summary 2 pages
hsj
PeerInteractionsinPreschoolInclusiveClassrooms-TheRolesofPragmaticLanguageandSelf-Regulation.pdf
https://doi.org/10.1177/0014402919828364
Exceptional Children 2019, Vol. 85(4) 432 –452 © The Author(s) 2019 DOI: 10.1177/0014402919828364 journals.sagepub.com/home/ecx
Original Research
The number of inclusive preschool class- rooms has been rising in the United States since the 1960s in response to legislative man- dates promoting placement of children with disabilities in the least restrictive environ- ment. According to the National Center for Education Statistics (2016), approximately 740,000 preschool-age children in the United States (6%) have at least one identified dis- ability. In 2014–2015, nearly 62% of children ages 6 to 21 with disabilities spent at least 80% of their school day in inclusive early childhood programs alongside their typically developing peers, compared to 33% of chil- dren with disabilities in 1990. A 2015 policy statement by the U.S. Department of Health and Human Services and the U.S. Department of Education (2015) underscored inclusion in preschool as a national priority in preschool expansion efforts.
Despite decades of effort to promote diver- sity and social inclusion for children with and without disabilities in preschool inclusive classrooms (Kwon, Elicker, & Kontos, 2011; Odom, Klingerman, & Jakowski, 2004), research continues to show mixed findings regarding whether children with disabilities can form frequent and positive play interac- tions with their typically developing peers in the classroom (Chen, Lin, Justice, & Sawyer, 2017; Diamond, 2001). A lack of understand- ing concerning whether and how interactions
828364ECXXXX10.1177/0014402919828364Exceptional ChildrenLin et al. research-article2019
1The Ohio State University 2Lehigh University
Corresponding Author: Tzu-Jung Lin, Department of Educational Studies, The Ohio State University, 145A Ramseyer Hall, 29 W Woodruff Ave., Columbus, OH 43210, USA. E-mail: [email protected]
Peer Interactions in Preschool Inclusive Classrooms: The Roles of Pragmatic Language and Self-Regulation
Tzu-Jung Lin1, Jing Chen1, Laura M. Justice1, and Brook Sawyer2
Abstract Drawing from a social network perspective, we examined the extent to which children with and without disabilities play with each other in preschool inclusive classrooms and identified malleable child characteristics that would support children forming these cross-status play interactions. A total of 200 children with disabilities and 301 children without disabilities participated in this study (M
age = 52.39 months, SD
age = 6.13). Results showed that children with and without
disabilities did not differ in the extent to which they formed cross-status play interactions after pragmatic language and self-regulation were taken into account. However, typically developing children were more likely to form same-status play interactions than children with disabilities. Children’s cross-status play interactions were predicted by self-regulation ability, which was fully mediated by their pragmatic language. The impact of pragmatic language on the formation of cross-status play interaction was greater for children with disabilities than their typically developing peers.
Lin et al. 433
between children with and without disabilities take place in preschool inclusive classrooms shakes the assumption that inclusive class- rooms benefit children with disabilities, as stated in the Salamanca Statement that “regu- lar schools with this inclusive orientation are the most effective means of combating dis- criminatory attitudes, creating welcoming communities, building an inclusive society and achieving education for all” (UNESCO, 1994, p. 8).
The primary goals of this study were to examine the extent to which children with dis- abilities and children without disabilities play with each other, hereafter called cross-status interaction, in preschool inclusive classrooms and to identify malleable child characteris- tics—characteristics that can be influenced by education—that would uniquely or jointly afford children to form more cross-status inter- actions in the classroom. In particular, we focused on children’s pragmatic language skills and self-regulation abilities as potential con- tributors to cross-status interactions, given the existing findings that these child characteristics are associated with the social functioning of children with and without disabilities in peer contexts (e.g., Eisenberg, Valiente, & Eggum, 2010; Staikova, Gomes, Tartter, McCabe, & Halperin, 2013).
Play and Peer Interactions Play offers many opportunities for peer interac- tions (Howes, Rubin, Ross, & French, 1988), and the qualities of play interactions are asso- ciated with academic engagement and moti- vation (Coolahan, Fantuzzo, Mendez, & McDermott, 2000). For children with and with- out disabilities, play interactions may occur with individuals with the same disability sta- tus or individuals who have a different status, and those interactions may be mediated by linguistic competence, especially in the area of pragmatics.
Researchers who have examined the promise of inclusive education have gener- ally reached a consensus that the benefits of inclusive education for children with and without disabilities cannot be fulfilled by
simply increasing the diversity of the class- room (e.g., Gottfried, 2016). Rather, benefits are contingent upon opportunities for chil- dren to form positive and productive social interactions within classrooms (e.g., Rafferty, Piscitelli, & Boettcher, 2003; Weiland, 2016). These findings encouraged us to examine pat- terns of social interactions between children with and without disabilities within inclusive classrooms.
In the current study, we examined the social interactions between preschool children with and without disabilities in the context of play. Play is a central activity in preschool class- rooms and affects all aspects of their learning and development (Barnett et al., 2008; Howes, 1988; Lillard, Hopkins, Dore, Palmquist, Lerner, & Smith, 2013). In play, especially interactive play in which children are interde- pendent and responsive to one another (Rubin, Bukowski, & Parker, 2006), children have opportunities to enact their social competence to pursue effective social exchanges or social construction of knowledge, such as perspec- tive taking, problem solving, social communi- cation, and self-regulation (Howes, 2011). More competent peers also play a role as scaf- folds that help a child progress to more com- plex levels of play through the development of different skills (Vygotsky, 1978).
Play Interactions Between Children With and Without Disabilities Forming play interactions with typically developing peers appears to benefit the social development of children with disabilities. Bauminger et al. (2008) conducted a labora- tory study to assess the qualitative differences of social interactions between three forms of friendship dyads among 8- to 12-year-old chil- dren: friendships between children with autism and their typically developing peers (cross- status dyads), friendships between children with autism and peers with a disability (same- status dyads), and friendships among typically developing children (typical dyads). Cross- status dyads showed more goal-oriented social behaviors, more positive social orientation and cohesion, and more complex coordinated play
434 Exceptional Children 85(4)
than same-status dyads. Moreover, the cross- status dyads did not differ from typical dyads on several interaction patterns they displayed, such as prosocial behaviors, conversations, social play, or time unoccupied. Children with autism in the cross-status dyads had signifi- cantly higher receptive language ability than those in the same-status dyad groups.
When children with disabilities engage in cross-status interactions, typically developing peers can motivate children with disabilities to demonstrate more complex and responsive social-emotional strategies. The literature of peer-mediated interventions provides reliable causal evidence to this account (e.g., Płatos & Wojaczek, 2018; Watkins et al., 2015). In peer-mediated interventions, typically devel- oping children are trained to model, tutor, and monitor positive social interactions for chil- dren with disabilities in natural social con- texts (Strain & Odom, 1986). Zhang and Wheeler (2011) conducted a meta-analysis of 45 single-subject-design studies and found that peer-mediated interventions are effective vehicles for promoting positive social interac- tions among children younger than 8 years diagnosed with autism. These results indicate that cross-status interactions can leverage the quality of social communication and coordi- nation for children with disabilities.
These findings also permit the inference that forming cross-status play interactions requires more linguistic, social, or cognitive demands from children both with and without disabilities (Bauminger et al., 2008). Children with disabilities have more limited capacities than their typically developing peers to enter into, coordinate, and sustain interactive play (Schuler & Fletcher, 2002). Children without disabilities often find it difficult to understand the emotions, motivation, and thoughts of children with disabilities (Yu, Ostrosky, & Fowler, 2015). The developmental differences in language, emotional regulation, and social understanding between children with and without disabilities often cause barriers for them to form successful and positive play interactions or relationships (Son et al., 2014; Walker & Berthelsen, 2008).
Yet, even with these barriers, children with disabilities have been found to form friend- ships with children with typically developing peers. Buysse (1993) and Guralnick (1997) asked parents and teachers to report children’s friendship patterns in community-based set- tings. Most preschool children, including chil- dren with disabilities, had at least one playmate or a mutual friend from their community. The majority of these playmates were typically developing children, which was consistent with the availability of typically developing peers in their social groups. Guralnick, Gott- man, and Hammond’s (1996) laboratory study provided more causal interpretations for chil- dren’s selection of playmates based on disabil- ity status. Children initially unacquainted with one another were formed into play groups similar in disability status (communication disorders or developmental delays) or contain- ing primarily typically developing children with a small number of children with disabili- ties. Results of a 2-week observation of chil- dren’s friendship formation showed that children with disabilities were apt to play with children either with or without disabilities, whereas typically developing children were more likely to interact or form friendships with typically developing peers.
One major methodological limitation of these studies is the reliance on linear and child-focused approaches, which often lim- ited researchers’ observations within dyads, ignoring the more complex nature of intercon- nectedness among children in a social system. Another limitation is that many of the peer interactions between children with and with- out disabilities were observed in laboratory settings. Less is known about the frequency of naturally occurring cross-status versus same- status interactions in regular social settings, such as preschool classrooms. Farmer and Farmer (1996) argued that to understand social inclusion issues, research needs to look beyond individuals’ characteristics to examine the classroom social processes within which chil- dren with disabilities are positioned. They conducted a descriptive social network analy- sis in three elementary inclusive classrooms.
Lin et al. 435
Over 90% of students with and without dis- abilities in the classrooms belonged to one of the peer groups naturally formed in these classrooms based on students’ personal or social characteristics. Chen et al. (2017) examined social networks to unpack the com- plex social process of peer interactions among a large number of preschool inclusive class- rooms. They found that children with disabili- ties were less likely to interact with their classmates than were their typically develop- ing peers. Furthermore, children tended to play with peers who shared the same disabil- ity status than with peers of a different disabil- ity status. However, the study did not explore whether this disability homophily effect (Mcpherson, Smith-lovin, & Cook, 2001), which counters Buysse’s (1993) and Gural- nick et al.’s (1996) findings, applies to chil- dren with different disability status.
Research needs to look beyond individuals’ characteristics to examine the classroom social
processes within which children with disabilities are positioned.
To this end, more research is needed to deepen our understanding about cross-status and same-status interactions between children with and without disabilities. Through the lens of a social network perspective, we explored the extent to which cross-status interactions occur in the context of preschool inclusive classrooms. We expected to see a significant disability homophily effect but hypothesized that typically developing children might be more likely than children with disabilities to play with same-sta- tus peers rather than cross-status peers due to the linguistic, social, and cognitive demands of playing with their different-status peers (Baum- inger et al., 2008; Guralnick et al., 1996).
The Role of Pragmatic Language and Self-Regulation in Peer Interaction Pragmatic language skills represent one’s use of language for social purposes (Prutting &
Kittchner, 1987; Roth & Spekman, 1984a, 1984b) to deliver communicative, social, and emotional functions during interactions, such as using social contextual cues to understand others’ intentions. In the current study, prag- matic language was operationalized as chil- dren’s functional communication skills at school (e.g., can appropriately respond to a familiar person’s smiles, frowns, or looks of surprise). Children rely on pragmatic lan- guage skills to initiate reciprocal and positive interactions, express needs, understand oth- ers’ perspectives, provide social support, or resolve conflicts (Asher, Parker, & Walker, 1996). Children who are able to initiate effec- tive communication with others are more likely to display prosocial behavior and main- tain their relationships with others (Dionne, Tremblay, Boivin, Laplante, & Perusse, 2003). In contrast, poor pragmatic language is associated with problem behaviors, social impairment, and impoverished relationships with others, and is more commonly seen in children with disabilities (Staikova et al., 2013). The development of pragmatic lan- guage skills becomes more rapid and sophisti- cated during the preschool years (O’Neill, 2007).
The preschool years also represent a developmental stage in which children’s self- regulation skills develop. Self-regulation refers to the ability to control one’s emo- tional, behavioral, and cognitive responses to environmental demands (Smith-Donald, Raver, Hayes, & Richardson, 2007). Research shows that children with lower ability to regulate their own emotions, behav- ior, and cognitive functioning often engage in lower-level participation in class (Valiente, Lemery-Chalfant, Swanson, & Reiser, 2008). They may have fewer opportunities to inter- act with peers (McClelland & Morrison, 2003) and suffer from poor quality of peer and teacher interactions (Eisenberg et al., 2010; Guralnick, 2006).
Language and self-regulation should be investigated for their mutual influences (e.g., Bohlmann, Maier, & Palacios, 2015). Accord- ing to Vygotsky’s (1978) theory, language is the primary means by which children regulate
436 Exceptional Children 85(4)
thoughts and their own behavior. This is evi- dent in young children’s active use of self- regulatory private speech—the think-aloud strategy—to regulate their own behavior (Winsler, Diaz, & Montero, 1997). The use of private speech interacts significantly with emotional regulation to predict children’s sadness and anger during a frustrating task (Day & Smith, 2013). Other studies found that language is a supporting factor of self- regulation development (e.g., Vallotton & Ayoub, 2011).
Bohlmann et al. (2015) systematically identified bidirectional associations between expressive vocabulary and self-regulation in both typically developing monolingual Eng- lish-speaking and bilingual preschoolers. They argued that both language and self-regu- lation grow in a coactive and iterative process as the rate of growth rapidly increases during the preschool period.
Although cumulative evidence supports the intricate relationship between language and self-regulation, most of these studies were conducted with typically developing children (Bohlmann et al., 2015). We hypoth- esize that pragmatic language skills and self- regulation may exert both unique and joint influences on the nature of play interactions between children with and without disabili- ties in preschool inclusive classrooms. We speculate that children who can communi- cate with others in socially acceptable and pragmatically effective ways are better at using language to regulate their attention and emotions while interacting with others. Alternatively, children with better self-regu- lation strategies may be better able to attend to others’ needs and feelings, which orient them toward more effective ways to commu- nicate with others. Both of these abilities— pragmatic language and self-regulation—may jointly and positively predict the extent to which children initiate cross-status interac- tions, considering that interacting with peers with whom they share a different disability status may involve higher levels of positive social orientation and more complex levels of cognitive and behavioral coordination (Bauminger et al., 2008).
The Current Study Social integration, namely, the meaningful par- ticipation between children with and without disabilities in classroom settings, is one of the key goals of early childhood inclusion (Gural- nick & Bruder, 2016). Understanding how likely cross-status interactions can evolve from preschool inclusive classrooms and what mal- leable individual factors contribute to cross-sta- tus play interactions may inform better educational practices that create preschool classrooms that are more socially integrated. We examined two specific research questions: (a) To what extent do children form cross-sta- tus play interactions and same-status play inter- actions in preschool inclusive classrooms? (b) To what extent do children’s pragmatic lan- guage skills and self-regulation ability uniquely or jointly predict their tendency to form cross- status and same-status play interactions, con- trolling for other child and classroom factors?
To address these aims, we used data col- lected during a large-scale, randomized con- trolled trial project that aimed to promote early language and literacy skills for children with disabilities. Classrooms were randomly assigned to a business-as-usual condition or a condition in which teachers implemented two whole-class lessons per week for 30 weeks. These lessons were designed to teach children skills in four areas: narrative (e.g., characters in stories), vocabulary (e.g., rare vocabulary words), phonological awareness (e.g., rhym- ing patterns), and print knowledge (e.g., names of alphabet letters). The lessons did not target children’s development of pragmatic skills or social competence; thus for the present pur- poses, we assessed our primary research aims using children in both conditions.
Method
Setting We included a total of 66 preschool inclusive classrooms from Head Start, state-funded, and private (tuition-based) programs in this study. These classrooms met the following criteria for a social network analysis: (a) inclusion
Lin et al. 437
(classrooms had to have at least two children with disabilities and two typically developing children) and (b) gender diversity (the class- rooms contained both boys and girls. All of the classrooms were full-day programs adopting a full-inclusion model (i.e., children with disabili- ties participated in all aspects of the curriculum, rather than joining part of the curriculum with typically developing children during the day). Consent and assent of all participating children were obtained based on approved institutional review board procedures.
Among the 66 lead teachers, 63% were female and 96% were Caucasian. Nineteen percent attained an associate’s degree as their highest level of completed education, and the remaining teachers held a bachelor’s degree or higher. Teachers had one or multiple majors in early childhood education (52%), special edu- cation (43%), elementary education (13%), child development (5%), English as a second language (5%), or other (14%).
Participants A total of 501 children participated in this study (288 boys, 213 girls). Children were sampled from classrooms that met several cri- teria, including caregiver consent and English proficiency (i.e., English is child’s primary language or child understands and speaks English well). Additionally, children needed to have sufficient verbal (i.e., speak in at least two-word utterances), cognitive, and sensory functioning to participate in the child-level assessments in the study; as such, children with disabilities sampled would be considered moderate to high functioning. Children’s age varied from 33 to 68 months (M = 52.39, SD = 6.13). According to parent report, the majority of these children were Caucasian (75%, n = 367), and the other children were identified as African American (16%), Asian (1%), Native Hawaiian (0.2%), or other race- ethnicity (5%). In terms of maternal educa- tion, specifically the highest earned degree, 7% of mothers did not have a high school diploma, 27% had a high school diploma, 37% had some college or a 2-year degree, 16% had a bachelor’s degree, 12% had a mas-
ter’s degree, and 1% had a doctoral degree. The median overall annual household income level was $40,001 to 60,000, and the mode income level was $20,000 or less.
Among the 501 children, 40% (n = 200) were children with disabilities as operational- ized by having an individualized educational plan (IEP). Parents were asked to report chil- dren’s primary disability type, if known. Thirty-six percent (n = 72) of the children with IEPs had a speech-language impairment, 17% (n = 34) had a developmental delay, 15.5% (n = 31) had multiple disabilities, 5.5% (n = 11) had an autism spectrum disor- der, and 5% (n = 10) had either emotional disturbance, orthopedic impairment, specific learning disability, visual impairment, hearing impairment, or other health impairment. For the remaining 21% of children with disabili- ties (n = 42), no specific diagnosis was reported by parents.
As presented in Table 1, class size ranged from 4 to 11 children (M = 7.59, SD = 2.04). The proportion of children with disabilities in each classroom ranged from 0.10 to 0.80 (M = 0.42, SD = 0.16). The proportion of girls in each classroom varied from 14% to 80% (M = 0.43, SD = 0.15).
Measures Pragmatic language skills. We assessed chil- dren’s pragmatic language skills in the win- ter and spring of the year using the Descriptive Pragmatics Profile (DPP; Wiig, Secord, & Semel, 2004). The spring scores were highly correlated with (r = .85) and greater than the fall scores (p < .001), sug- gesting that preschool children’s pragmatic language skills developed from winter to spring in a steady rate. For the purpose of this study, we chose the spring measure and not the fall measure as the predictor of chil- dren’s play networks because it better reflected children’s skill level at the time when their play networks were assessed. The DPP included 26 items in a 4-point Likert format (1 = never to 4 = always). It was completed by each teacher for each child in the classroom. Test authors reported high
438 Exceptional Children 85(4)
internal consistency (alpha = .95) and test– retest reliability (.87), as well as content validity, convergent validity, and diagnostic accuracy (Reilly et al., 2010; Wiig et al., 2004). In the current study, concurrent valid- ity of the DPP measure was confirmed by its significant correlation with children’s expressive and receptive language skills measured by Clinical Evaluation of Lan- guage Fundamentals Preschool 2 (r = .57, p <.001; Wiig et al., 2004). The alpha was .97. We used average scores to represent a child’s pragmatic language skill in the analysis.
Self-regulation skills. The Preschool Self-Regu- lation Assessment (Smith-Donald et al., 2007) was used to examine two forms of children’s self-regulation: attention-impulse control (e.g., “The child has difficulty waiting between tasks) and positive emotion (e.g., “Child shows intense positive emotions or behaviors”). Trained assessors scored each child’s behavior from 0 to 3 based on 25 items during test-tak- ing sessions in the fall. Testing was conducted one-on-one in a private area in the child’s pre- school. Raters received a two-step training process prior to assessing. First, raters
Table 1. Descriptive Statistics of Child and Classroom Characteristics in Preschool Inclusive Classrooms.
Variable Missing n % M SD Range
Children characteristics Gender 0 Boy 288 57 Girl 213 43 Race 11 Caucasian 367 75 Other 123 25 Disability status 0 Without IEP 301 60 With IEP 200 40 Maternal education 29 High school or less, no diploma 33 7 High school 129 27 Some college or a 2-year degree 173 37 Bachelor’s degree 74 16 Master’s degree 56 12 Doctoral degree 7 1 Annual household income 136 $20,000 or less 92 25 $20,001–$40,000 83 23 $40,001–$60,000 53 15 $60,001–$100,000 67 18 $100,000–$150,000 44 12 More than $150,000 26 7 Age in months 0 501 52.39 6.13 33.00–68.00 Pragmatic language 9 492 3.36 0.59 1.50–4.00 Attention-impulse control 6 495 2.45 0.64 0.07–3.00 Positive emotion 9 492 2.12 0.75 0.00–3.00 Classroom characteristics 0 Classroom size 66 7.59 2.04 4.00–11.00 Proportion of children with disabilities 66 0.42 0.16 0.10–0.80 Proportion of girls 66 0.43 0.15 0.14–0.80
Note. IEP = individualized education program.
Lin et al. 439
completed an online training that provided information on (a) the structure and materials of the measure and (b) how to administer the measure reliably. Raters had to score 100% on a quiz on the materials. Second, raters scored a minimum of 95% on the fidelity-of-assess- ment checklist during a live assessment ses- sion. The constructs were internally consistent (attention-impulse control, α = .89; positive emotion, α = .87). Interrater reliability was high (intraclass correlation = .82 for attention- impulse control and .83 for positive emotion). The construct and concurrent validity of this measure have been reported by the test authors (Smith-Donald et al., 2007). In the analysis, we averaged children’s attention-impulse con- trol and positive emotion across the corre- sponding items.
Classroom play networks. Toward the end of the school year, teachers rated the frequency with which each pair of children in the class- room reciprocally played with each other dur- ing a typical school day on a 5-point rating scale (0 = never to 4 = always). Teachers were asked to rate children’ play interactions based on their daily observations of low-infer- ence behaviors, such as engaging in pretend play, giving and sharing toys, exploring objects together, and building blocks together. This measure was first developed and imple- mented in a randomized controlled trial study (see Lin, Justice, Paul, & Mashburn, 2016, for more information). To test its construct valid- ity, we calculated the total number of peer interactions received by each individual child from all of the classmates participating in the current study. This measure was positively correlated with children’s learning-related behaviors measured by the Preschool Learn- ing Behaviors Scale (McDermott, Leigh, & Perry, 2002) for children both with (p < .01) and without disabilities (p < .05). These cor- relations support previous studies that chil- dren with better learning-related behaviors engaged in more interactive play (e.g., Bulotsky-Shearer, Bell, Romero, & Carter, 2012). The consistent findings suggest that our teachers were good informants of chil- dren’s social interaction in the classrooms.
Analytic Approach In the current study, we applied an inferential social network analysis called exponential random graph models (ERGMs; Pattison & Wasserman, 1999) to examine the dynamics of classroom social networks. The unit of analysis was the play interaction between a pair of children (e.g., the frequency of play between Mary and Lily), and the interdepen- dence among these play interactions was taken into account. The software we used was the statnet package in R (Handcock et al., 2016). Two types of network effects were tested simultaneously in ERGMs: actor covariate effect and structural effect. In the current study, actor covariate effects refers to child or dyadic characteristics that potentially influence the network formation process. There were two forms of actor covariate effects in ERGMs, main actor effects and homophily effects. Main actor effects explained whether a child with a particular characteristic was more or less likely to form play interactions with peers. Homophily effects measured how likely play interactions were formed between dyads with similar characteristics.
We used structural effects as a statistical control for the interdependency among play interactions by the endogenous characteristics of a network. Network structural effects refers to the extent to which existing patterns of local network configurations influence subse- quent network formation. In this study, we considered two types of structural effects, edge and transitivity. The edge effect referred to the overall probability of the network for- mation in a logit format when all the other variables were controlled, which was similar to the intercept effect in regular linear models. A negative edge effect has been commonly observed in ERGM studies, especially when the network size is large (e.g., Mcfarland, Moody, Diehl, Smith, & Thomas, 2014). The transitivity effect (e.g., Robins, Pattison, & Wang, 2009) examined the extent to which children expanded their play interactions. The specific type of transitivity effect used in this study was the effect of triangle closing (also
440 Exceptional Children 85(4)
called geometrically weighted edgewise shared partnerships, or GWESP; Hunter, 2007), which modeled the probability for child A to interact with child C, given child A interacted with child B, who interacted with child C. In this case, the dyadic interactions among children A, B, and C were interdepen- dent with each other and could not be treated as independent observations as assumed in regular linear models.
In the current study, we conducted three data transformation steps to prepare the net- work data for ERGMs. First, all the classroom social networks were valued networks in which play interactions between pairs of chil- dren rated by the teachers ranged from 0 (never play) to 4 (always play). Due to ERGM’s requirement, these valued networks were then transformed into binary networks where the values of play interactions were either 1 or 0 based on the cutoff score of 2 (sometimes play). A play interaction between a pair of chil- dren rated at the value of 1 in the binary net- work indicated that the dyad played with each other at or above the average level. The cutoff score was selected based on the grand mean value of all play interactions across classrooms (M = 1.97). Using this cutoff score would make the results more interpretable in the framework of a pooled social network that is illustrated subsequently. Second, 66 small binary classroom social networks were pooled into one large social network, yielding reliable estimates of play interactions across all class- rooms. Between-classroom interactions were not plausible to occur because play interac- tions were assessed by teachers within their own classrooms. Hence, these implausible play interactions were specified as structural zeros in the ERGMs. Third, this large pooled network was recoded into a cross-status net- work and a same-status network to examine the mechanisms underlying the formation of cross-status and same-status interactions. In the cross-status play network, interactions formed between children with and without dis- ability status at or above the average level were coded as 1; otherwise, as 0. In the same- status network, interactions formed between children with the same disability status at or
above the average level were coded as 1; oth- erwise, as 0. ERGMs were specified for the cross-status and the same-status play net- works, respectively.
The data-pooling approach was built upon the assumption that network formation pro- cesses were homogeneous across classrooms. However, this approach might fail to account for potential between-classroom variations due to the fact that children were nested within classrooms. To account for this limitation, we included classroom characteristics as control variables, including classroom size, classroom proportion of children with disabilities, and proportion of girls in each classroom. To exam- ine the model fits of ERGMs, the Markov chain Monte Carlo (MCMC) diagnoses and the goodness-of-fit (GOF) functions were applied (Hunter, Goodreau, & Handcock, 2008).
To address the first research question, we calculated the frequencies of cross- and same- status interactions, accompanied with ERGMs showing the effects of disability status on the formation of cross- and same-status interac- tions, controlling for classroom characteristics and children’s gender and age (e.g., Hestenes & Carroll, 2000; Martin et al., 2013). To address the second question, pragmatic lan- guage and self-regulation were entered into the ERGMs to examine the unique and joint effects of these two constructs. If the main effects of pragmatic language and self-regula- tion were significant, we further examined whether these effects differed by children’s disability status.
Results In this section, we report three sets of analy- ses. First, we describe the proportions of cross-status and same-status interactions in the preschool classrooms. Second, we pres- ent two ERGMs showing the main effects of disability status on children’s cross-status and same-status interactions. Third, we report additional ERGMs showing the main and mediating effects of pragmatic lan- guage and self-regulation on the formation of cross-status and same-status play interac- tions.
441
T ab
le 2
. D
es cr
ip ti ve
S ta
ti st
ic s
o f C
hi ld
C ha
ra ct
er is
ti cs
b y
D is
ab ili
ty S
ta tu
s.
C hi
ld re
n w
it h
di sa
bi lit
ie s
(n
= 2
00 )
C hi
ld re
n w
it ho
ut
di sa
bi lit
ie s
(n =
3 01
) C
o rr
el at
io ns
V ar
ia bl
e %
M SD
% M
SD t
C o he
n’ s
d A
ge PL
A C
PE
G en
de r
M
al e
71 48
.5
Fe
m al
e 29
51 .5
R
ac e
C au
ca si
an 77
73
O
th er
s 23
26
A ge
in m
o nt
hs 52
.2 0
7. 06
52 .5
1 5.
44 −
0. 55
0. 05
—
Pr ag
m at
ic la
ng ua
ge (
PL )
3. 01
0. 61
3. 59
0. 45
− 12
.0 7*
** 1.
08 .1
4* *
—
A tt
en ti o n-
im pu
ls e
co nt
ro l (
A C
) 2.
18 0.
75 2.
64 0.
47 −
8. 28
** *
0. 74
.1 9*
* .3
6* *
—
Po si
ti ve
e m
o ti o n
(P E)
1. 88
0. 80
2. 28
0. 66
− 5.
91 **
* 0.
55 .1
7* *
.2 6*
* .3
4* *
—
*p <
.0 5.
* *p
< .0
1. *
** p <
.0 01
.
442 Exceptional Children 85(4)
Preliminary Analysis As shown in Table 2, the correlations among children’s pragmatic language skills and two types of self-regulation skills (attention- impulse control, positive emotion) were all significant at the alpha level of .01. Table 2 also presents the descriptive statistics of child characteristics by disability status. The major- ity of children with disabilities were male (71%), whereas the gender ratio was relatively even for typically developing children. Most of the children with and without disabilities were Caucasian. Independent t tests indicated that children with and without disabilities did not differ significantly by age (t = −0.55). However, typically developing children had higher pragmatic language skills and self-reg- ulation ability than children with disabilities (ps < .001, Cohen’s ds = 0.55–1.08; see Table 2).
Cross-Status and Same-Status Play Interactions Our first research question was to understand the extent to which children with and without disabilities engaged in cross-status and same- status play interactions. The descriptive statis- tics based on the large binary network is presented in Table 3. Across 66 classrooms, we observed 1,153 interactions, among which 44.93% occurred between children with dif- ferent disability status, 12.06% occurred between children with disabilities, and 43.02% were between typically developing
children. We noted that the number of same- status or cross-status interactions a child could have depended on the number of peers with or without disabilities in the classrooms. Hence, we calculated the proportions of observed interactions for each type of interaction given the maximum possible interactions. Maximum possible interactions was the number of dyadic play interactions that children can pos- sibly form in the classroom. In a classroom where there were m children with disabilities and n typically developing children, the maximum possible interactions for the play interactions, regardless of cross-status or same-status interactions, would be (m + n)(m + n – 1)/2. The maximum possible interac- tions for the cross-status interactions would be m × n, and the maximum possible interac- tions for the same-status interactions would be [m (m – 1) + n (n – 1)]/2. Of the 1,786 maximum possible interactions from 66 class- rooms, we observed 64.56% interactions. Of the 875 possible cross-status interactions, we observed 59.20% cross-status interactions. Of the 245 possible same-status interactions for children with disabilities, we observed 56.73% interactions, which was smaller than the proportion of same-status interactions for typically developing children (74.47%).
Two sets of ERGMs were conducted to examine the effect of disability status on the probability for children to form cross-status or same-status interactions with peers. In the cross-status model (Table 4, Model 1), the effect of disability status was significant and
Table 3. Descriptive Statistics of Classroom Social Networks by Disability Status.
Variable Total
interactions
Interactions between children
with different disability status
Interactions between
children with disabilities
Interactions between
children without disabilities
Frequency of observed interactions 1,153 518 139 496 Observed interactions divided by
total interactions — 44.93% 12.06% 43.02%
Maximum possible interactions 1,786 875 245 666 Observed interactions divided by
maximum possible interactions 64.56% 59.20% 56.73% 74.47%
Note. Frequencies of interactions were calculated from the binary networks, where the interactions between pairs of children were coded as 1 if the frequency of interactions ≥ 2 (sometimes play) and otherwise coded as 0.
443
T ab
le 4
. ER
G M
s Pr
ed ic
ti ng
t he
F o rm
at io
n o f C
ro ss
-S ta
tu s
Pl ay
In te
ra ct
io ns
.
V ar
ia bl
e M
o de
l 1 M
o de
l 2 M
o de
l 3 M
o de
l 4 M
o de
l 5
C hi
ld a
tt ri
bu te
s
D is
ab ili
ty s
ta tu
s (1
= w
it h
di sa
bi lit
ie s,
0 =
t yp
ic al
ly d
ev el
o pi
ng )
− 0.
22 (
0. 09
)* −
0. 06
( 0.
09 )
− 0.
12 (
0. 09
) −
0. 02
( 0.
10 )
− 0.
38 (
0. 15
)*
G
en de
r (1
= g
ir l,
0 =
b o y)
− 0.
08 (
0. 09
) −
0. 25
( 0.
09 )*
* −
0. 11
( 0.
09 )
− 0.
26 (
0. 09
)* *
− 0.
28 (
0. 10
)* *
A
ge (
in m
o nt
hs )
0. 02
( 0.
01 )*
** 0.
00 (
0. 01
) 0.
02 (
0. 01
)* *
0. 00
( 0.
01 )
0. 00
( 0.
01 )
Pr
ag m
at ic
la ng
ua ge
0. 52
( 0.
06 )*
** 0.
51 (
0. 06
)* **
0. 49
( 0.
06 )*
**
A tt
en ti o n-
im pu
ls e
C o nt
ro l
0. 12
( 0.
06 )*
0. 04
( 0.
06 )
0. 02
( 0.
06 )
Po
si ti ve
e m
o ti o n
0. 15
( 0.
06 )*
0. 08
( 0.
06 )
0. 09
( 0.
06 )
D
is ab
ili ty
S ta
tu s ×
P ra
gm at
ic L
an gu
ag e
0. 15
( 0.
05 )*
* D
ya d
at tr
ib ut
es
G en
de r
ho m
o ph
ily e
ffe ct
0. 60
( 0.
10 )*
** 0.
61 (
0. 11
)* **
0. 61
( 0.
10 )*
** 0.
62 (
0. 11
)* **
0. 61
( 0.
11 )*
** C
la ss
ro o m
a tt
ri bu
te s
Pr
o po
rt io
n o f c
hi ld
re n
w it h
di sa
bi lit
ie s
− 0.
23 (
0. 19
) −
0. 13
( 0.
21 )
− 0.
27 (
0. 20
) −
0. 14
( 0.
21 )
− 0.
14 (
0. 21
)
Pr o po
rt io
n o f g
ir ls
0. 16
( 0.
20 )
0. 39
( 0.
20 )
0. 17
( 0.
20 )
0. 40
( 0.
21 )
0. 46
( 0.
21 )*
C
la ss
s iz
e 0.
03 (
0. 02
) 0.
02 (
0. 02
) 0.
03 (
0. 02
) 0.
02 (
0. 02
) 0.
02 (
0. 02
) St
ru ct
ur al
e ffe
ct s
Ed
ge s
− 1.
81 (
0. 65
)* *
− 3.
47 (
0. 71
)* **
− 2.
69 (
0. 70
)* **
− 3.
80 (
0. 75
)* **
− 3.
61 (
0. 76
)* **
T
ri an
gl e
cl o si
ng (
G W
ES P)
− 0.
36 (
0. 05
)* **
− 0.
40 (
0. 05
)* **
− 0.
36 (
0. 05
)* **
− 0.
40 (
0. 05
)* **
− 0.
41 (
0. 05
)* **
N ot
e. S
ta nd
ar d
er ro
rs s
ho w
n in
p ar
en th
es es
. E R
G M
= e
xp o ne
nt ia
l r an
do m
g ra
ph m
o de
l; G
W ES
P =
g eo
m et
ri ca
lly w
ei gh
te d
ed ge
w is
e sh
ar ed
p ar
tn er
sh ip
. *p
< .0
5. *
*p <
.0 1.
* **
p <
.0 01
.
444 Exceptional Children 85(4)
negative (β = −0.22, SE = 0.09, p < .05), suggesting that the probability for children with disabilities to form cross-status interac- tions was 80.25% (= exp[−0.22]) of that for typically developing children. As all the class- rooms in our sample contained at least two children with disabilities and two children without disabilities, cross-status interactions were plausible in all of the classrooms. The negative effect of disability status thus sug- gested that children with disabilities were less likely to form cross-status interactions than typically developing children. In the same- status model (Table 5, Model 1), the main effect of disability status showed a significant negative effect (β = –1.02, SE = 0.09, p < .001), which suggested that the likelihood of forming same-status play interactions for chil- dren with disabilities was only 36.06% (= exp[–1.02]) of that for children without dis- abilities.
The Effect of Language Skills and Self-Regulation Skills on Peer Interactions To address the second research question, we examined the effect of pragmatic language and self-regulation on the formation of cross- status or same-status play interactions. Prag- matic language and self-regulation were entered into the ERGMs one at a time. When the ERGMs included only pragmatic language skills (Model 2 in Tables 4 and 5), pragmatic language was a significant positive predictor of both cross-status and same-status play interactions (β = 0.52, SE = 0.06, p < .001; β = 0.12, SE = 0.05, p < .05, respectively). When the models contained self-regulation skills only (Model 3 in Tables 4 and 5), atten- tion-impulse control and positive emotion positively predicted cross-status play interac- tions (β = 0.12, SE = 0.06, p < .05; β = 0.15, SE = 0.06, p < .05). Neither of these two types of self-regulation was a significant predictor of same-status play interactions. When both pragmatic language and self-regu- lation were included in the model (Model 4 in Tables 4 and 5), the significant effects of prag- matic language remained, whereas the effects
of self-regulation disappeared. These findings indicated that the effects of self-regulation on peer interactions were fully mediated by prag- matic language.
We then examined the interplay between pragmatic language and disability status given the robust effect of pragmatic language on cross- and same-status interactions. As shown in Table 4 (Model 5), the interaction between pragmatic language and disability status was positively significant (β = 0.15, SE = 0.05, p < .01), and the results of the main effect of pragmatic language remained significant. This finding suggested that pragmatic lan- guage had a stronger positive effect on the probability for children with disabilities to form cross-status interactions than it did for typically developing children. This interac- tion effect was not significant in the same- status model (Table 5, Model 5).
When all the other variables were taken into account, the main effect of gender was significant in the cross-status interaction model (β = −0.28, SE = 0.10, p < .01). Com- pared with boys, girls were less likely to play with peers who had a different disability sta- tus. The gender homophily effect was evident in both cross-status and same-status play interactions (β = 0.61, SE = 0.11, p < .001; β = 0.62, SE = 0.11, p < .001, respectively), indicating that children were more likely to play with peers who had the same gender. Age had a positive effect on cross-status interac- tions when pragmatic language was not included in the model (Table 4, Models 1 and 3) but became nonsignificant once we accounted for pragmatic language.
Classroom size was a negative predictor of same-status interactions (Table 5, Model 5; β = −0.05, SE = 0.01, p < .001), suggest- ing that the tendency for children to form same-status play interactions decreased with class size. The proportion of children with disabilities in a classroom did not signifi- cantly predict cross- or same-status interac- tions. The effect of gender ratio was significant in the cross-status model only after the interaction between disability status and pragmatic language was considered (Table 4, Model 5).
445
T ab
le 5
. ER
G M
s Pr
ed ic
ti ng
t he
F o rm
at io
n o f S
am e-
St at
us P
la y
In te
ra ct
io ns
.
V ar
ia bl
e M
o de
l 1 M
o de
l 2 M
o de
l 3 M
o de
l 4 M
o de
l 5
C hi
ld a
tt ri
bu te
s
D is
ab ili
ty s
ta tu
s (1
= w
it h
di sa
bi lit
ie s,
0 =
t yp
ic al
ly d
ev el
o pi
ng )
− 1.
02 (
0. 09
)* **
− 0.
99 (
0. 09
)* **
− 0.
99 (
0. 09
)* **
− 0.
98 (
0. 09
)* **
− 0.
93 (
0. 14
)* **
G
en de
r (1
= g
ir l,
0 =
b o y)
0. 07
( 0.
08 )
0. 03
( 0.
07 )
0. 06
( 0.
07 )
0. 03
( 0.
08 )
0. 03
( 0.
08 )
A
ge (
in m
o nt
hs )
0. 00
( 0.
00 )
− 0.
00 (
0. 00
) 0.
00 (
0. 00
) −
0. 00
( 0.
00 )
− 0.
00 (
0. 00
)
Pr ag
m at
ic la
ng ua
ge 0.
12 (
0. 05
)* 0.
10 (
0. 04
)* 0.
11 (
0. 04
)*
A tt
en ti o n-
im pu
ls e
C o nt
ro l
0. 03
( 0.
05 )
0. 02
( 0.
05 )
0. 02
( 0.
05 )
Po
si ti ve
e m
o ti o n
0. 06
( 0.
05 )
0. 04
( 0.
05 )
0. 04
( 0.
05 )
D
is ab
ili ty
S ta
tu s ×
P ra
gm at
ic L
an gu
ag e
− 0.
02 (
0. 04
) D
ya d
at tr
ib ut
es
G en
de r
ho m
o ph
ily e
ffe ct
0. 62
( 0.
10 )*
** 0.
61 (
0. 11
)* **
0. 62
( 0.
11 )*
** 0.
61 (
0. 11
)* **
0. 62
( 0.
11 )*
** C
la ss
ro o m
a tt
ri bu
te s
Pr
o po
rt io
n o f c
hi ld
re n
w it h
di sa
bi lit
ie s
0. 25
( 0.
17 )
0. 27
( 0.
17 )
0. 25
( 0.
17 )
0. 28
( 0.
18 )
0. 27
( 0.
17 )
Pr
o po
rt io
n o f g
ir ls
0. 08
( 0.
16 )
0. 13
( 0.
16 )
0. 10
( 0.
17 )
0. 12
( 0.
17 )
0. 12
( 0.
17 )
C
la ss
s iz
e −
0. 05
( 0.
01 )*
** −
0. 05
( 0.
01 )*
** −
0. 04
( 0.
01 )*
** −
0. 05
( 0.
01 )*
** −
0. 05
( 0.
01 )*
** St
ru ct
ur al
e ffe
ct s
Ed
ge s
− 1.
94 (
0. 56
)* **
− 2.
28 (
0. 60
)* **
− 2.
21 (
0. 62
)* **
− 2.
42 (
0. 60
)* **
− 2.
42 (
0. 61
)* **
T
ri an
gl e
cl o si
ng (
G W
ES P)
1. 37
( 0.
12 )*
** 1.
35 (
0. 12
)* **
1. 36
( 0.
12 )*
** 1.
35 (
0. 12
)* **
1. 36
( 0.
13 )*
**
N ot
e. S
ta nd
ar d
er ro
rs s
ho w
n in
p ar
en th
es es
. E R
G M
= e
xp o ne
nt ia
l r an
do m
g ra
ph m
o de
l; G
W ES
P =
g eo
m et
ri ca
lly w
ei gh
te d
ed ge
w is
e sh
ar ed
p ar
tn er
sh ip
. *p
< .0
5. *
*p <
.0 1.
* **
p <
.0 01
.
446 Exceptional Children 85(4)
Regarding the structural effects in the ERGMs (Tables 4 and 5), the edges effect was negative across all the models, suggesting that children did not form play interactions by chance. The positive triangle-closing (GWESP) effect in the same-status model suggested that same-status interactions tended to spread from child to child transitively. In contrast, the negative GWESP effect in the cross-status model suggested that cross-status interactions were less likely to be formed tran- sitively.
The MCMC graphical statistics suggested that the models were free of degeneracy and the results were trustworthy. The GOF func- tions compared the distributions of the observed and the simulated networks. All the p values in the GOF examinations were close to 1 (range = 0.90–1.00), which suggested that the models well represented the observed network in terms of the statistics that had been specified. Thus, the model fits were acceptable.
Discussion We examined cross-status play interactions in preschool inclusive classrooms and factors associated with those interactions through the lens of social network analysis as a means of studying the effects of inclusive education. Overall, our findings showed that although cross-status play interactions occurred less frequently than same-status interactions, cross-status interactions did occur at a promi- nent rate in preschool inclusive classrooms. In addition, this study went beyond previous findings about the disability homophily phe- nomenon (Chen et al., 2017) to demonstrate that the level at which children engaged in cross-status or same-status play interactions differed by children’s disability status, prag- matic language, and self-regulation skills.
One major finding of the study was that typically developing children showed a greater likelihood than children with disabilities to form same-status interactions (Table 5, Model 1), after controlling for base rate differences. Interestingly, the disability effect on cross-sta- tus interactions disappeared after accounting for pragmatic language and self-regulation (Table 4). This finding suggests that the vari-
ance in children’s cross-status interactions was more likely to attribute to pragmatic language and self-regulation skills rather than disability status itself, and that children with and without disabilities did not differ in their likelihood to form cross-status interactions. These findings were aligned with findings by Buysse (1993) and Guralnick et al. (1996) that children with disabilities did not favor cross-status or same- status interactions, whereas typically develop- ing children showed a strong preference for same-status interactions. One possible reason might be that for typically developing chil- dren, forming same-status interactions is less demanding than forming cross-status interac- tions (Bauminger et al., 2008). Even though the linguistic, cognitive, and social demands for children with disabilities to form cross- status interactions remain greater than same- status interactions, the difference might not be as substantial compared with their typically developing peers. If typically developing peers are capable of offering quality support, chil- dren with disabilities might even favor cross- status interactions over same-status interactions, although this is beyond the scope of the current study.
Children with disabilities did not favor cross-status or same-status interactions, whereas typically developing children showed a
strong preference for same-status interactions.
Another key finding of this study was about the roles of pragmatic language and self-regu- lation in the formation of classroom play net- works. Both pragmatic language and self-regulation positively predicted the likeli- hood of cross-status play interactions, but only pragmatic language predicted the likelihood of same-status play interactions. Furthermore, pragmatic language seemed to play a more important role for children with disabilities than their typically developing peers in the for- mation of cross-status interactions. These find- ings reinforce the idea that forming cross-status interactions is more challenging than forming same-status interactions. Another possible
Lin et al. 447
explanation could be drawn from the homoph- ily literature. Studies of social homophily indi- cate that children prefer to interact with peers with whom they share similar characteristics (e.g., Martin et al., 2013). Interpersonal simi- larities contribute not only to more likeness with others but also to more shared experi- ences, which could potentially lead to more mutual understanding and effective communi- cation. Children with and without disabilities may have insufficient shared experiences to form interactive play and therefore require greater self-regulation and pragmatic language skills to coordinate their play.
Pragmatic language seemed to play a more important role for children with disabilities than their typically developing peers in the formation
of cross-status interactions.
The influence of self-regulation on the for- mation of cross-status interactions was found to be fully mediated by pragmatic language, suggesting that children with higher self- regulation are better at applying their prag- matic language to form cross-status play interactions. This finding partially supports the mutual development of language and self-regulation in the early childhood stage (Bohlman et al., 2015) but is inconsistent with previous studies showing either moderating (Day & Smith, 2013) or supporting (Vallotton & Ayoub, 2011) effects of language on self- regulation. It is important to note that previ- ous research mostly focused on self-regulation and its relationships with children’s expres- sive or receptive language skills (e.g., Day & Smith, 2013; Valloton & Aboub, 2011). In contrast, the current study addressed the rela- tionship between self-regulation and prag- matic language given that pragmatic language was more pertinent to our outcome of interest, children’s play interactions (Dionne et al., 2003). Compared with expressive or receptive language, pragmatic language might require extra social-cognitive resources (e.g., selec- tive attention to important social contextual cues or emotional regulation toward others’
social actions) in order to be more adequately used in complex social situations. It stands to reason that self-regulation is a necessary capacity to empower children’s use of prag- matic language in their social networks.
Children with higher self- regulation are better at applying their pragmatic language to form
cross-status play interactions.
Regarding potential between-classroom differences of classroom play networks, we included proportions of children with disabili- ties, gender ratio, and class size in the ERGMs as control variables. None of the classroom- level variables significantly predicted cross- status or same-status play interactions except for class size. Even though gender ratio showed a significant effect in the cross-status model, the effect seemed to be tangential and unstable. As shown in Table 4, the effect of gender ratio was not significant in Model 1 to Model 4 and was barely significant in Model 5. The negative effect of class size on same- status interactions supports previous social network research that the tendency for chil- dren to form an additional social network tie decreases as the network size increases (Sni- jders et al., 2010). However, this finding needs to be interpreted with caution, as the class sizes reported in this study were constrained by the number of participating children and might not represent the total class size or even the actual ratio of children with and without disabilities in these inclusive classrooms.
The study has several limitations. First, certain interactions that were not plausible to occur, called structural zeros, could not be fully constrained in the ERGMs. In this study, all the between-classroom interactions for both cross- and same-status models have been fully constrained as structural zeros. To con- strain structural zeros within classrooms, however, ERGMs required a binary matrix where all the implausible interactions were filled with 0s and all the estimable cells were filled with 1s. These estimable cells in the structural zero matrix must be symmetric.
448 Exceptional Children 85(4)
However, the structural zero matrix in the cross-status model in each classroom was an off-diagonal rectangle, where the length and width were the number of children with dis- abilities (width) and the number of typically developing children (length). When implausi- ble play interactions were treated as plausible and absent interactions (i.e., real zeros), the probability of cross-status interactions might have been underestimated, which in turn attenuated the effects of main predictors (e.g., pragmatic language, self-regulation) on class- room social networks.
The timing of the child measures in this study might potentially bias our study find- ings. Children’s pragmatic language was assessed in the winter and spring semesters, but self-regulation skills were assessed only in the fall. The rapid development in these skills during the preschool year might change the relationship between these malleable child characteristics and their interactions in the classroom. To test this hypothesis, we replaced the spring language scores with the winter language scores in the ERGMs, and the results led to the same conclusions. The spring scores were related to (r = .85) and greater than the fall scores (p < .001). These results indicate that even though children’s pragmatic lan- guage skills developed rapidly, the rank ordering of children on the posttest was simi- lar to the rank ordering on the pretest, which yielded a stable relationship between prag- matic language and classroom play networks. We conjecture that this stable relationship might apply to children’ self-regulation skills as well, providing that pragmatic language and self-regulation were positively correlated.
Even though the significant influence of pragmatic language and self-regulation skills on children’s play interactions empirically supports the social learning theory (Vygotsky, 1978), this study cannot exclude the possibil- ity of overlooking other malleable child or dyadic characteristics, such as children’s atti- tudes toward peers with disabilities (De Boer, Pijl, Post, & Minnaert, 2013) or interest in similar play activities (Dietrich, 2005). Fur- thermore, the peer social networks were derived from teacher report. The extent to
which teachers introduced their own biases toward certain students in the measurement was uncertain. However, the teacher rating scale has shown construct validity in this study, which is consistent with research sug- gesting that teachers are reliable evaluators of preschool children’s behaviors and skills based on their ongoing observations (Cabell, Justice, Zucker, & Kilday, 2009; Meisels, Liaw, Dorfman, & Nelson, 1995). Still, these teachers may have introduced biases into their perceptions of classroom social networks in some meaningful ways (e.g., Gest, 2006; Neal, Cappella, Wagner, & Atkins, 2011). Future research should identify factors associ- ated with these teacher biases.
It is important to note that children’s social networks were based on one-time measure- ment assessed at the end of the academic year. Longitudinal studies are needed to tap into the reciprocal causal relationships between self- regulation and pragmatic language and their joint influence on classroom play networks. Specific types of disability for at least 21% of children were unknown, which prevented us from comparing the social interactions of children of different types of disabilities. This issue should be investigated in the future. Last, the majority of the classrooms in our sample had fewer than 41% of children with disabilities who were moderate or high func- tioning. The result might not be generalizable to classrooms in which a large proportion of children have disabilities or to classrooms that include children with severe disabilities.
The current study investigated young chil- dren’s social networks underlying inclusive preschool classrooms based on pooled class- room play networks. Although this approach allowed us to derive more reliable statistical estimates of children’ social interactions in ERGMs, it contradicts the nested nature of the classroom data. Future research is therefore needed to identify unique classroom-level factors associated with cross-status and same- status social interactions in inclusive class- rooms.
As the percentage of children with disabili- ties who spend a majority of educational time in inclusive classrooms continues to increase
Lin et al. 449
over time, findings from the present study underscore the importance of fostering chil- dren’s ability to form positive cross-status interactions to establish more socially accept- ing and integrative classroom environments. The current study particularly calls for the need to expose children to inclusive practices in their early preschool age when their lan- guage and self-regulation skills are the most malleable and adaptive to change (Bohlmann et al., 2015). Early intervention may poten- tially lower the rate of peer victimization and its negative influences on children’s academic and social development in the long run (Ladd, Ettekal, & Kochenderfer-Ladd, 2017).
Although research on inclusion of children with disabilities in general education class- rooms has strongly focused on risk factors and negative effects, the present study took educa- tional and developmental perspectives on the influences of pragmatic language and self- regulation skills in the social integration of children with and without disabilities in the classroom. Our findings thus foreshadow the need for teachers to adapt their inclusion prac- tices based on children’s pragmatic language and self-regulation skills or, more broadly speaking, social cognitive skills. Considering that language and self-regulation have been reliably associated with children’s academic and social outcomes (Duncan et al., 2007), our findings indicate that educational programs designed to scaffold children’s development in pragmatic language and self-regulation, such as collaborative small-group work featuring high-quality social interactions, may yield synergistic effects on the development of chil- dren with and without disabilities.
References Asher, S. R., Parker, J. G., & Walker, D. L. (1996).
Distinguishing friendship from acceptance: Implications for intervention and assessment. In W. M. Bukowski, A. F. Newcomb, & W. W. Hartup (Eds.), The company they keep: Friendships in childhood and adolescence (pp. 366–405). New York, NY: Cambridge University Press.
Barnett, W. S., Jung, K., Yarosz, D. J., Thomas, J., Hornbeck, A., Stechuk, R., & Burns, S. (2008). Educational effects of the Tools of the
Mind curriculum: A randomized trial. Early Childhood Research Quarterly, 23, 299–313. https://doi.org/10.1016/j.ecresq.2008.03.001
Bauminger, N., Solomon, M., Aviezer, A., Heung, K., Brown, J., & Rogers, S. J. (2008). Friendship in high-functioning children with autism spectrum disorder: Mixed and non-mixed dyads. Journal of Autism and Developmental Disorders, 38, 1211–1229. https://doi.org/10.1007/s10803-007-0501-2
Bohlmann, N. L., Maier, M. F., & Palacios, N. (2015). Bidirectionality in self-regulation and expressive vocabulary: Comparisons between monolingual and dual language learners in pre- school. Child Development, 86, 1094–1111. https://doi.org/10.1111/cdev.12375
Bulotsky-Shearer, R. J., Bell, E. R., Romero, S. L., & Carter, T. M. (2012). Preschool interac- tive peer play mediates problem behavior and learning for low-income children. Journal of Applied Developmental Psychology, 33, 53–65. https://doi.org/10.1016/j.appdev.2011.09.003
Buysse, V. (1993). Friendships of preschoolers with disabilities in community-based child care set- tings. Journal of Early Intervention, 17, 380–395. https://doi.org/10.1177/105381519301700404
Cabell, S. Q., Justice, L. M., Zucker, T. A., & Kilday, C. R. (2009). Validity of teacher report for assessing the emergent literacy skills of at-risk preschoolers. Language, Speech, and Hearing Services in Schools, 40, 161–173. https://doi.org/10.1044/0161-1461
Chen, J., Lin, T.-J., Justice, L., & Sawyer, B. (2017). The social networks of children with and without disabilities in early childhood spe- cial education classrooms. Journal of Autism and Developmental Disorders. Advance online publication. https://doi.org/10.1007/s10803- 017-3272-4
Coolahan, K., Fantuzzo, J., Mendez, J., & McDermott, P. (2000). Preschool peer interactions and readi- ness to learn: Relationships between class- room peer play and learning behaviors and conduct. Journal of Educational Psychology, 92, 458–465. http://dx.doi.org/10.1037/0022- 0663.92.3.458
Day, K. L., & Smith, C. L. (2013). Understanding the role of private speech in children’s emo- tion regulation. Early Childhood Research Quarterly, 28, 405–414. https://doi.org/10.1016 /j.ecresq.2012.10.003
De Boer, A., Pijl, S. J., Post, W., & Minnaert, A. (2013). Peer acceptance and friendships of students with disabilities in general education:
450 Exceptional Children 85(4)
The role of child, peer, and classroom variables. Social Development, 22, 831–844. https://doi. org/10.1111/j.1467-9507.2012.00670.x
Diamond, K. E. (2001). Relationships among young children’s ideas, emotion under- standing and social contact with classmates with disabilities. Topisc in Early Childhood Special Education, 21, 104–113. https://doi. org/10.1177/027112140102100204
Dietrich, S. L. (2005). A look at friendships between preschool-aged children with and without dis- abilities in two inclusive classrooms. Journal of Early Childhood Research, 3, 193–215. https://doi.org/10.1177/1476718X05053933
Dionne, G., Tremblay, R., Boivin, M., Laplante, D., & Perusse, D. (2003). Physical aggression and expressive vocabulary in 19-month-old twins. Developmental Psychology, 39, 261. https://doi.org/10.1037/0012-1649.39.2.261
Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P., . . .Japel, C. (2007). School readiness and later achievement. Developmental Psychology, 43, 1428–1446. https://doi.org/10.1037/0012- 1649.43.6.1428
Eisenberg, N., Valiente, C., & Eggum, N. D. (2010). Self-regulation and school readiness. Early Education and Development, 21, 681–698. https://doi.org/10.1080/10409289.2010.497451
Farmer, T. W., & Farmer, E. (1996). Social rela- tionships of students with exceptionalities in mainstream classrooms: Social networks and homophily. Exceptional Children, 62, 431–450. https://doi.org/10.1177/001440299606200504
Gest, S. D. (2006). Teacher reports of children’ s friendships and social groups: Agreement with peer reports and implications for studying peer Similarity. Social Development, 15, 248–259.
Gottfried, M. A. (2016). Peer sameness and peer diversity: The influence of breadth and depth of classmates with high academic needs in kin- dergarten. Early Education and Development, 27, 655–675. https://doi.org/10.1080/1040928 9.2016.1127087
Guralnick, M. J. (1997). Peer social networks of young boys with developmental delays. American Journal of Mental Retardation, 101, 595–612. Retrieved from http://www.ncbi. nlm.nih.gov/pubmed/9152475
Guralnick, M. J. (2006). Peer relationships and the mental health of young children with intellec- tual delays. Journal of Policy and Practice in Intellectual Disabilities, 3, 49–56.
Guralnick, M. J., & Bruder, M. B. (2016). Early childhood inclusion in the United States: Goals,
current status, and future directions. Infants & Young Children, 29, 166–177. https://doi .org/10.1097/IYC.0000000000000071
Guralnick, M. J., Gottman, J. M., & Hammond, M. A. (1996). Effects of social setting on the friendship formation of young children differ- ing in developmental status. Journal of Applied Developmental Psychology, 17, 625–651. https:// doi.org/10.1016/S0193-3973(96)90019-2
Handcock, M., Hunter, D., Butts, C., Goodreau, S., Krivitsky, P., Bender-deMoll, S., & Morris, M. (2016). statnet: Software tools for the statistical analysis of network data. R packgage Version 2016, 9, CRAN.R-project.org/package=statnet. Retrieved from http://www.statnet.org
Hestenes, L. L., & Carroll, D. E. (2000). The play interactions of young children with and without disabilities: Individual and environ- mental influences. Early Childhood Research Quarterly, 15, 229–246. https://doi.org/10.1016 /S0885-2006(00)00052-1
Howes, C. (1988). Peer interaction of young chil- dren. Monographs of the Society for Research in Child Development, 53, 1–92. http://dx.doi. org/10.2307/1166062
Howes, C. (2011). Social play of children with adults and peers. In P. Nathan, & A. D. Pellegrini (Eds.), The Oxford handbook of the development of play (pp. 231–244). New York, NY: Oxford University Press.
Howes, C., Rubin, K., Ross, H., & French, D. (1988). Peer Interaction of Young Children. Monographs of the Society for Research in Child Development, 53, I-92. http://doi.org /10.2307/1166062
Hunter, D. R. (2007). Curved exponential family models for social networks. Social Networks, 29, 216–230. https://doi.org/10.1016/j.socnet.2006 .08.005
Hunter, D. R., Goodreau, S. M., & Handcock, M. S. (2008). Goodness of fit of social network models. Journal of the American Statistical Association, 103, 248–258. https://doi.org /10.1198/016214507000000446
Kwon, K.-A., Elicker, J., & Kontos, S. (2011). Social IEP objectives, teacher talk, and peer interaction in inclusive and segregated pre- school settings. Early Childhood Education Journal, 39, 267–277. https://doi.org/10.1007/ s10643-011-0469-6
Ladd, G. W., Ettekal, I., & Kochenderfer-Ladd, B. (2017). Peer victimization trajectories from kindergarten through high school: Differential pathways for children’s school engagement and achievement? Journal of Educational
Lin et al. 451
Psychology, 109, 1–16. https://doi.org/10.1037 /edu0000177
Lillard, A. S., Hopkins, E. J., Dore, R. A., Palmquist, C. M., Lerner, M. D., & Smith, E. D. (2013). Concepts and theories, methods and reasons: Why do the children (pretend) play? Reply to Weisberg, Hirsh-Pasek, and Golinkoff (2013); Bergen (2013); and Walker and Gopnik (2013). Psychological Bulletin, 139, 49–52. https://doi.org/10.1037/a0030521
Lin, T.-J., Justice, L. M., Paul, N., & Mashburn, A. J. (2016). Peer interaction in rural pre- school classrooms: Contributions of children’s learning-related behaviors, language and lit- eracy skills, and problem behaviors. Early Childhood Research Quarterly, 37, 106–117. https://doi.org/10.1016/j.ecresq.2016.04.001
Martin, C. L., Kornienko, O., Schaefer, D. R., Hanish, L. D., Fabes, R. A., & Goble, P. (2013). The role of sex of peers and gender-typed activities in young children’s peer affiliative networks: A longitudinal analysis of selection and influence. Child Development, 84, 921– 937. https://doi.org/10.1111/cdev.12032
McClelland, M. M., & Morrison, F. J. (2003). The emergence of learning-related social skills in preschool children. Early Childhood Research Quarterly, 18, 206–224.
McDermott, P. A., Leigh, N. M., & Perry, M. A. (2002). Development and validation of the Preschool Learning Behaviors Scale. Psychology in the Schools, 39, 353–365. https://doi.org/10.1002/pits.10036
Mcfarland, D. A., Moody, J., Diehl, D., Smith, J. A., & Thomas, R. J. (2014). Network ecol- ogy and adolescent social structure. American Sociological Review, 79, 1088–1121. https:// doi.org/10.1177/0003122414554001
Mcpherson, M., Smith-lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415–444. https://doi.org/10.1146/annurev. soc.27.1.415
Meisels, S. J., Liaw, F., Dorfman, A., & Nelson, R. F. (1995). The work sampling system: Reliability and validity of a performance assess- ment for young children. Early Childhood Research Quarterly, 10, 277–296. https://doi .org/10.1016/0885-2006(95)90008-X
National Center for Education Statistics. (2016). Digest of education statistics. Washington, DC: Institute of Education Sciences. Retrieved from https://nces.ed.gov/programs/digest/d16/ tables/dt16_204.50.asp
Neal, J. W., Cappella, E., Wagner, C., & Atkins, M. S. (2011). Seeing eye to eye: Predicting teacher- student agreement on classroom social networks. Social Development, 20, 376–393. https://doi. org/10.1111/j.1467-95072010.00582.x.
Odom, S. L., Klingerman, K., & Jakowski, M. (2004). Investigating inclusion: A review of research methods for individuals with intellec- tual disability. In E. Emerson & C. H. Travis (Eds.), The international handbook of applied research in intellectual disabilities (pp. 281– 295). Chichester, UK: Wiley.
O’Neill, D. K. (2007). The Language Use Inventory for young children: A parent-report measure of pragmatic language development for 18- to 47-month-old children. Journal of Speech Language and Hearing Research, 50, 214–228. https://doi.org/10.1044/1092-4388(2007/017)
Pattison, P., & Wasserman, S. (1999). Logit models and logistic regressions for social networks: II. Multivariate relations. British Journal of Mathematical and Statistical Psychology, 52, 169–193.
Płatos, M., & Wojaczek, K. (2018). Broadening the scope of peer-mediated intervention for individuals with autism spectrum disor- ders. Journal of Autism and Developmental Disorders, 48, 747–750.
Prutting, C. A., & Kittchner, D. M. (1987). A clinical appraisal of the pragmatic aspects of language. Journal of Speech and Hearing Disorders, 52, 105–119. https://doi.org/10.1044/jshd.5202.105
Rafferty, Y., Piscitelli, V., & Boettcher, C. (2003). The impact of inclusion on language develop- ment and social competerne among preschool- ers with disabilities. Exceptional Children, 69, 467–479. https://doi.org/10.1177/0014402 90306900405
Reilly, S., Wake, M., Ukoumunne, O. C., Bavin, E., Prior, M., Cini, E., . . . Bretherton, L. (2010). Predicting language outcomes at 4 years of age: Findings from Early Language in Victoria Study. Pediatrics, 126, e1530–e1537. https:// doi.org/10.1542/peds.2010-0254
Robins, G., Pattison, P., & Wang, P. (2009). Closure, connectivity and degree distributions: Exponential random graph (p*) models for directed social net- works. Social Networks, 31, 105–117. https://doi. org/10.1016/j.socnet.2008.10.006
Roth, F. P., & Spekman, N. J. (1984a). Assessing the pragmatic abilities of children: Part 1. Organizational framework and assessment parameters. Journal of Speech and Hearing Disorders, 49, 2–11.
452 Exceptional Children 85(4)
Roth, F. P., & Spekman, N. J. (1984b). Assessing the pragmatic abilities of children: Part 2. Guidelines, considerations, and specific eval- uation procedures. Journal of Speech and Hearing Disorders, 49, 12–17.
Rubin, K. H., Bukowski, W. M., & Parker, J. G. (2006). Peer interactions, relationships, and groups. In N. Eisenberg (Ed.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (pp. 571–645). New York, NY: Wiley.
Schuler, A. L., & Fletcher, E. C. (2002). Making communication meaningful: cracking the lan- guage interaction code. In R. L. Gabriels & D. E. Hill (Eds.), Autism: From research to indi- vidualized practice (pp. 127–154). London, PA: Jessica Kingsley.
Smith-Donald, R., Raver, C. C., Hayes, T., & Richardson, B. (2007). Preliminary construct and concur- rent validity of the Preschool Self-Regulation Assessment (PSRA) for field-based research. Early Childhood Research Quarterly, 22, 173– 187. https://doi.org/10.1016/j.ecresq.2007.01.002
Snijders, T. A. B., Van de Bunt, G. G., & Steglich, C. E. G. (2010). Introduction to stochastic actor-based models for network dynamics. Social Networks, 32, 44–60.
Son, E., Peterson, N. A., Pottick, K. J., Zippay, A., Parish, S. L., & Lohrmann, S. (2014). Peer victimization among young children with disabilities: Early risk and protective factors. Exceptional Children, 80, 368–384. https:// doi.org/10.1177/0014402914522422
Staikova, E., Gomes, H., Tartter, V., McCabe, A., & Halperin, J. M. (2013). Pragmatic deficits and social impairment in children with ADHD. Journal of Child Psychology and Psychiatry, 54, 1275–1283. https://doi.org/10.1111/jcpp.12082
Strain, P. S., & Odom, S. L. (1986). Peer social initiations: Effective intervention for social skills development of exceptional children. Exceptional Children, 52, 543–551. https:// doi.org/10.1177/001440298605200607
UNESCO. (1994). The Salamanca statement and the framework for action on special needs edu- cation. Paris, France: Author.
U.S. Department of Health and Human Services & U.S. Department of Education (2015). Policy statement on inclusion of children with dis- abilities in early childhood programs. Infants & Young Children, 29, 3–24. https://doi. org/10.1097/IYC.0000000000000054
Valiente, C., Lemery-Chalfant, K., Swanson, J., & Reiser, M. (2008). Prediction of children’s aca- demic competence from their effortful control,
relationships, and classroom participation. Journal of Educational Psychology, 100, 67–77. https:// doi.org/10.1037/0022-0663.100.1.67.Prediction
Vallotton, C., & Ayoub, C. (2011). Use your words: The role of language in the development of toddlers’ self-regulation. Early Childhood Research Quarterly, 26, 169–181. https://doi. org/10.1016/j.ecresq.2010.09.002
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological pro- cess. Cambridge, MA: Harvard University Press.
Walker, S., & Berthelsen, D. (2008). Children with autistic spectrum disorder in early child- hood education programs: A social construc- tivist perspective on inclusion. International Journal of Early Childhood, 40, 33–51. https:// doi.org/10.1007/BF03168362
Watkins, L., O’Reilly, M., Kuhn, M., Gevarter, C., Lancioni, G. E., Sigafoos, J., & Lang, R. (2015). A review of peer-mediated social inter- action interventions for students with autism in inclusive settings. Journal of Autism and Developmental Disorders, 45, 1070–1083. https://doi.org/10.1007/s10803-014-2264-x
Weiland, C. (2016). Impacts of the Boston Prekindergarten Program on the school readi- ness of young children with special needs. Developmental Psychology, 52, 1763–1776. https://doi.org/10.1037/dev0000168
Wiig, E. H., Secord, W., & Semel, E. M. (2004). CELF Preschool 2: Clinical evaluation of lan- guage fundamentals preschool. San Antonio, TX: Pearson/PsychCorp.
Winsler, A., Diaz, R. M., & Montero, I. (1997). The role of private speech in the transition from collaborative to independent task performance in young children. Early Childhood Research Quarterly, 12, 59–79. https://doi.org/10.1016/ S0885-2006(97)90043-0
Yu, S., Ostrosky, M. M., & Fowler, S. A. (2015). The relationship between preschoolers’ atti- tudes and play behaviors toward classmates with disabilities. Topics in Early Childhood Special Education, 35, 40–51. https://doi. org/10.1177/0271121414554432
Zhang, J., & Wheeler, J. J. (2011). A meta-analysis of peer-mediated interventions for young chil- dren with autism spectrum disorders. Education and Training in Autism and Developmental Disabilities, 46, 62–77. Retrieved from https:// www.jstor.org/stable/23880031
Manuscript received January 2018; accepted January 2019.
Copyright of Exceptional Children is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
