Journal of Communication Disorders 46 (2013) 428–439

Contents lists available at ScienceDirect

Journal of Communication Disorders

Quantitative and qualitative characteristics of the school and

home language environments of preschool-aged childrenwith ASD

Sloane Burgess *, Lisa Audet, Sanna Harjusola-Webb

Kent State University, Speech Pathology & Audiology, A104 Music and Speech Building, Kent, OH 44242, United States

A R T I C L E I N F O

Article history:

Received 4 September 2012

Received in revised form 20 September 2013

Accepted 21 September 2013

Available online 11 October 2013

Keywords:

Autism Spectrum Disorder

Naturalistic language sampling

Preschool

Language environment

LENA

A B S T R A C T

The purpose of this research was to begin to characterize and compare the school and

home language environments of 10 preschool-aged children with Autism Spectrum

Disorders (ASD). Naturalistic language samples were collected from each child, utilizing

Language ENvironment Analysis (LENA) digital voice recorder technology, at 3-month

intervals over the course of one year. LENA software was used to identify 15-min segments

of each sample that represented the highest number of adult words used during

interactions with each child for all school and home language samples. Selected segments

were transcribed and analyzed using Systematic Analysis of Language Transcripts (SALT).

LENA data was utilized to evaluate quantitative characteristics of the school and home

language environments and SALT data was utilized to evaluate quantitative and

qualitative characteristics of language environment. Results revealed many similarities

in home and school language environments including the degree of semantic richness, and

complexity of adult language, types of utterances, and pragmatic functions of utterances

used by adults during interactions with child participants. Study implications and

recommendations for future research are discussed.

Learning outcomes: The reader will be able to, (1) describe how two language

sampling technologies can be utilized together to collect and analyze language samples,

(2) describe characteristics of the school and home language environments of young

children with ASD, and (3) identify environmental factors that may lead to more positive

expressive language outcomes of young children with ASD.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Language learning is a dynamic process that requires the active participation of both language partners to obtainoptimum results. Young children use eye gaze, gestures such as pointing and showing, and words to initiate interactionswith adults their environment and communicate a variety of functions (Halliday, 1978). A typical adult response to a childinitiation may be to provide the child with information about the object or event of interest using the type and amount oflanguage that he/she deems appropriate in relation to the child’s development (Murray, Johnson, & Peters, 1990). In this way,the adult is creating a language environment that is individualized to the child in terms of the content as well as the quantityand quality of the language provided. When a child has a social-communicative impairment, such as an Autism Spectrum

* Corresponding author. Tel.: +1 330 672 0610; fax: +1 330 672 2512.

E-mail addresses: sburges8@kent.edu (S. Burgess), laudet@kent.edu (L. Audet), shwebb@kent.edu (S. Harjusola-Webb).

0021-9924/$ – see front matter � 2013 Elsevier Inc. All rights reserved.

http://dx.doi.org/10.1016/j.jcomdis.2013.09.003

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 429

Disorder (ASD), there may be a disruption to the language learning partnership. Adults are charged with remediating theimpairment and minimizing the disruption to the partnership across multiple contexts. Young children with ASD areincreasingly likely to interact with communicative partners in early intervention (school) environments, however, to date,little is known about characteristics of their school language environments. This study begins to characterize and compareschool and home language environments.

ASD is a neurodevelopmental disorder that results in impairments in social interaction and communication, andrepetitive and stereotyped patterns of behavior, interests, and activities (APA, 2000). Among the earliest indicators of ASD isthe absence of typical pre-verbal social-communicative behaviors including gestures such as pointing or showing to directthe attention of others, gaze monitoring, sharing affect, and responding to the initiations of others (Baron-Cohen et al., 1996;Osterling, Dawson, & Munson, 2002; Wetherby et al., 2004). The development of verbal language is often delayed as well(Weismer, Lord, & Isler, 2010). It has been hypothesized that young children with ASD may be provided with a decreasedamount and/or different quality of adult ‘‘talk’’ than that provided to their typically developing peers during the criticalperiod for language development, as a result of their social communicative impairments (Warren, Gilkerson, et al., 2010). Forexample, an adult may talk less to a child who does not initiate interactions and is content to be alone, or if the child does notdirect the adult’s attention to objects/activities of his or her liking, may talk about topics outside of the child’s interest,limiting the child’s engagement and motivation to actively participate. High quality language environments have been foundto be positively related to expressive language outcomes of typically developing children as well as those withdevelopmental disabilities including ASD.

1.1. Home language environments

The quantity of adult language children are exposed to is an important factor in their subsequent language development.In their seminal study, Hart and Risley (1995) collected monthly 1-h naturalistic language samples from 42 families ofchildren aged 7–36 months over the course of approximately 2½ years, resulting in a total of 1318 h of language sample data.Language samples were collected during the evening meal to capture of time of day that was likely to include a high numberof parent–child communicative interactions. The study group included professional and working class families, and familiesliving in poverty. Results of this study indicated that children from professional families were exposed to the greatestnumber (2153 per hour) and diversity of words, followed by working class families (1251 per hour), and then families livingin poverty (616 per hour), and that professional families used more encouragement during interactions with their childrenthan their counterparts (Hart & Risley, 1995). Children of professional parents had developed more advanced vocabulariesthan their peers at age 3 and longitudinal data collected from this group indicated that scores at age 3 predicted language andliteracy skills at age 9 (Walker, Greenwood, Hart, & Carta, 1994).

A follow up to the Hart and Risley study was conducted more recently utilizing LENA digital voice recording technology(Gilkerson & Richards, 2009). Multiple full-day (12 h) naturalistic language samples were collected from 314 families oftypically developing young children once monthly for a period of 6–11 months, resulting in a total of over 18,000 h oflanguage sample data. Consistent with the results of Hart and Risley (1995) study, there was a positive relationship betweenthe number of adult words a child was exposed to and child expressive language skills (Gilkerson & Richards, 2009). Motherswho had graduated from college used significantly more words per hour and participated in more conversational turns withtheir child than mothers who did not earn a high school diploma although there was substantial variability. Unlike the Hartand Risley (1995) study, this study did not evaluate the quality of the adult language such as the use of encouragement orother adult language facilitation techniques. Studies evaluating the quantity of adult talk directed to children with variouslanguage-learning and developmental disabilities including ASD, have consistently demonstrated that these groups areexposed to an equal number of words during interactions with their mothers in their home environments as childrenwithout disabilities (Cunningham, Siegal, Van der Spuy, Clark, & Bow, 1985; Warren, Gilkerson, et al., 2010; Whitehurst et al.,1988; Wolchick, 1983).

In addition to the importance of exposure to a high quantity of adult language there is evidence that child languagedevelopment is enhanced by exposure to high quality adult language. For typically developing children, languagedevelopment has been shown to be positively related to the frequency of joint attention experiences (i.e., caregiver andchild sharing attention on shared object/activity) and decreased caregiver demandingness (Tomasello, Mannle, &Kruger, 1986; Tomasello & Todd, 1983). Typically developing children learn best when they are actively engaged ininteracting with their environment with non-demanding encouragement and scaffolding by a parent or caregiver. Thisis true for young children with disabilities as well (Dunst, Trivette, & Masiello, 2011; Siller & Sigman, 2002; Warren,Brady, Sterling, Fleming, & Marquis, 2010; Warren, Gilkerson, et al., 2010). Siller and Sigman (2002) investigated therelationship between parent ‘‘synchrony’’, defined by the authors as ‘‘the extent to which the caregiver’s behavior issynchronized with the child’s focus of attention and ongoing activity’’ (p. 77) and language outcomes of 18 childrenwith ASD, 18 children with developmental disabilities (DD), and 18 typically developing (TD) children matched onmental age over the course of 16 years. Parents who demonstrated increased synchrony and decreased demandingness(i.e., verbal interactions with the child that were not directive such as labels, comments, and sharing information)had children who developed better joint attention and language skills (Siller & Sigman, 2002). Parents of children withASD demonstrated equal levels of synchronized behavior as parents of children with DD and TD children (Siller &Sigman, 2002).

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439430

Warren, Brady, et al. (2010) investigated the relationship between maternal responsivity and child language outcomes in55 young children with Fragile X Syndrome (FXS). Maternal ‘‘responsivity’’ was defined as ‘‘a healthy, growth-producingrelationship characterized by warmth, nurturance, and stability as well as specific behaviors, such as contingent positiveresponses to child initiations.’’ (p. 54) and was measured based on maternal use of gestures, requests for verbal behavior,comments, and recodes or verbal statements of a child’s behavior. Mothers who were more responsive had children withbetter language outcomes in terms of total communication and number of different words (Warren, Brady, et al., 2010).Further, maternal responsivity partially mediated the effects of increased ASD symptomatology on language development;children with FXS and high ASD symptomatology with highly responsive mothers showed greater growth in expressive andreceptive language than similar children with less responsive mothers (Warren, Brady, et al., 2010).

Dunst et al. (2011) investigated the effects of interest based learning on several child outcomes including languagedevelopment. Mothers of 17 preschool aged children with ASD were asked to broadly identify their child’s interests and thenparticipate with their child in 8–10 of these activities weekly for 14–16 weeks. Members of the research team followed upwith participants regularly to assess fidelity of implementation. The group of 17, parent/child dyads were divided into lowand high interest groups (based on frequency of participation in high-interest activities) for data analysis. The high interestgroup performed better in language, cognitive, and social skills on the Developmental observation checklist system (DOCS;Hresko, Miguel, Sherbenough, & Burton, 1994) than the low interest group (Dunst et al., 2011). Consequently, there is agrowing body of research supporting the importance of naturalistic, child-centered approaches within the homeenvironment to promote more positive language outcomes in young children with disabilities (Kaiser & Roberts, 2013;Pajareya & Nopmaneejumruslers, 2011).

This brief review of home language environment literature provides substantial support for the importance of highquantity and quality of adult talk in terms of supporting positive language outcomes for typically developing children as wellas children with developmental disabilities including ASD. Prior research has focused predominantly on naturalisticinteractions between children and their mothers in home environments. However, research suggests that a vast majority ofyoung children with ASD spend a substantial amount of time each week participating in educational programs (M = 19.5 h;Bitterman, Daley, Misra, Carlson, & Markowitz, 2008).

1.2. School language environment

There are numerous publications describing general characteristics of quality educational programs for young childrenwith ASD including systematic instruction, highly trained teaching staff, structured environments, planned generalization ofskills, and specialized curriculum emphasizing language and communication (Iovannone, Dunlap, Huber, & Kinkaid, 2003;Kaiser, Hester, & McDuffie, 2001; National Autism Center, 2009; Paul, 2008). There are also many published studiessupporting the utility of a variety of intervention strategies to enhance the language and communication skills of youngchildren with ASD including naturalistic teaching (Wolery, 2000), peer mediated intervention (Garfinkle & Schwartz, 2002;Odom & Strain, 1986), picture exchange communication system (PECS; Ganz & Simpson, 2004; Yoder & Stone, 2006) anddiscrete trial teaching (Rogers & Vismara, 2008). However, there has been a paucity of research undertaken to describe howteachers are actually communicating with and eliciting communication from young children with ASD in their educationalenvironments on a day-to-day basis.

A recent study utilized LENA technology to evaluate characteristics of the school language environments of 14 preschoolaged children with ASD who attended ‘‘high quality’’ preschool programs (Dykstra et al., 2013). Results of this studyindicated that the study participants were exposed to a relatively high number of adult words, averaging 1700 per hour, withhigher adult word counts evident on days in which the children participated in speech-language therapy (Dykstra et al.,2013). This study did not include comparisons between home and school language environments or include analysis of thequality of adult language. A second study evaluating characteristics of the school language environment investigatedteacher’s responses to the communicative initiations of 8 young children with severe characteristics of ASD (Keen, Sigafoos,& Woodyat, 2005). The results of this study indicated that teachers responded verbally or physically to child initiations anaverage of 62% of opportunities, although there was much variability among teachers (Keen et al., 2005). Teachers were mostlikely to respond to children’s use of social conventions (M = 87%) and least likely to respond to child rejections (M = 51%). Athird study investigating characteristics of the school language environment, focused on identifying the frequency and typeof prompts (e.g., verbal prompt, modeling, physical prompt) teachers used to elicit communication from 34 minimally/non-verbal children with ASD, aged 3–16 years. Two hours of data was collected and analyzed for each student in a variety ofnaturalistic school settings. Teacher elicitations for communication ranged from 0 to 179, with a mean of 20.4 per 2 h, or oneevery 6 min (Chiang, 2009). Results suggested that the school language environments of these participants were relativelyimpoverished in terms of the number of teacher/child communicative interactions. The most common teacher strategies foreliciting student communicative behavior were verbal prompts and modeling. Teachers tended to initiate less frequentlywith students with a greater number of ASD symptoms (Chiang, 2009). Although Chiang (2009) study is useful in describingone piece of quantitative data from the school language environment, it is fairly limited in its scope, as it provides very littledata specific to the quality of the language environment.

This brief review of literature highlights that more is known about the home language environments of young childrenASD than the school language environments. However, the home language environment research has focused on eitherquantity or quality of language, but not both. Additionally, there have been no studies comparing characteristics of school

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 431

and home language environments of young children with ASD. This exploratory research aims to describe and comparequantitative and qualitative characteristics of the school and home language environment of young children with ASD.

1.3. Study questions

1. In what contexts are young children with ASD exposed to the greatest number of adult words?

Ta

Pa

A

A

V

V

V

C

C

C

No

in

sta

2. A

re young children with ASD being exposed to semantically rich language environments? 3. H ow often is language being directed to young children with ASD? 4. D oes complexity of adult language vary between language environments? 5. W hat types of utterances are used during interactions with young children with ASD? 6. W hat types of pragmatic functions are being used during interactions with young children with ASD?

2. Method

2.1. Participants

Participants were 10 male children ages 35–67 months at the beginning of the study who attended centered-basededucational programs. Participants were required to be diagnosed with ASD (i.e., autism, Asperger Syndrome, or PDD-NOS)by a medical professional, as reported by their parent, and to receive a score at or above the cutoff score at or above the cutofffor ASD on the Autism Diagnostic Observation Schedule (ADOS, Lord, Rutter, DiLavore, & Risi, 2001) that was administered aspart of the research protocol. All participants who met the inclusion criteria were included in this study. Participants wererecruited from programs serving preschool aged children with ASD in a suburban area of the Midwest, in the United States.All 10 children had delayed language development as assessed by the Vineland adaptive behavior scales – Second Edition(Sparrow, Cicchetti, & Balla, 2005) and Clinical evaluation of language fundamentals – Preschool, Second Edition (CELF-P2;Wiig, Secord, & Semel, 2004). Participant characteristics are summarized in Table 1.

Nine of the participants were Caucasian and one was Hispanic. All but one participant were from 2-parent families andthe mothers of all participants had completed a minimum of 12 years of schooling (graduated from high-school) with sevencompleting a secondary degree (associate’s degree or higher). Yearly family incomes ranged from below $25,000 to between$100,000 and $125,000. All of the participants received special education services and speech-language therapy in theireducational programs. Three participants also received 30 min of speech-language therapy outside of their school programs.

2.2. Preschool settings

Participants were recruited from four private (i.e., non-public school) preschool programs in the Midwest. Threeparticipants attended a program that served only children with ASD. All three of these children were in the same classroomthat served a total of 5 children. Three participants attended a program that served children identified with ASD and typicallydeveloping peers. All three children were in the same classroom that served a maximum of 12 students, half of whom wereidentified with ASD. Four participants attended programs that served children with various developmental disabilitiesincluding ASD in addition to typically developing peers. These four participants were served in three different classrooms,each of which had a maximum of 16 children, at least half of whom were identified with developmental disabilities.

Program curriculum varied across settings. One program serving three participants was described as using a combinationof the Treatment and Education of Autistic and Related Communication handicapped CHildren (TEACCH) andDevelopmental, Individual Difference, Relationship-based (DIR1) models for intervention. One program serving thereparticipants was described as using a combination of the SCERTS1 and HighScope models for intervention. The other twoprograms, serving the remaining four participants were described as being ‘‘eclectic’’ in their approaches.

ble 1

rticipant age, ADOS, Vineland adaptive behavior scales language scores, and CELF-P language scores at entry of study.

Participants

1 2 3 4 5 6 7 8 9 10

ge in months 35 39 46 57 58 58 60 60 67 67

DOS total score 11 17 18 21 21 19 11 11 19 19

ineland receptive language AE 18 3 15 23 14 23 26 26 9 29

ineland expressive language AE 13 5 16 30 9 34 53 30 6 47

ineland communication SS 69 44 63 72 54 76 87 76 40 79

ELF-P receptive language SS – – – 81 – 59 59 77 – –

ELF-P expressive language SS – – – 69 – 67 61 59 – –

ELF-P core language SS – – – 71 – 59 65 63 – –

te: ADOS scores of 12 or above indicate presence of autism, scores of 7 or above indicate presence of ASD. Although participant 10 was able to participate

standardized assessment using the CELF-P, it was not administered and scores are not reported. Participants 1, 2, 3, 5, and 9 were not able to participate in

ndardized language assessment. Vineland age equivalency (AE) scores reported in months. CELF-P standard scores, M = 100, SD = 15.

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439432

Each classroom was lead by a master’s level teacher and a minimum of two para-educators. Occupational and speech-language therapists worked directly with the students within the classroom setting and/or in a separate setting inaccordance to their Individualized Education Plans (IEPs). Eight participants attended school 5½ h per day, 5 days a week andthe other two students attended 2½ h per day, 5 days per week.

2.3. Measures

2.3.1. LENATM

Naturalistic language samples were collected using LENA Digital Language Processors (DLPs). The LENA DLP is a recordingdevice that is about the size of a small cell phone that fits into a pocket of specially designed clothing (e.g., vest, shirt, oroveralls) and is used to record vocal interactions between the target child and his or her communicative partners withinnaturalistic environments (Gilkerson & Richards, 2009). The audio recordings are then downloaded into the LENA softwareand analyzed using advanced algorithms and statistical modeling to provide the researcher, clinician or parent with a varietyof data including; adult word count (AWC), child turn count (CTC), and child vocalization count (CVC; Gilkerson & Richards,2009). Data is presented graphically and can be viewed by day, or by 1-h or 5-min segments.

2.3.2. Systematic Analysis of Language Transcripts (SALT)

Naturalistic language samples were analyzed using SALT software (Miller, Andriacchi, & Nockerts, 2011). Audio files weretranscribed within the SALT software using standard conventions to provide the researchers with a variety of data includingadult and child Mean Length of Utterance (MLU), utterance types, and number of utterances, words, and different words.Additionally, a coding system was developed specifically for this study to allow for the identification and subsequentanalysis of the number and percent of adult utterances that were not directed to the target child, yes/no (opposed to open-ended questions), and different pragmatic functions utilized by adults during their interactions with the study participants.To allow for identification of the number and percent of utterances directed toward the target child each adult utterance thatwas not directed toward the target child was assigned one of four unique codes indicating that the utterance was directedtoward (a) another child; (b) an adult; (c) an unidentifiable target; or (d) a group of children. The percent of utterancesdirected to the target child was calculated by subtracting the non-child directed utterances from the total number of adultutterances, dividing the difference by the total number of utterances and multiplying by 100. To allow for differentiation ofyes/no and open-ended questions, all yes/no questions were coded. The percent of yes/no questions was calculated bydividing the number of yes/no questions by the total number of questions and multiplying by 100. Finally, to identify thepragmatic function of adult statements directed to the target child, utterances were assigned one of eight pragmatic functioncodes. The pragmatic functions coded in this study are summarized in Table 2. Not infrequently, an utterance contained morethan one pragmatic function and when this occurred, the utterance was coded with the last pragmatic function and was alsoassigned a second code to indicate that the utterance contained ‘‘multiple pragmatic functions’’. For example the utterance;‘‘Give me your finger, I made you a pretend band aid for a pretend booboo.’’ contains 2 pragmatic functions, first a directivefor motor behavior followed by a comment. This utterance was coded as a comment and an utterance containing multiplepragmatic functions. The decision was made to code the final function of utterances with multiple functions, as the initialfunction was most frequently a response to a previous child utterance and the final function was most frequently the ‘‘pieceof talk’’ to which the adult expected a child response.

2.4. Procedure

This study received approval from the appropriate institutional review boards and informed written consent wasobtained from each participant’s parent(s) prior to his involvement in the study. Additionally, although neither verbal norwritten assent was obtained from the participants due to their age and disability status, if a participant refused to wear theclothing (e.g., vest holding the DLP) associated with the study after three trials in a three-day period, refusal of assent was

Table 2

Pragmatic function codes, descriptions, and examples.

Pragmatic function Description; example

Comment Statement about a person’s affective or physical state, an object or event in the environment, or event that

is remote that the child participated in: ‘‘That is a silly dog.’’, ‘‘You played on the playground.’’

Directive for motor behavior Statement used to direct motor behavior: ‘‘Sit down.’’, ‘‘Touch. . ..’’, ‘‘Give me . . .’’

Directive for prohibition Statement used to inhibit a motor behavior: ‘‘Stop.’’, ‘‘No.’’

Labeling Statement used to label an object, person, or activity: ‘‘triangle’’, ‘‘blue’’

Requests for verbal behavior Statement that requests verbal behavior: ‘‘Say mommy.’’, ‘‘Tell me. . .’’, and intonation prompts

Sharing information Statement about a person, activity or object that is not obvious to the child or that he or she does not

already know: ‘‘You are going to work with x today.’’, ‘‘Dinosaurs lived a long time ago.’’ or response to

a child’s question

Verbal correction Statement that provides a correction to child’s verbal output: ‘‘No, that is green.’’, ‘‘It is getting shorter not longer.’’

Verbal reinforcement Statement that provides positive feedback to child: ‘‘Awesome work,’’, ‘‘Yes that is Cinderella.’’, repeating the

child’s utterance to reinforce

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 433

assumed. One child refused to participate and was not included in this study.After consent was obtained, an initial data collection session was scheduled with the participant’s parent; in all cases the

respondent was the child’s mother. The participant’s parent was interviewed using the Vineland adaptive behavior scales –Interview Edition (Sparrow et al., 2005) and was asked to complete a background information form and the Parenting stressindex (PSI; Abidin, 1995). Data presented in this study was collected as part of a larger investigation examining languageenvironment characteristics and expressive language development of preschoolers with ASD, and analysis of PSI data is notincluded as part of this study. After the initial data collection session with the parent, the CELF-P2 and ADOS (Lord et al.,2001) were administered to the participants on separate days, within 2 weeks of each other. Each participant was evaluatedindividually in a quiet therapy room in his school or in the lab of the primary investigator. Five participants did notsuccessfully complete the trial items on the CELF-P (Wiig et al., 2004) and therefore no scores were reported for theseparticipants. Each testing session lasted between 30 and 60 min. The investigator or a trained undergraduate or graduatestudent administered all tests. Parent and child assessments were repeated 10.5–11 months after the initial data collectionsessions.

Naturalistic language samples were obtained via LENA technology for each study participant at approximately 3-monthintervals (range 2–6 months) for one year, resulting in a 2–4 language samples for each participant. A total of 31 languagesamples are included in this study. Although participants were asked to wear the DLP in both home and schoolenvironments, this did not always occur. Twenty-two language samples represent full day recordings including both schooland home. The remaining nine samples were obtained from either home or school but not both on the same day. Within theschool environment the majority of language samples were recorded during ‘‘circle’’ group and direct instruction periodswith the target child. Within the home environment, the majority of samples were recorded during ‘‘play’’ interactions withfamily members and instruction periods emphasizing academic and self-help skills. Only participants with a minimum oftwo school and home samples were included in data analysis for study. The duration of recordings ranged from 2.5 to 15.4 h.

Language samples were downloaded into the LENA software and the resulting bar graphs were visually analyzed in the 5-min segment view. The primary investigator identified the three consecutive 5-min segments with the highest number ofadult words, as indicated by LENA AWC data, for all school and home samples. Segments with the highest AWC were selectedfor SALT analysis as they were hypothesized to contain the ‘‘richest’’ examples of child-directed adult language. Previousresearch has used a similar approach in evaluating characteristics of home language environments (Hart & Risley, 1995).

The primary investigator and trained undergraduate and graduate students transcribed audio files. Utterance boundarieswere determined by a change in turns (e.g., child turn, adult turn) or a 2 s pause between utterances from the samecommunicative partner (e.g., adult utterance, 2 s pause, adult utterance), consistent with recommendations from Miller(1981, p.14) and procedures utilized in other language sampling research (Rice, Smolik, Thompson, Rytting, & Blossom,2010). Coding conventions were consistent with those described in the SALT manual (Miller et al., 2011) and the specificcoding procedure, described above, was developed and utilized for this project. Two practice transcripts were scored by boththe lead and second author with discrepancies discussed until 100% agreement was reached. Student transcribers weretrained using practice transcripts until they reached a minimum of 85% agreement in total utterances with the lead or secondauthor. Once they reached 85% reliability they were assigned novel 15-min samples to transcribe. For each set of 10completed transcripts, three were randomly chosen for reliability checks. The second author provided inter-rater reliabilitychecks by listening to the audio files as she read each transcript and indicating agreement on utterance boundaries, content,adherence to coding conventions, and assignment of specialized codes for each utterance. Utterances identified as beingdiscrepant in any of the above described characteristics were identified and reliability was calculated by dividing the numberof utterances with discrepancies by the total number of utterances and multiplying by 100. Inter-rater reliability was 97% fortranscripts included in this study.

3. Results

3.1. In what contexts are young children with ASD exposed to the greatest number of adult words in school and home

environments?

Twenty-one of the 26 school language samples across participants with the highest AWC contained adult-lead activitiesincluding 12 ‘‘circle time’’ segments and nine 1-to-1 instruction segments. Four samples included adult–child interactionsduring lunch, and one segment included adult–child and child–child interactions during a shared reading activity including apeer. Although school segments captured predominantly adult-lead/adult-directed activities, there were clear examples ofthe adult building on a child interest within activities (e.g., child expressing his interesting in getting an M&M candy for hishard work and the subsequent discussion about the details of this request including the color and amount of candy the childdesired).

Thirteen of the 26 home language samples across participants with the highest AWC occurred during play/conversationwith a family member or family members, five occurred within typical family routines (e.g., mealtime, driving home fromschool, getting ready for bed), seven occurred during 1 to 1 ‘‘instruction times’’ emphasizing language/academic skills, andactivities of daily living with a parent or caregiver/tutor, and one segment contained adult to adult conversation in closeproximity to, but not including, the child. The home language samples more frequently represented child-oriented/directedinteractions than those obtained in the school environment.

Table 3

LENA AWC per hour and 15-min transcribed segments, and SALT number of adult words and different words, in school and home environments.

School Home

M SD M SD

AWC per hour 2132 655 1307 592

AWC 15-min school segment 1493 394 1042 469

SALT number of adult words 1158 247 836 402

SALT number different adult words 279 54 239 91

Note: SALT data is based on 15 min transcribed samples.

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439434

3.2. Are young children with ASD being exposed to semantically rich language environments?

LENA data including mean adult word count (AWC) per hour and 15-min transcribed segments and SALT dataincluding mean number of adult words and number of different words per 15-min transcribed segments, were usedevaluate the semantic richness of the school and home language environments. Young children with ASD were exposedto relatively high numbers of adult words and diversity of words in both their home and school environments although adifference between contexts was identified. T-Test statistics were conducted to compare LENA AWC per hour and 15-mintranscribed segments, and SALT number of adult words and number of different words, between school and homeenvironments. There was a significant difference in the AWC between school and home such that participants wereexposed to a greater number of words per hour in their school environments t(50) = 5.25, p < .0001, d = 1.48 and during15 min transcribed segments t(50) = 3.80, p = .00104, d = 1.07. There was also a significant difference between the SALTnumber of adult words spoken during 15-min transcribed segments between school and home environments. Againparticipants were exposed to greater number of words in their school environment, t(49) = 3.33, p = .002, d = 95. Themagnitude of these differences was large based on the Cohen’s effect size values. The difference between the SALTnumber of different adult words spoken during 15-min transcribed segments at school and home was found to be non-significant (see Table 3).

3.3. How often is language being directed to young children with ASD in school and home environments?

Coded SALT data was used to evaluate the number and percent of utterances directed toward the target child during 15-min highest AWC segments at school and home. Percentage data is reported to compensate for variability in the number oftotal utterances in each transcribed segment. A majority of adult utterances were directed toward the target children ratherthan to other adults or children in both school (50%) and home (62%) environments. In the school setting the percent ofutterances directed toward the target child ranged from 9 to 100% and in the home setting ranged from 14 to 100%. There wasa greater percentage of adult utterances directed to other adults in the home setting than the school setting t(49) = 2.74,p = .009, d = .78, and a greater percentage of utterances directed to groups of children in the school setting than the homesetting t(49) = 3.58, p = .001, d = 1.02. The magnitude of these differences was large based on the Cohen’s effect size values.There were no differences in the percentage of utterances directed to another child or an unidentifiable target betweensettings.

3.4. Does complexity of adult language vary between language environments?

Analysis of SALT adult Mean Length of Utterance (MLU) data from the 15-min segments with the highest AWC at schooland home was utilized to evaluate language complexity. Adult MLU in morphemes was the same in both settings (7.80) andsubstantially higher than child MLU both in the school (1.98) and the home (2.50) settings.

3.5. What types of utterances are used during interactions with young children with ASD?

SALT analyses of the 15-min segments with the highest AWC at school and home were utilized to identify the numberand percentages of different types of utterances used by adults during their interactions with young children with ASD.Percentage data is reported to compensate for variability in the number of total utterances in each transcribed segment.Utterance types included statements, questions, yes/no questions, exclamations, intonation prompts, and imitations. Thepredominant adult utterance type used during interactions in both school and home environments was statementsfollowed by questions. Adults in the home environment were more likely to ask yes/no questions, t(49) = 2.98, p = .005,d = .85 and use exclamations t(49) = 2.04, p = .05, d = .58 than adults in the school environment. Adults in the schoolenvironment were more likely to use intonation prompts t(49) = 2.60, p = .01, d = .74 than adults in the home environment.Exclamations, intonation prompts, and imitations were used infrequently in both the school and home environments (seeTable 4).

Table 4

Percentage of utterance types used during interactions in school and home environments.

School Home

M SD M SD

Statements 67.36 11.57 68.04 6.94

Questions 25.84 9.10 24.77 8.04

Yes/no questions 26.56 20.34 43.46 20.21

Exclamations 2.20 2.18 4.12 4.18

Intonation prompts 2.64 4.21 .42 1.14

Imitations .96 1.13 .58 1.13

Note: Percentages do not add up to 100 due to rounding.

Table 5

Percentage of pragmatic functions used during interactions in school and home environments.

School Home

M SD M SD

Comments 31.68 19.27 35.00 20.21

Directives for motor behavior 24.64 16.40 23.42 16.87

Directives for prohibition 2.16 3.63 5.42 8.86

Labels 2.72 3.32 2.54 3.86

Requests for verbal behavior 7.72 7.36 4.15 7.48

Social niceties 7.92 7.25 12.08 19.85

Sharing information 6.68 10.60 6.50 7.44

Verbal corrections 2.80 5.15 .92 3.12

Verbal reinforcement 13.16 10.48 10.27 10.70

Multiple pragmatic functions 30.28 20.55 22.92 14.60

Note: Percentages do not add up to 100 due to rounding.

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 435

3.6. What types of pragmatic functions are being used during interactions with young children with ASD?

Coded SALT data was used to analyze the types of pragmatic functions used by adults during interactions with youngchildren with ASD. Percentage data is reported to compensate for variability in the number of total utterances in eachtranscribed segment. As can be seen in Table 5, the most frequently used functions in both settings were comments anddirectives for motor behavior. Far smaller percentages of utterances were used to prohibit behavior, label, provide verbalcorrections, share information, or participate in social niceties/routines. There was little variability in functions betweenschool and home. There was a somewhat greater percentage of directives for prohibition and use of social niceties/routines inthe home environments and somewhat greater use of requests for verbal behavior in the school environments, but none ofthese differences approached statistical significance. Of the total number of adult utterances coded for pragmatic function,57% of school and 60% of home contained pragmatic functions that have been identified as being facilitative of child languageincluding comments, labels, requests for verbal behavior, sharing information, and social niceties and routines (Girolametto,Weitzman, & Lieshout, 2000; Kaiser & Roberts, 2013).

4. Discussion

The purpose of this study was to begin to characterize and compare the quantity and quality of language used by adultsduring interactions with young children with ASD in their school and home environments. LENA technology allowed for thecollection of naturalistic language samples where specific segments could be selected based on certain criteria (e.g., highestAWC) and then analyzed using both LENA and SALT technology. In both school and home contexts, adults directed themajority of their language to the young children with ASD (did not just talk in proximity to them) and used a large numberand variety of words. In fact, based on LENA AWC data, the children in this study were exposed an equal number of words perhour in their school environment as the children of ‘‘professional’’ parents in the Hart and Risley (1995) study, and more intheir home environments than the children of college graduates in the Gilkerson and Richards (2009) study. This studyidentified a much more enriched language environment than that experienced by the school-aged children with ASD withsevere characteristics of ASD described by Chiang (2009) with whom, teachers were observed to interact with on averageonly once every 6 min.

In addition to being exposed to a large number of adult words, the children who participated in this study were alsoexposed to high quality language. One measure of quality language is the degree to which it is child-centered. In bothenvironments a mean of at least 50% of adult utterances were directed to the target child although there was a high degree ofvariability in this data among participants. It is concerning that in some segments as little as 9% of utterances were target-child directed, especially given the fact that segments were chosen to represent the greatest amount of adult talk in closeproximity to each participant. It will be important for future research to identify child and/or program characteristics that

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439436

may have influenced these results including child language level and intervention approaches utilized within the classroomsetting.

Another measure of quality language is the degree to which it facilitates child expressive language use and/orparticipation in reciprocal conversation. Approximately 25% of adult utterances at school and home were questions and amajority of those questions were open-ended, suggesting that they were used to encourage more complex child expressivelanguage responses than those required of yes/no questions. There was some variability in utterance type use betweenschool and home environments including increased use of intonation prompts to request verbal behavior in schoolenvironments and use of yes/no questions in home environments. It is unclear what, if any, effect the use of different adultutterance types has on child language use. This is an area in need of specific study. The fact that the majority of adultstatements expressed pragmatic functions including commenting and sharing information that have been found to besupportive of expressive language outcomes in young children with and without disabilities (Venuti, de Falco, Esposito,Zaninelli, & Bornstein, 2012; Weiss, 1981) is a positive finding. It is however, somewhat concerning that over 25% of adultutterances were used to direct and prohibit behavior rather than interact. Data comparing the language environments ofyoung children with ASD to those of peers with other developmental disabilities and peers who are typically developing willbe needed to determine if this is ‘‘typical’’ of adult interactions with young children generally or specific to young childrenwith ASD.

Overall, this study highlighted the many similar characteristics between school and home language environments foryoung children with ASD including adult language complexity, the percent of utterances directed to the target child, andthe types of pragmatic functions used during interactions. One difference in the school and home language environmentsof young children with ASD included the types of interactions that predominated. School segments were much morelikely to capture adult-lead/directed activities while home environments were more likely to capture routine- and play/conversation-based activities. Given the relatively strong support for the efficacy of child-centered, naturalisticinterventions to support language and communication skills of young children with disabilities including ASD in earlychildhood programs (Harjusola-Webb & Robbins, 2012; Ingersoll, Lewis, & Kroman, 2007; Kaiser et al., 2001; Yoder &Stone, 2006) it was somewhat surprising that none of the school segments with the highest AWC occurred during childdirected play.

Based on past research findings, it would be expected that more child-centered contexts would lead to more positivechild-language outcomes. As home language samples more frequently represented interactions in child-centered contextsand school language samples more frequently represented interactions in adult-lead contexts, post hoc analysescomparing child language characteristics including LENA CTC and CVC for 15 min transcribed segments and SALT child MLUin morphemes and number of words between environments were conducted. Participants had a higher number of meanturns (31 versus 28), vocalizations (72 versus 45), and words (114 versus 104) in their home environments than in theirschool environments and also used a higher MLU (2.50 versus 1.98). Although none of these differences reachedsignificance, they do support previous findings of better child language outcomes in more child-centered interactions(Siller & Sigman, 2002; Warren, Brady, et al., 2010). The children in this study used a higher MLU and vocalized more in theirhome environments, suggesting that the nature of the interactions (e.g., more play-based and child-directed) may havebeen facilitative of greater child language complexity and vocal play. Given the criteria for segment selection in this study(15-min segments of highest adult word count), it would hold to reason that adult-directed 1:1 educational segmentswould predominate. However, transcription and analysis of 15-min segments based on highest level of child output(vocalizations or turns) may yield important information regarding the relationship between child expressive language useand context.

A second difference highlighted by this study was that the children were exposed to a greater number of words in theirschool environments compared to their home environments. Prior research has suggested that family SES and maternaleducation level were positively related to child language outcomes (Gilkerson & Richards, 2009; Hart & Risley, 1995). It ispossible that the difference in quantity of adult talk observed in this study was influenced by the advanced education level ofthe lead teacher and graduate-level specialists who interacted with the participants each day. It is also possible thatdifferences were influenced by the contexts in which the participants were interacting (i.e., more adult versus child directed)or differences in the language focus within environments; that while educational programs focused more on content/semantics, home environments focused more on pragmatics and sustained interaction. Certainly, more detailed analysis ofthis is warranted. It is a positive finding, however, that participants were exposed to a high number and diversity of word inboth environments.

An unexpected result of this research was the finding that many adult utterances contained multiple pragmatic functions.Often adults responded or initiated with a child making a comment, followed by a directive for motor behavior, for example.Given the nature of the language learning difficulties of young children with ASD, it is possible that adult pragmaticcomplexity results in greater processing demands being placed on the child. Pragmatic complexity has not previously beendiscussed in the research literature and it is unknown how controlling for pragmatic complexity may or may not providebenefit to the child linguistically, pragmatically, educationally, or behaviorally. The pragmatic complexity identified in thisstudy is similar in a way to the language and communication facilitating technique of utilizing turnabouts (e.g., providing aresponse to a child’s initiation followed by a ‘‘press’’ for additional ‘‘talk) to encourage additional child verbalization, in thatmany of the utterances that contained pragmatic complexity in this study consisted of adult praise or comment followed by a‘‘press’’. The main difference was that the most frequent type of ‘‘press’’ utilized by the adults in this study tended to be

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 437

directives for motor behavior rather than for verbal behavior. Future studies will be needed to explore if and how pragmaticcomplexity is related to child expressive language output in young children with ASD as well as typically developing childrenand children with other disabilities.

Although the results of this study provide beginning evidence that young children with ASD are exposed to rich languageenvironments in both school and home settings, there are limitations that must be considered. The first limitation has to dowith the small sample size that limits generalizability of results. The sample also included only young children with ASD,limiting interpretation of results in relationship to how the language environments of the children included in this studycompared to ‘‘typical’’ school and home language environments for preschool aged children in general. It will be importantfor future studies to include a greater number of children and compare children with ASD to those from other populations(e.g., language delayed, typically developing) matched according to level of functioning and/or age. The sample also onlyrepresented children from private educational settings, which although are representative of some educational programs foryoung children with ASD, do not represent the most ‘‘typical placement’’ (i.e., pubic school preschool programs). Publicschool preschool programs generally have a lower teacher to student ratio than the programs described in this study and amore diverse student population in terms of the inclusion of typically developing peers and students with varyingdisabilities, many of which are less severe than ASD. It will be important for future studies to include children who areparticipating in public school preschool programs and to compare results across educational program utilizing differentevidenced-based intervention strategies to better evaluate the range of language environments provided to young childrenwith ASD.

A second limitation of this study concerns the procedure for selecting language segments for analysis. Given that thepurpose of this study was to examine adult language directed to young children with ASD, 15-min segmentsrepresenting the highest number of adult words according to the LENA AWC data were selected for analysis. It is likelythat this decision accounted for the high frequency of selected samples in the school setting representing segments witha relatively high degree of adult direction. It will be important in the future to additionally select segments that mightbe more likely to represent more child-initiated or directed interactions to determine if there are differences in theamount and/or type of adult language used in that context compared to direct instruction. Furthermore, it will beimportant to determine if language used within different contexts (e.g., direct instruction, play) is more or lessfacilitative of more positive child expressive language use. Additionally, this study did not compare characteristics ofthe school and home language environments in relation to specific child or environmental characteristics, such as childlanguage level or intervention approach used within the classroom. I also did not identify language environmentcharacteristics that are predictive of positive child language outcomes over time. Each of these aspects would beimportant to examine in future research.

5. Conclusion

Overall, the results of this preliminary investigation of the school and home language environments of young childrenwith ASD suggest that the young children with ASD who participated in this study were exposed to a high quantity ofadult language in terms of the number and diversity of words used with them during interactions, and high quality adultlanguage in terms of the percent that is child-directed and facilitative of positive child language outcomes, which are verypositive findings. In addition the data provide beginning evidence of differences in language environments in terms of thecontexts in which interactions are taking place and the types of utterances used during interactions that may beimportant in terms of child language outcomes. The combination of technologies used in this study made it possible torelatively easily collect full-day naturalistic language samples, identify communication segments of interest, and analyzequantitative and qualitative aspects of adult and child language. Future studies utilizing this combination of technologiesis needed to characterize and compare language environment characteristics of larger and more diverse groupsof children and to identify language environment characteristics that predict most positive child language outcomesover time.

Acknowledgements

We greatly acknowledge partial support for this project from a Federal Education Grant awarded to the HattieLarlham Research Foundation, Mantua, OH and the University Research Council at Kent State University. We wouldalso like to acknowledge the graduate and undergraduate students who so graciously and diligently transcribed audiofiles.

Conflict of interest statement

The authors received financial support for this project from a Federal Education Grant awarded to the Hattie LarlhamResearch Foundation, Mantua, OH and the University Research Council at Kent State University. There are no nonfinancialrelationships to disclose.

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439438

Appendix A. Continuing education

CEU Questions1. True or False, SALT technology was used to collect full-day naturalistic language samples?2. True or False, Participants were exposed to a greater number of words per hour and per 15 min transcribed segment in their

school environments3. The most commonly used utterance type used during interactions at both home and school were:

a. S

tatements b. O pen-ended questions c. Y es/no questions d. E xclamations

4. The utterance type used more frequently in the school environment than the home environment was:

a. In

tonation prompts b. Y es/no questions c. Im itation d. E xclamations

5. The pragmatic functions used most frequently in the home environment were:

a. C

omments and directives for motor behavior b. D irectives for motor behavior and directives for prohibition c. C omments and sharing information d. L abeling and directives for motor behavior

References

Abidin, R. (1995). Parenting stress index (3rd ed.). Torrance, CA: Western Psychological Services.American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., Text Revision (IV-TR)). Washington, DC: American

Psychiatric Association.Baron-Cohen, S., Cox, A., Baird, G., Swettenham, J., Nightingale, N., Morgan, K., et al. (1996). Psychological markers in the detection of autism in infancy in a large

population. British Journal of Psychiatry, 168(2), 158–163.Bitterman, A., Daley, T. C., Misra, S., Carlson, E., & Markowitz, Y. J. (2008). A National sample of preschoolers with autism spectrum disorders: Special education

services and parent satisfaction. Journal of Autism and Developmental Disorders, 38, 1509–1517.Chiang, H. (2009). Naturalistic observations of elicited expressive communication of children with autism. Autism, 13(2), 165–178.Cunningham, C. E., Siegal, L. S., Van der Spuy, H. I. J., Clark, M. L., & Bow, S. J. (1985). The behavioral and linguistic interactions of specifically language-delayed and

normal boys and their mothers. Child Development, 56, 1389–1403.Dunst, C. J., Trivette, C. M., & Masiello, T. (2011). Exploratory investigation of the effects of interest-based learning on the development of young children with

autism. Autism, 15(3), 295–305.Dykstra, J., Sabatos-De Vito, M. G., Irvin, D. W., Boyd, B. A., Hume, K. A., & Odom, S. L. (2013). Using the language environment analysis (LENA) system in preschool

classrooms with children with autism spectrum disorders. Autism http://dx.doi.org/10.1177/1362361312446206Ganz, J. B., & Simpson, R. L. (2004). Effects on communicative requesting and speech development of the picture exchange communication system in children with

characteristics of autism. Journal of Autism and Developmental Disorders, 34(4), 395–409.Garfinkle, A. N., & Schwartz, I. S. (2002). Peer imitation: Increasing social interactions in children with autism and other developmental disabilities in inclusive

preschool classrooms. Topics in Early Childhood Special Education, 22(1), 26–38.Gilkerson, J., & Richards, J. A. (2009). The power of talk (2nd ed.). Impact of adult talk, conversational turns, and TV during the critical 0–4 years of child

developmentDenver, Colorado: Infoture Inc.Girolametto, L., Weitzman, E., & Lieshout, R. (2000). Directiveness in teacher’s language input to toddlers and preschoolers in day care. Journal of Speech, Language,

and Hearing Research, 43, 1101–1114.Halliday, M. A. K. (1978). Language as a social semiotic: The social interpretation of language and meaning. Baltimore: University Press.Harjusola-Webb, S., & Robbins, S. H. (2012). The effects of teacher-implemented naturalistic intervention on the communication of preschoolers with autism.

Topics in Early Childhood http://dx.doi.org/10.1177/0271121410397060Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore, MD: Brookes.Hresko, W., Miguel, S., Sherbenough, R., & Burton, S. (1994). Developmental observation checklist system: A systems approach to assessing very young children:

Examiner’s manual. Texas: Pro-ed.Ingersoll, B., Lewis, E., & Kroman, E. (2007). Teaching the imitation and spontaneous use of descriptive gestures in young children with autism using a naturalistic

behavioral intervention. Journal of Autism and Developmental Disorders, 37, 1446–1456.Iovannone, R., Dunlap, G., Huber, H., & Kincaid, D. (2003). Effective educational practices for students with autism spectrum disorders. Focus on Autism and other

Developmental Disorders, 18(3), 150–163.Kaiser, A. P., Hester, P. P., & McDuffie, A. S. (2001). Supporting communication in young children with developmental disabilities. Mental Retardation and

Developmental Disabilities Research Reviews, 7, 143–150.Kaiser, A. P., & Roberts, M. Y. (2013). Parent-implemented enhanced milieu teaching with preschool children who have intellectual disabilities. Journal of Speech

Language and Hearing Research, 56(1), 295–309.Keen, D., Sigafoos, J., & Woodyatt, G. (2005). Teacher responses to the communicative attempts of children with autism. Journal of Developmental and Physical

Disabilities, 17(1), 19–33.

S. Burgess et al. / Journal of Communication Disorders 46 (2013) 428–439 439

Lord, C., Rutter, M., DiLavore, P. C., & Risi, S. (2001). The autism diagnostic observation schedule (ADOS). Torrance, CA: Western Psychological Services.Miller, J. F. (1981). Assessing language production in children. Baltimore, MD: University Park Press.Miller, J. F., Andriacchi, K., & Nockerts, A. (2011). Assessing language production using SALT software: A clinician’s guide to analysis. Madison, WI: SALT, LLC.Murray, A., Johnson, J., & Peters, J. (1990). Fine-tuning of utterance length to preverbal infants: Effects on later language development. Journal of Child Language, 17,

511–525.National Autism Center. (2009). National standards report. Randolph, MA: National Autism Center.Odom, S. L., & Strain, P. S. (1986). A comparison of peer-initiation and teacher-antecedent interventions for promoting reciprocal social interactions of autistic

preschoolers. Journal of Applied Behavior Analysis, 19, 58–72.Osterling, J. A., Dawson, G., & Munson, J. A. (2002). Early recognition of 1-year-old infants with autism spectrum disorder versus mental retardation. Developmental

Psychology, 14(2), 239–251.Pajareya, K., & Nopmaneejumruslers, K. (2011). A pilot randomized controlled trial of DIR/FloortimeTM parent training intervention for preschool children with

autistic spectrum disorders. Autism, 15(5), 563–577.Paul, R. (2008). Interventions to improve communication. Child and Adolescent Psychiatric Clinics of North America, 17(4), 835–840 http://dx.doi.org/10.1016/

j.chc.2008.06.011Rice, M., Smolik, D., Thompson, T., Rytting, N., & Blossom, M. (2010). Mean length of utterance levels in 6-month intervals for children 3 to 9 years with and

without language impairments. Journal of Speech Language and Hearing Research, 53(2), 333–349.Rogers, S. J., & Vismara, L. A. (2008). Evidence based comprehensive treatments for early autism. Journal of Clinical Child and Adolescent Psychology, 37(1), 8–38.Siller, Y., & Sigman, M. (2002). The behaviors of parents of children with autism predict subsequent development of their children’s communication. Journal of

Autism and Developmental Disorders, 32(2), 77–89.Sparrow, S. S., Cicchitti, D. V., & Balla, D. A. (2005). Vineland adaptive behavior scales (2nd ed.). San Antonio, TX: Pearson.Tomasello, M., Mannle, S., & Kruger, A. (1986). Linguistic environment of 1- to 2-year-old twins. Developmental Psychology, 22, 169–176.Tomasello, M., & Todd, J. (1983). Joint attention and lexical acquisition style. First Language, 4, 197–212.Venuti, P., de Falco, S., Esposito, G., Zaninelli, M., & Bornstein, M. H. (2012). Maternal functional speech to children: A comparison of autism spectrum disorder,

Down Syndrome, and typical development. Research in Developmental Disabilities, 33, 506–517.Walker, D., Greenwood, C., Hart, B., & Carta, J. (1994). Prediction of school outcomes based on early language production and socioeconomic factors. Child

Development, 65, 606–621.Warren, S. F., Brady, N., Sterling, A., Fleming, K., & Marquis, J. (2010). Maternal responsivity predicts language development in young children with fragile x

syndrome. American Association on Intellectual and Developmental Disabilities, 15(1), 54–75.Warren, S. F., Gilkerson, J., Richards, J. A., Oller, D. K., Xu, D., Yapanel, U., et al. (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, 555–569.Weismer, S. E., Lord, C., & Esler, A. (2010). Early language patterns of toddlers on the autism spectrum compared to toddlers with developmental delay. Journal of

Autism and Developmental Disorders, 40, 1259–1273.Weiss, R. S. (1981). INREAL intervention for language handicapped and bilingual children. Journal of the Division for Early Childhood, 4, 40–52.Wetherby, A. M., Woods, J., Allen, L., Cleary, J., Dickinson, H., & Lord, C. (2004). Early indicators of autism spectrum disorders in the second year of life. Journal of

Autism and Developmental Disorders, 34(5), 473–493.Whitehurst, G. J., Fischell, J. E., Lonigan, C. J., Valdez-Manchaca, M. C., DeBaryshe, B. D., & Caulfield, M. B. (1988). Verbal interaction in families of normal and

expressive-language delayed children. Developmental Psychology, 24(5), 690–699.Wiig, E. H., Secord, W. A., & Semel, E. (2004). Clinical evaluation of language fundamentals Preschool (2nd ed.). Needham Heights, MA: Allyyn & Bacon.Wolchick, S. (1983). Language patterns of parents of young autistic and normal children. Journal of Autism and Developmental Disorders, 13(2), 167–180.Wolery, M. (2000). Recommended practices in child-focused interventions. In S. Sandall, M. McLean, & B. Smith (Eds.), DEC recommended practices in early

intervention/early childhood special education (pp. 29–38). Longmont, CO: Sopris West.Yoder, P., & Stone, W. L. (2006). Randomized comparison of two communication interventions for preschoolers with autism spectrum disorders. Journal of

Consulting and Clinical Psychology, 74(3), 426–435.