

Irvin

(2012) Using the language environment analysis (LENA) system in preschool classrooms with children with autism spectrum disorders. Autism 17(5): 582–594.

15.

File

(1994) Children’s play, teacher-child interactions, and teacher beliefs in integrated early childhood programs. Early Childhood Research Quarterly 9(2): 223–240.

16.

File

Kontos

(1993) The relationship of program quality to children’s play in integrated early intervention settings. Topics in Early Childhood Special Education 13(1): 1–18.

17.

Girolametto

Weitzman

(2002) Responsiveness of child care providers in interactions with toddlers and preschoolers. Language, Speech, and Hearing Services in Schools 33: 268–281.

18.

Girolametto

Hoaken

Weitzman

(2000) Patterns of interaction in integrated day care groups. Language, Speech, and Hearing Services in Schools 31: 154–167.

19.

Gotham

Pickles

Lord

(2009) Standardizing ADOS scores for a measure of severity in autism spectrum disorders. Journal of Autism and Developmental Disorders 39: 693–705.

20.

Hart

Risley

(1995) Meaningful Differences in the Everyday Experience of Young American Children. Baltimore, MD: Paul H Brookes.

21.

Hart

Risley

(1999) The Social World of Children: Learning to Talk. Baltimore, MD: Paul H Brookes.

22.

Hauck

Fein

Warehouse

(1995) Social initiations by autistics children to adults and other children. Journal of Autism and Developmental Disorders 25(6): 579–595.

23.

Hestenes

Cassidy

Niemeyer

(2004) A microanalysis of teachers’ verbalizations in inclusive classrooms. Early Education and Development 15: 23–38.

24.

Hestenes

Cassidy

Shim

(2008) Quality in inclusive preschool classroom. Early Education and Development 19(4): 519–540.

25.

Hume

Boyd

McBee

(2011) Assessing implementation of comprehensive treatment models for young children with ASD: reliability and validity of two measures. Research in Autism Spectrum Disorders 5: 1430–1440.

26.

Hume

Loftin

Lantz

(2009) Increasing independence in autism spectrum disorders: a review of three focused interventions. Journal of Autism and Developmental Disorders 39(9): 1329–1338.

27.

Irvin

Hume

Boyd

(2013) Child and classroom characteristics associated with the adult language provided to preschoolers with autism spectrum disorder. Research in Autism Spectrum Disorders 7(8): 947–955.

28.

Kohler

Strain

(1999) Maximizing peer-mediated resources in integrated preschool classrooms. Topics in Early Childhood Special Education 19(2): 92–102.

29.

Kontos

(1999) Preschool teachers’ talk, roles, and activity settings during free play. Early Childhood Research Quarterly 14(3): 363–382.

30.

Kontos

Wilcox-Herzog

(1997) Teachers’ interactions with children: why are they so important? Research in review. Young Children 52(2): 4–12.

31.

Kontos

Moore

Giorgetti

(1998) The ecology of inclusion. Topics in Early Childhood Special Education 18: 38–49.

32.

Mashburn

Pianta

Hamre

(2008) Measures of classroom quality in prekindergarten and children’s development of academic, language, and social skills. Child Development 79: 732–749.

33.

Matson

Wilkins

Macken

(2009) The relationship of challenging behaviors to severity and symptoms of autism spectrum disorders. Journal of Mental Health Research in Intellectual Disabilities 2: 29–44.

34.

Mullen

(1995) Mullen Scales of Early Learning. Los Angeles, CA: Western Psychological Services.

35.

Phillips

McCartney

Scarr

(1987) Child-care quality and children’s social development. Developmental Psychology 23(4): 537–543.

36.

Pianta

Howes

Burchinal

(2005) Features of prekindergarten programs, classrooms, and teachers: do they predict observed classroom quality and child teacher interactions? Applied Developmental Science 9: 144–159.

37.

Powell

Burchinal

File

(2008) An eco-behavioral analysis of children’s engagement in urban public school preschool classrooms. Early Childhood Research Quarterly 23: 108–123.

38.

Professional Development in Autism Center (2008) PDA Program Assessment. Seattle, WA: PDA Center at University of Washington.

39.

Raudenbush

Bryk

(2002) Hierarchical Linear Models: Applications and Data Analysis Methods. 2nd ed. Thousand Oaks, CA: SAGE.

40.

Reszka

(2010) Ecological features of preschool environments and the social engagement of children with autism spectrum disorder. Doctoral Dissertation, The University of North Carolina at Chapel Hill, Chapel Hill, NC (Retrieved from ProQuest Dissertations and Theses. (AAT 3418605)).

41.

SAS Institute Inc. (2010) SAS (Version 9.2). Cary, NC: SAS Institute Inc.

42.

Schopler

Reichler

Renner

(1988) The Childhood Autism Rating Scale. Los Angeles, CA: Western Psychological Services.

43.

Sontag

(1997) Contextual factors influencing the sociability of preschool children with disabilities in integrated and segregated classrooms. Exceptional Children 63(3): 389–404.

44.

Stahmer

Collings

Palinkas

(2005) Early intervention practices for children with autism: descriptors from community providers. Focus on Autism and Other Developmental Disabilities 20: 66–79.

45.

Strain

Schwartz

Bovey

(2008) Social competence interventions for young children with autism: programmatic research findings and implementation issues. In: Brown

Odom

McConnell

(eds) Social Competence of Young Children: Risk, Disability, and Intervention. Baltimore, MD: Paul H Brookes, pp. 253–272.

46.

Strain

Hoyson

(2000) The need for longitudinal, intensive social skills intervention: LEAP follow-up outcomes for children with autism. Topics in Early Childhood Special Education 20: 116–122.

47.

Tapp

Walden

(2000) Procoder: a system for collection and analysis of observational data from videotape. In: Thompson

Felce

Symons

(eds) Behavioral Observation: Technology and Applications in Developmental Disabilities. Baltimore, MD: Paul H Brookes, pp.61–70.

48.

Tsao

Odom

Brown

(2001) Code for Active Student Participation and Revised (CASPER III): A Training Manual for Observers. Bloomington, IN: Indiana University.

49.

Twisk

JWR

(2006) Applied Multilevel Analysis: A Practical Guide. Cambridge: Cambridge University Press.

50.

Warren

Gilkerson

Richards

(2010) What automated vocal analysis reveals about the vocal production and language learning environment of young children with autism. Journal of Autism and Developmental Disorders 40(5): 555–569.

51.

Wilcox-Herzog

Kontos

(1998) The nature of teacher talk in early childhood classrooms and its relationship to children’s competence with objects and peers. Journal of Genetic Psychology 159(1): 30–44.

52.

Zimmerman

Steiner

Pond

(2002) The Preschool Language Scale. 4th ed. San Antonio, TX: The Psychological Corporation.