Friendships in preschool children with autism spectrum disorder: What holds them back,child characteristics or teacher behavior?

Abstract

Children begin to show preferences for specific playmates as early as the first 2 years of life. Children with autism spectrum disorder have difficulty making friends, even in elementary and middle school. However, very little is known about earlier friendships in children with autism such as preschool friendships. This study examined friendships in preschool children with autism and explored how joint attention contributes to these friendships in mainstream settings. A secondary aim was to determine the extent to which teachers used strategies to facilitate friendship development. The participants were 31 mainstreamed preschool children (ages 2–5 years) with autism spectrum disorder. School observations were conducted individually to capture participants’ interactions with peers and adults during free play. The results indicated that 20% of the participants had friendships at school. Children with friends were more likely than children without friends to be jointly engaged with their peers during free play, and they used higher joint attention skills. Teachers used few friendship facilitating strategies, and more often used behavioral management strategies within the classrooms. Future studies may want to examine the effects of early interventions and/or teacher training on the development of friendships in preschool children with autism spectrum disorder within the school setting.

Keywords

autism spectrum disorders friendship joint attention joint engagement preschool

Much has been written about the difficulties in developing friendships for children with autism spectrum disorder (ASD). About 20% of elementary aged children with ASD in mainstream classes report reciprocal friendships compared to 60% of neurotypical children in the same classes (Chamberlain et al., 2007; Kasari et al., 2011). Children with ASD report larger social networks at younger ages than at older ages (Rotheram-Fuller et al., 2010) with the majority of adolescents and adults reporting few if any friends (Howlin et al., 2013). The extent of the difficulties in friendship development has been well documented with greater difficulty in friendship development at the older ages, but studies have not examined friendship development in the youngest children. The aim of this study was to first determine whether reciprocal friendships exist in preschool aged children with ASD according to observers, parents, and teachers. Second, we examined both child social communication characteristics and teacher strategies that predicted friendships among children with ASD and their typical classmates.

Friendship development in preschool children

Friendships have been identified in neurotypical children as young as 2 years of age or even younger (Vandell and Mueller, 1980). Infants and toddlers begin to show preference for specific playmates by sharing affect and engaging in prosocial behaviors differentially to some children versus others (Howes, 1983; Vandell and Mueller, 1980). Even at a young age, friendships are important. In neurotypical development, friendships have been associated with social and emotional competence in young children. Compared to children without friends, children with friends exhibit more prosocial behaviors and are more likely to be accepted and less likely to be rejected and bullied by peers (Howes, 1988; Sebanc, 2003; Vaughn et al., 2001). Children who have reciprocated friendships also demonstrate higher play levels, higher language levels, better conflict resolution skills, and more positive affect (Fujisawa et al., 2008; Gottman, 1983; Howes, 1988; Vaughn et al., 2001). Studies examining friendships in young children have mainly used observational measures. These methods of using behavioral indicators of friendships before children can reliably verbalize their preferences have been used across multiple studies of neurotypical children. As such, this measurement approach may have applicability to preschool aged children with ASD who have difficulty giving verbal preferences.

Rarely have preschool children with ASD been observed in their mainstream settings for evidence of friendships. Most studies that have examined friendships have been between typically developing children and children with disabilities, not specific to autism. These studies indicate that typically developing children are more preferred playmates than children with disabilities (Guralnick and Groom, 1987). In recent years, there have been more studies looking specifically at children with autism. However, these studies examining friendships in preschool children with ASD have all been based on parent and teacher reports (Bauminger-Zviely and Agam-Ben-Artzi, 2014; Bauminger-Zviely et al., 2014; Kimhi and Bauminger-Zviely, 2012). These studies indicate that interactions with friends compared to non-friends are qualitatively different. Interactions are more socially complex between friends; children show higher reciprocity and increased shared enjoyment.

Joint attention and social development in children with autism

Several early core impairments may affect the extent to which young children with autism can engage with their peers. Joint attention is a core deficit that may not develop or is diminished in frequency and quality for young children with autism (Loveland and Landry, 1986; Mundy et al., 1987; Sigman et al., 1986). This early social communication skill is a core deficit specific to children with autism and is also a significant predictor for peer engagement and prosocial behavior during elementary and middle school years for children with autism (Sigman and Ruskin, 1999; Travis et al., 2001). Children with autism who initiate more joint attention at a young age engage in higher levels of social play during elementary and middle school years.

These studies suggest that deficits in joint attention can potentially negatively impact the social competence of children with autism throughout their lives. They are less likely to engage in helpful and prosocial behaviors (Travis et al., 2001), and these social difficulties can extend well into their middle school years. For example, if a child is not using gestures and/or language for the purpose of sharing an object or an experience with another person (joint attention), he or she may have more difficulty initiating and maintaining a social interaction with a peer. Most children engage in a game or activity with a peer to initiate an interaction, which can be a beginning of a potential friendship. To do so, the child will need to know how to comment appropriately (joint attention language) and to share and take turns (joint attention give) during these activities.

Adult strategies for peer relationships in mainstream classrooms

In addition to child characteristics, teachers play an important role in the social development of preschool children with disabilities, particularly in promoting friendships in mainstream classrooms. Studies have found that teachers use various strategies to promote friendships in children with disabilities in their classrooms (Buysse et al., 2008; Hollingsworth and Buysse, 2009). For example, teachers can use passive strategies such as supervising dyadic play in their classrooms or suggesting specific playmates for children with autism. Teachers also utilize more active strategies which can range from inserting themselves into the dyad’s interaction (i.e. taking a role in the children’s play) to facilitating the dyad’s interaction (i.e. suggesting a specific activity, redirecting inappropriate behaviors). Other strategies to promote friendships include using friendship stories, puppets to role-play activities, and modeling appropriate social skills for the children (Buysse et al., 2008; Hollingsworth and Buysse, 2009).

However, adults can sometimes impede social interaction and friendship development in mainstream settings. Studies have found that children with disabilities are more likely to interact with adults than their peers in mainstream settings (Anderson et al., 2004; Brown et al., 1999; Kishida and Kemp, 2009). Anderson et al. (2004) speculated that typically developing children may be more hesitant to play with children with autism when there is an adult “shadowing” them because they either want to be away from the adults or they may find the adult stigmatizing.

Most of the studies on inclusion for preschool children with disabilities have mainly examined the social interactions between children with and without disabilities (Brown et al., 1999; Odom et al., 2006). To date, only one study has specifically examined social interactions of preschool children with autism in mainstream settings (Kishida and Kemp, 2009). This study examined the overall engagement and quality of interaction of children with autism. This study builds on the previous study by specifically focusing on friendships in mainstream settings for preschool aged children with ASD. The goals were to (a) identify the extent to which friendships exist in preschool children with autism, (b) examine how early social communication skills affect friendships, and (c) how teachers support friendship development in mainstream classrooms.

Method

Participants

The sample included 31 preschool children (26 males and 5 females) with ASDs, their parents (n = 31), and teachers (n = 25). The children’s age ranged from 31 to 60 months (M = 46.16, standard deviation (SD) = 7.00). Children were recruited from mainstream preschool classrooms in the Greater Los Angeles area (14 schools, 25 classrooms). Flyers were given to school administers and teachers, and families who were interested in the study contacted the research team and were given an eligibility phone screen.

All participants met inclusion criteria for an ASD based on the Autism Diagnostic Observation Schedule (ADOS): autism (n = 18) and ASD (n = 13). The sample was ethnically heterogeneous, comprising Caucasians (25.8%), Asians (29%), Hispanics (25.8%), African Americans (9.7%), and other ethnic groups (9.7%). Child characteristics are presented in Table 1.

Table 1.

Child demographics (N = 31).

Variables	M	SD	Range	Ratio	%
Chronological age (months)	46.16	7.00	31–60
MSEL (age equivalent)
Mental age (months)	32.83	5.89	24–51
Developmental quotient	72.64	16.38	50–116
Receptive language	30.97	7.71	16–55
Expressive language	28.39	7.93	12–55
Gender (male:female)				26:5
Ethnicity
Caucasian					25.8
Asian					29.0
Hispanic					25.8
African American					9.7
Other					9.7
Class size	14.31	3.22	8–18
Classroom ratio (adult:child)				3:14

SD: standard deviation; MSEL: Mullen Scales of Early Learning.

Child measures

Diagnostic assessment

The ADOS (Lord et al., 2001) was used to assess whether children met criteria for autism. The assessment was administered at entry by the research team (graduate students who are research reliable on the ADOS). The ADOS is a semi-structured, standardized assessment of communication, social interaction and play, and imaginative use of materials. It is designed to assess for a clinical presentation of autism or other pervasive developmental disorders. The ADOS yields four domains: language/communication, reciprocal social interaction, play and imagination/creativity, and stereotyped behaviors and restricted interests.

Cognitive assessment

The Mullen Scales of Early Learning (MSEL) (Mullen, 1989) was used to assess general cognitive ability. The MSEL yields age-equivalent scores for young children from birth to 6 years of age. Four cognitive domains were assessed: visual reception, fine motor, receptive, and expressive language. T-scores and age equivalents are available for each subscale. A developmental quotient was also calculated from the combined subscales.

Joint attention assessment

The Early Social-Communication Scales (ESCS; Mundy et al., 1986; Seibert et al., 1982) is a semi-structured assessment used to assess joint attention and behavior regulation (i.e. requesting) skills. Studies have shown good reliability and validity for this measure (Mundy et al., 1986). The ESCS was videotaped and later coded for frequencies of both initiating and responding to joint attention and behavior regulation. Reliability for blind coders was established at 80% for each category: initiating joint attention and behavior regulation. The overall intraclass coefficient of coders was 0.91, range 0.87–0.98.

Behavioral observation

This behavioral coding system is designed to capture the target child’s interactions with their peers and adults during free play in the classroom. Each child was observed for two 15-min interactions with their peers at his or her preschool. Both observations were completed within 1 month of enrollment in the study.

The coding system is adapted from Kasari et al. (unpublished). Data were collected using interval-coding methods (Alberto and Troutman, 2006). Each observation was divided into 50-s time intervals. After the end of each interval, the observer noted a series of skills for each target child: engagement states, child initiations, and adult strategies are used to engage the child (see Appendix 1 for coding details). There were six different adult strategies that were measured: adult use of joint attention gestures or language, behavioral regulation gestures or language, environmental arrangement, prompting target child, recruit other children to engage target child, and whether target child was with the adult for half or more of the interaction session. These strategies were selected from autism-specific interventions that increased young children’s social communication skills (Kasari et al., 2012; Lawton and Kasari, 2012).

This observational measure was also used to define friendship. Friendship is defined by three criteria: (a) at least 50% of the target child’s social initiation attempts were responded to, (b) at least one unit of joint engagement or games during the interaction, and (c) at least one positive affective exchange (measure adapted from Howes, 1983).

Blind coders were trained to reliability criteria of 80% for each category. Of all the observations, 20% were coded by two blind observers. Intraclass correlations were used to estimate reliability; the overall intraclass coefficient between two independent coders was 0.85, range 0.64–0.99.

Parent and teacher measure

Friendship questionnaire

The friendship questionnaire is adapted from Goldman and Buysse (2005). Both the child’s parent and teacher were asked to identify three of the child’s friends in class. From the three identified friends, parents and teachers were asked to rate the quality of the child’s best friendship. The quality of the best friendship was rated on a 5-point Likert scale on 19 items, including positive characteristics (i.e. share with each other, play happily together) and negative characteristics (i.e. fight verbally, grab and take things from each other).

Statistical methods

Comparisons in demographic variables between the groups (with and without friends) were evaluated using analysis of variance (ANOVA) for continuous variables, and χ² or Fisher exact tests were used to evaluate distribution of categorical variables between the groups. Skewness and kurtosis were examined for each variable. Time spent in different engagement states and initiating joint attention exhibited deviations from normality because of the high prevalence of zeroes since many of the children did not spend any time in joint engagement or display any joint attention skills.

Due to violation of the normality assumption, permutation test was conducted instead of non-parametric test (i.e. Mann–Whitney test) to examine differences between the two groups because of the zeroes (i.e. ties). In addition, associations between continuous variables were assessed using Pearson’s correlation. Multiple comparisons were all adjusted using Dunn (1961) multiple-comparison test.

Results

Definition of friendship

Parent reports, teacher reports, and school observations were used to identify friendships in preschool children with autism. Of the parents, 42% (n = 13) in the sample reported that their child had a friend in class, and 54% (n = 17) of the teachers in the sample reported that the child with autism had a friend in class. Using a more stringent criteria to define friendship, observational data revealed that approximately 20% of the sample had friendships (n = 7) as defined by three criteria: (a) at least 50% of their social initiation attempts were responded to, (b) at least one unit of joint engagement or games during the interaction, and (c) at least one positive affective exchange (adapted from Howes, 1983).

Child characteristics of children with ASD with and without friends

There were no significant differences in age, cognition, language, social communication, or play skills between children with and without friends as identified by the parents or the teachers. Therefore, subsequent analyses between groups with and without friends are based on the definition adapted from Howes (1983) using the observation measure.

A series of correlations were conducted to determine associations between child characteristics and social communication skills (see Table 2). Notably, mental age was correlated with total requesting skills (r = 0.386, p = 0.032 and response to joint attention (r = 0.420, p = 0.019). Both receptive and expressive language skills were positively correlated with response to joint attention (r = 0.434, p = 0.015; r = 0.463, p = 0.009, respectively).

Table 2.

Correlations of child demographics and social communication skills (N = 31).

	Joint attention skills (total)	Behavioral regulation skills (total)	Response to joint attention
Chronological age	−0.246	−0.018	−0.087
Mental age (Mullen)	0.325	0.386*	0.420*
Receptive language (Mullen)	0.251	0.255	0.434*
Expressive language (Mullen)	0.515**	0.339	0.463**

p ⩾ 0.05.

p ⩾ 0.01.

Fisher’s exact tests were also conducted to determine differences between the two groups in gender, ethnicity, and diagnosis. Although there were no group differences in gender or ethnicity, the analysis revealed that children with friends were more likely to be diagnosed with ASD rather than autism when compared to children without friends (p = 0.012). In addition, there were no cognitive or language differences between children who were diagnosed with autism and children who were diagnosed with ASD.

Social engagement, social initiations, and social communication outcomes

Joint attention skills were assessed with the ESCS. Both low-level (eye contact and giving) and high-level (pointing and showing) joint attention skills were examined. There were no differences in low-level joint attention skills between the two groups, but there was a group difference for high-level joint attention skills (p = 0.01). Six of seven of the children with friends displayed high-level joint attention, whereas only 12 of 24 children without friends displayed high-level joint attention skills.

Engagement states were measured using the behavioral observation coding system. Six different engagement states were examined to evaluate differences in the percentage of time children with and without friends spent being engaged during free play: solitary, proximity, onlooker, parallel, parallel aware, and joint engagement (see Graph 1). As a group, children in the sample spent the most time in solitary play (M = 33.71, standard deviation (SD) = 24.43) compared to other engagement states. There were no group differences between children with and without friends in the percentage of time they spent in solitary play.

Graph 1.

Observation of free play in classroom: percentage of time in engagement states.

Children without friends (M = 24.54, SD = 21.67) were spending significantly more time in parallel play than children with friends (M = 7.86, SD = 11.92), p = 0.05, whereas children with friends (M = 18.57, SD = 18.17) were spending significantly more time in parallel aware, p = 0.018, and joint engagement, p = 0.001, than children without friends.

Adult strategies outcomes

There were no significant differences between the types of strategies the adults used between the two groups of children (children with friends vs children without friends) (see Graph 2). The two strategies most frequently utilized by teachers to engage children with autism who did and did not have friends were behavior regulation skills (M = 19.68, SD = 15.72) and environmental arrangement (M = 11.58, SD = 12.42), whereas the least used strategies were recruiting other children to engage the child with autism (M = 0.61, SD = 1.12) and prompting the child with autism to engage with other children in the class (M = 0.87, SD = 1.82) (See Table 3).

Graph 2.

Observation of free play in the classroom: frequency of adult strategies.

Table 3.

Descriptives of number of adult strategies used in classroom.

	Minimum	Maximum	Mean	SD
Behavioral regulation	1	85	19.68	15.721
Joint attention	0	13	4.06	3.838
Environment	0	30	11.58	12.417
Prompt target	0	7	0.87	1.821
Recruit other children	0	4	0.61	1.116
50% Time with adult	0	29	3.10	5.816

SD: standard deviation.

The rate of the number of strategies being used per minute by the teachers was also examined. There were no significant differences between the two friend groups. The teachers on average are using 0.66 behavioral regulation per minute. This strategy is much more used than the other five strategies that were examined. The second most frequent strategy was environmental arrangement, and the rate of use was 0.37 per minute, which was half as frequent as behavioral regulation.

Because teachers rarely used any of the strategies, except for behavioral regulation, all strategies were dichotomized into high low binary measures for easier interpretation. Behavioral regulation was dichotomized using a median (m = 18) split, while the other five teacher strategies (joint attention, environment, prompt target, recruit other children, and 50% time with adult) were dichotomized into not used or used at least once due to the limited use of these strategies. Teachers who prompted the child at least once had children who spent significantly more time in parallel aware engagement state compared to teachers who did not prompt the child to play with their peers (p = 0.015), but no differences in time spent in joint engagement state. There were no significant differences in the time children spent in parallel aware or joint engagement between teachers who used at least one joint attention, environment, recruit other children, or 50% time with adult and teachers who did not use the strategies at all. There were also no significant differences in the time children spent in parallel aware or joint engagement between teachers who used more behavioral regulations compared to teachers who used less.

Subgroup analysis

Friendship quality outcomes

Both children with and without friends (based on Howes’ definition using the behavioral observation) were rated on their best friendship quality by parents and teachers. Nine sets of parents and teachers identified one common friend (out of the three top friends they were asked to identify) and only five sets reported the same best friend. However, about 50% of both parents and teachers were able to rate the quality of the identified best friendships. Of the parents (n = 17) and teachers (n = 12) who were able to rate the best friendship quality, there were no significant differences between the two groups of children (with or without friends). In addition, out of the 19-item friendship quality questionnaire, parents were, on average, only able to rate 7 of the items to be either positive or negative. On average, 12 of the items were rated as “Haven’t seen.” Similarly, teachers were only able to rate 10 of the items on the questionnaire as either positive or negative with the rest of the items rated as “Haven’t seen.” Examples of behaviors that were not seen include ignore each other’s suggestion, settle arguments peacefully, and say they like each other or are friends.

Parents also identified other settings besides school as sources for playmates. Other settings include family friends, outside activities, and neighborhood. The most identified setting for a “friend” was through family friends (n = 13), followed by outside activities (n = 6) and the neighborhood (n = 6).

Discussion

The primary goal of the study was to determine the extent to which preschool children with autism develop friendships in mainstream school settings. Despite the efforts of mainstream classrooms to socially integrate children with autism (Boutot and Bryant, 2005), the study revealed that only a fifth of the children from the sample had friends (as defined by Howes’ definition for typical children) in the classrooms, which is similar to the percentage of children with friends in elementary school (Kasari et al., 2011). This finding regarding friendships is in line with the literature for children with other developmental disabilities (Guralnick et al., 2007) where friendship formations have been a challenge.

Children’s joint attention skills were associated with whether children had a friend in the class. Compared to children without friends, children who had friends were using higher level joint attention skills such as pointing and showing. In this study, nearly all children with friends (6/7) used high-level joint attention skills, and only half (12/24) of the children without friends used high-level joint attention skills. There were no differences in cognitive and language levels, highlighting the importance of social communication skills. These qualitative differences appear important to young children who are more likely to notice and respond to a clear, directed gesture like a show or point. These results should be interpreted with caution despite their clinical relevance because of the small sample size. This association warrants further study.

Another distinct difference between the two groups is how the children spend their time during free play. Children with friends spent significantly more time in parallel aware and joint engagement (i.e. playing together with an object), whereas children without friends spent more of their time in parallel play. Children with friends may be more aware of their environment and their proximity to other children, thus increasing the likelihood of engaging other children in play and making new friends.

Parents and teachers seem to have a more positive perception of the children’s friendships when compared to blind observers. Twice as many parents and teachers than blind observers identified a “friend” in the target child’s class. Although parents and teachers identified these “friends,” they had difficulty rating the friendship quality between the friend and the target child. For example, when asked to rate the quality of specific social interactions between the dyad such as whether the pair shared toys or argued verbally, many parents and teachers reported that they had never seen these behaviors to rate them. Furthermore, parents and teachers did not necessarily identify the same “friend” at school for the same child even when they did report a “friend” for the target child. These findings raise issues as to the “reliability” of parent and teacher report of friendships, or alternatively suggest that our definitions of friendships may need to be reconsidered.

Finally, the study examined six different teacher strategies that were used to support friendships in the classroom. The most common strategy was behavioral regulation. Despite the fact that many children in the sample spent the majority of the time in solitary play, very few adults in the classrooms tried to intervene and have these children engage with their peers. Consequently, children spent a majority of their free play wandering around the classroom or playing by themselves. The adults usually intervened only when the child was behaving inappropriately (e.g. throwing toys) or when it was time to clean up. This may be due to the teacher-to-student ratio in the classrooms (3:14). Thus, adults in the classroom were mainly engaging the children for classroom management purposes and were rarely interacting with the children to promote social interaction with their peers. However, of the few instances where the adults intervened and prompted the child to play with their peers, data showed promising results. The children were more aware of each other, which may be the first step in developing meaningful friendships. This suggests that adults may not recognize the importance of free play as a time to develop peer relationships.

The findings from this study suggest that children may need more support from both parents and teachers to foster friendships with their peers at school. Even though these children may be considered “high functioning” and “mainstreamed,” there are already social skill deficits that are apparent in school settings. Preschool aged children with ASD may need early interventions that specifically target core deficits such as joint attention and communication skills to help foster meaningful relationships (Kasari et al., 2010; Lawton and Kasari, 2012).

Additionally, with the increase prevalence of autism (Centers for Disease Control and Prevention (CDC), 2012), many teachers will face the possibility of receiving more children with autism in their classrooms. Parents are finding inclusion as a better alternative for their children with autism (De Boer et al., 2010), particularly if their child is high functioning. However, teachers may not feel well equipped to work with this population (Robertson et al., 2003). Therefore, additional training and specific strategies may be needed for both special education and general education teachers to better meet the needs of young children with autism. In addition to standard behavioral management trainings, specific strategies to facilitate and foster friendship between children with autism and their peers should be included. For example, peer-mediated interventions can be implemented in these classrooms to increase social initiations and interactions of the children in the classroom (Chan et al., 2009, for review) since one of the key benefits of an inclusion classroom is the availability of neurotypical peer models (Mesibov and Shea, 1996).

Parents should also be more involved in their children’s friendship development at this early age. As previously mentioned, one of the findings from this study indicated that very few parents and teachers identified the same friend for a target child. Better communication between teachers and parents would help foster friendships between young children with autism and their peers. For example, if a teacher recognizes a potential friendship developing between a child with autism and a peer in class, the teacher could communicate this to the parents so that they can arrange outside activities (i.e. playdates) to further develop the friendship.

Strength of study

This study is one of the very first studies to explore friendships in children with autism in preschool settings (see review by Petrina et al., 2014). Although only a fraction of the sample met criteria for having a “friend,” this small sample size provided valuable information about the challenges in friendship formation for this population. In addition, some parents were unable to identify a peer at school that their child played with but were able to identify playmates outside of school including family and neighborhood children. Future studies examining friendships in preschool children with autism may want to expand the research by investigating friendships in different settings, including both school and home.

Limitation of study

A limitation of the study was its moderate sample size of 31 preschool children. There was minimal variability to capture the differences between the two groups of children (children who had friends vs child who did not). By Howes’ definition of friendship, only seven of the children in this study had friends. This small sample makes it difficult to generalize the findings. In the future, a larger sample of children should be assessed to further examine the characteristics of preschool children who are developing friendships. Additionally, having a comparison group of typically developing children may provide important benchmarks of friendship development within a preschool setting. Finally, this study conducted only two observations within a month. Future studies may want to conduct more observations throughout the year to capture the varying dynamics of friendship formation at different times of the school year (e.g. beginning vs end of the school year).

Footnotes

Appendix 1

Friendship observation coding key.

State/levels (the state displayed by the target child for the majority of the 50-s observation time)
Solitary (S)	Child plays alone, with no peers within 3 ft, and no mutual eye gaze with other children.
Proximity (X)	Child plays alone within 3-ft range of peer.
Onlooker (O)	Child has one-way awareness of child who is farther than 3 ft away.
Parallel (P)	Child and peer are engaged in a similar activity but there is no social behavior.
Parallel Aware (PA)	Child and peer engaged in similar activity and mutually aware of each other during activity (e.g. glances at the other child but no direct social behavior).
Joint engagement (JE)	Child and peer direct social behavior (e.g. offering objects, conversing, toy-taking, and other activities with a turn-taking structure).
Discrete behaviors
Social initiations (\|, +, −)	Child directs communication to peer—for example, offers toy, greets, asks to play game.
Response to social initiation (+, −)	Child responds to overture of peer with gesture, or language—response coded as accept or reject.
Shared affect (l, +, −)	Child and peer smile/laugh while looking at each other or sharing same activity or child and peer engage in a negative encounter (i.e. arguing, fighting).
Adult strategies
JA skills	Adult initiates a joint attention act (e.g. coordinated joint look, pointing, showing for the purpose of sharing, give to share).
BR skills	Adult initiates a behavior regulation act, or requesting, (e.g. pointing, giving for the purpose of requesting).
Environment	Adult arranges the environment to facilitate play (e.g. appropriate toy choice, creating play space).
Prompt target	Adult prompts the child to go play with his or her peers.
Recruitment of other children	Adult calls other children over to play with target child.
50% Time with adult	Adult allows target child to play with only him or her for more than 50% of the observation time interval.

JA: joint attention; BR: behavior regulation.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Alberto

Troutman

(2006) Applied Behavior Analysis for Teachers. 7th ed. Upper Saddle River, NJ: Merrill and Prentice Hall.

Anderson

Moore

Godfrey

. (2004) Social skills assessment of children with autism in free-play situations. Autism 8(4): 369–385.

Bauminger-Zviely

Agam-Ben-Artzi

(2014) Young friendship in HFASD and typical development: friend versus non-friend comparisons. Journal of Autism and Developmental Disorders 44(7): 1733–1748.

Bauminger-Zviely

Karin

Kimhi

. (2014) Spontaneous peer conversation in preschoolers with high-functioning autism spectrum disorder versus typical development. Journal of Child Psychology and Psychiatry 55(4): 363–373.

Boutot

Bryant

(2005) Social integration of students with autism in mainstream settings. Education and Training in Developmental Disabilities 40(1): 14–23.

Brown

Odom

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

Buysse

Goldman

West

. (2008) Friendships in early childhood: implications for early education and intervention. In: Brown

Odom

McConnell

(eds) Social Competence of Young Children: Risk, Disability, and Intervention. Baltimore, MD: Paul H. Brookes, pp.77–97.

Centers for Disease Control and Prevention (CDC) (2012) MMWR weekly: prevalence of autism spectrum disorders—Autism and Developmental Disabilities Monitoring Network, 14 Sites, United States, 2008. Available at: http://www.cdc.gov/mmwr/pdf/ss/ss6103.pdf

Chamberlain

Kasari

Rotheram-Fuller

(2007) Involvement or isolation? The social networks of children with autism in regular classrooms. Journal of Autism and Developmental Disorders 37: 230–242.

10.

Chan

Lang

Rispoli

. (2009) Use of peer-mediated interventions in the treatment of autism spectrum disorders: a systematic review. Research in Autism Spectrum Disorders 3: 876–889.

11.

De Boer

Pijl

Minnaert

(2010) Attitudes of parents towards mainstream education: a review of literature. European Journal of Special Needs Education 25(2): 165–181.

12.

Dunn

(1961) Multiple comparison among means. Journal of the American Statistical Association 56: 52–64.

13.

Fujisawa

Kutsukake

Hasegawa

(2008) Reciprocity of prosocial behavior in Japanese preschool children. International Journal of Behavioral Development 32(2): 89–97.

14.

Goldman

Buysse

(2005) Playmates and Friends Questionnaire for Teachers, Revised. Chapel Hill, NC: The University of North Carolina, FPG Child Development Institute.

15.

Gottman

(1983) How children become friends. Monographs of the Society for Research in Child Development 48(3): Serial No. 201.

16.

Guralnick

Groom

(1987) The peer relations of mildly delayed and nonhandicapped preschool children in mainstream playgroups. Child Development 58: 1556–1572.

17.

Guralnick

Neville

Hammond

. (2007) The friendships of young children with developmental delays: a longitudinal analysis. Journal of Applied Developmental Psychology 28: 64–79.

18.

Hollingsworth

Buysse

(2009) Establishing friendships in early childhood mainstream settings: what roles do parents and teachers play? Journal of Early Intervention 31: 287–307.

19.

Howes

(1983) Patterns of friendship. Child Development 54(4): 1041–1053.

20.

Howes

(1988) Peer interaction of young children. Monographs of the Society for Research in Child Development 53(1): Serial No. 217.

21.

Howlin

Moss

Savage

. (2013) Social outcomes in mid- to later adulthood among individuals diagnosed with autism and average nonverbal IQ as children. Journal of the American Academy of Child & Adolescent Psychiatry 52(6): 572–581.

22.

Kasari

Gulsrud

Wong

. (2010) Randomized controlled caregiver mediated joint engagement intervention for toddlers with autism. Journal of Autism and Developmental Disorders 40: 1045–1056.

23.

Kasari

Locke

Gulsrud

. (2011) Social networks and friendships at school: comparing children with and without autism. Journal of Autism and Developmental Disorders 41: 533–544.

24.

Kasari

Rotheram-Fuller Locke

(Unpublished) The development of the playground observation of peer engagement (POPE) Measure. Los Angeles, CA: University of California Los Angeles.

25.

Kasari

Rotheram-Fuller

Locke

. (2012) Making the connection: randomized controlled trial of social skills at school with children with autism spectrum disorder. Journal of Child Psychology and Psychiatry 53(4): 431–439.

26.

Kimhi

Bauminger-Zviely

(2012) Collaborative problem solving in young typical development and HFASD. Journal of Autism and Developmental Disorders 42: 1984–1997.

27.

Kishida

Kemp

(2009) The engagement and interaction of children with autism spectrum disorder in segregated and mainstream early childhood center-based settings. Topics in Early Childhood Special Education 29: 105–118.

28.

Lawton

Kasari

(2012) Brief report: teacher-implemented joint attention intervention: pilot randomized controlled study for preschoolers with autism. Journal of Counseling and Clinical Psychology 30: 687–693.

29.

Lord

Rutter

DiLavore

. (2001) Autism Diagnostic Observation Schedule Manual. Los Angeles, CA: Western Psychological Services.

30.

Loveland

Landry

(1986) Joint attention and language in autism and developmental language delay. Journal of Autism and Developmental Disorders 16(3): 335–349.

31.

Mesibov

Shea

(1996) Full inclusion and students with autism. Journal of Autism and Developmental Disorders 26(3): 337–346.

32.

Mullen

(1989) Mullen Scales of Early Learning. Cranston, RI: T.O.T.A.L. Child, Inc.

33.

Mundy

Sigman

Ungerer

. (1986) Defining the social deficits of autism. The contribution of nonverbal communication measures. Journal of Psychology and Psychiatry 27: 657–669.

34.

Mundy

Sigman

Ungerer

. (1987) Play and nonverbal communication correlates of language development in autistic children. Journal of Autism and Developmental Disorders 17: 349–363.

35.

Odom

Zercher

. (2006) Social acceptance and rejection of preschool children with disabilities: a mixed-method analysis. Journal of Educational Psychology 98(4): 807–823.

36.

Petrina

Carter

Stephenson

(2014) The nature of friendship in children with autism spectrum disorders: a systematic review. Research in Autism Spectrum Disorders 8: 111–126.

37.

Robertson

Chamberlain

Kasari

(2003) General education teachers’ relationships with included students with autism. Journal of Autism and Developmental Disorders 33(2): 123–130.

38.

Rotheram-Fuller

Kasari

Chamberlain

. (2010) Grade related changes in the social inclusion of children with autism in general education classrooms. Journal of Child Psychology and Psychiatry 51(11): 1227–1234.

39.

Sebanc

(2003) The friendship features of preschool children: links with prosocial behavior and aggression. Social Development 12(2): 249–268.

40.

Seibert

Hogan

Mundy

(1982) Assessing interactional competencies: the Early Social-Communication Scales. Infant Mental Health Journal 3: 244–259.

41.

Sigman

Ruskin

(1999) Continuity and change in the social competence of children with autism, Down syndrome, and developmental delays. In: Monographs of the Society for Research in Child Development (Serial No. 256, Vol. 64, Issue 1). Chicago, IL: University of Chicago Press.

42.

Sigman

Mundy

Sherman

. (1986) Social interactions of autistic, mentally retarded, and normal children and their caregivers. Journal of Psychology and Psychiatry 27: 647–656.

43.

Travis

Sigman

Ruskin

(2001) Links between social understanding and social behavior in verbally able children with autism. Journal of Autism and Developmental Disorders 31(2): 119–130.

44.

Vandell

Mueller (1980) Peer play and friendships during the first two years. In: Foot

Chapman

Smith

(eds) Friendships and Social Relationships in Children. New York: John Wiley, pp.181–208.

45.

Vaughn

Colvin

Azria

. (2001) Dyadic analyses of friendship in a sample of preschool-age children attending Head Start: correspondence between measures and implications for social competence. Child Development 72(3): 862–878.