Educators’ engagement with children with autism spectrum disorder in a learning environment with multiple technologies in Finland and China

ABSTRACT

Technology-enhanced environments are increasingly used to enhance the interaction skills of children with autism spectrum disorder (ASD), yet little is known about how technology use influences interactions between educators and children. This article examined how educators engage with children with ASD in a technology-enhanced learning environment (LE). More precisely, the article investigated how educators verbally guide children with ASD during activities at four different technology-based workstations and how the children respond verbally and behaviorally to educators’ guidance. Seven children with ASD and their educators participated in this exploratory study. Data consisted of video recordings collected during interactional situations in technology-enhanced LEs with multiple technologies in Finland and China. The recordings were analyzed by developing a coding scheme in order to capture the educators’ verbal guidance and the children’s verbal and behavioral responses. The study showed that the educators engaged with children with ASD differently at different workstations, and the technology used shaped the educators’ verbal guidance and the children’s verbal and behavioral responses. These differences in educators’ verbal guidance and children’s responses between the workstations showed how the design and features of technology-enhanced solutions affected interactions around their use. Thus, by using versatile technologies, it was possible to shape the interaction between educators and children during educational activities and to provide children with ASD a variety of opportunities to practice communication and interaction skills.

KEYWORDS:

• Autism spectrum disorder
• technology-enhanced learning environment
• social interaction
• communication
• learning environment

Introduction

Difficulties in social communication and interaction are commonly cited features of persons with autism spectrum disorder (ASD) (American Psychiatric Association, 2013; Arciuli & Brock, 2014). Consequently, teaching social skills has been emphasized in the education of children with ASD (Boyd, Conroy, Asmus, McKenney, & Mancil, 2008; Camargo et al., 2014). Numerous interventions and pedagogical approaches have been invented to support the acquisition of social communication and interaction skills at school (Alexander, Ayres, & Smith, 2014; Grow & LeBlanc, 2013; Kossyvaki, Jones, & Guldberg, 2016; Lindsay, Proulx, Scott, & Thomson, 2014; Waugh, Alberto, & Fredrick, 2011). According to Kossyvaki et al. (2016), developmental or relationship-based approaches typically stress the importance of adult interaction style in enhancing children’s social communication skills. For example, the adults are encouraged to respond to all the attempts by the children to communicate. On the other hand, behavioral or naturalistic approaches aim at teaching a child a specific skill, and the role of the adult is to provide necessary guidance (Schreibman et al., 2015).

In addition to interventions and pedagogical strategies, accommodating the physical learning environment (LE) (Martin, 2016; McAllister & Maguire, 2012; Smith & Brown, 2000) and utilizing various technology-enhanced solutions (Benitti, 2012; DiGennaro Reed, Hyman, & Hirst, 2011; Schmidt & Schmidt, 2008; Tetreault & Lerman, 2010; Wojciechowski & Al-Musawi, 2017) have been found to support the acquisition of social skills of children with ASD (Cormier & Natale, 2014; Ramdoss et al., 2012; Virnes, Kärnä, & Vellonen, 2015). For example, computers, laptops, smartphones, and various applications (Battocchi et al., 2009; Bauminger et al., 2007; De Leo & Leroy, 2008 DiGennaro Reed et al., 2011; Fletcher-Watson et al., 2016; Jacklin & Farr, 2005; Piper, O’Brien, Morris, & Winograd, 2006) as well as robots (Benitti, 2012; De Silva, Tadano, Saito, Lambacher, & Higashi, 2009; Kozima, Nakagawa, & Yasuda, 2005; Robins, Dautenhahn, & Dickerson, 2009) have been found to enhance the interaction skills of children with ASD. Moreover, virtual LEs and androids (Foster et al., 2010; Pioggia et al., 2005; Schmidt & Schmidt, 2008) as well as video modeling (Gul & Vuran, 2010; Tetreault & Lerman, 2010) have been used by children with ASD to practice interaction skills. Overall, previous studies have shown that there is enormous potential in technology as a means to enhance the interaction skills of children with ASD. Yet, more research is needed to show their efficacy (Den Brok & Sterkenburg, 2015; Ploog, Scharf, Nelson, & Brooks, 2013).

Most previous research on the use of technology with children with ASD has typically been carried out by using a single technology such as a tabletop computer (Battocchi et al., 2009), a mobile phone (De Leo & Leroy, 2008) or a tablet (Fletcher-Watson et al., 2016). In addition, as technologies have often been designed to be used by children themselves, research has focused on the interaction between children and technologies (Odom et al., 2015). However, very little is known about the impact of multiple technologies on teaching interaction skills when they are used side by side in a learning environment, and the role of a teacher in such learning situations. According to Bunning, Heath, and Minnion (2010), the role of the person providing support has been found to play a critical role in computer-based activities performed by children with ASD (see also Bunning, Kwiatkowska, & Weldin, 2012). Jacklin and Farr (2005) have also noted how important it is to consider social interactions around the technology used, rather than investigating the use of technology in isolation, focusing merely on the user of the technology. In their study involving children with ASD, they found that technology use without social interaction is at risk of resulting in an obsessive use of technology, whereas following the child’s lead results in positive and sustained interaction. In another study involving children with intellectual disabilities, educators were found to dominate the interactions around technology use (Bunning et al., 2010). In these interactions, the educators mainly used verbal requests with some gestures towards the technology.

These findings demonstrate how technology use always takes place in a specific social context: Bunning et al. (2010, p. 61) note that children’s engagement and work with technology is inevitably “inseparable from the communication context determined by the type of linguistic support given by the teacher.” Both of these studies highlight the importance of considering the role of human mediation in technology use and encourage further examination of technology as part of social interaction. As claimed by Jacklin and Farr (2005) and Bunning et al. (2010), children’s engagement with technology, and interaction and communication around technology use, crucially depend on the mediating role of the adult co-participants. As these two studies focus on single technologies rather than multiple solutions, it remains unknown how the features of different technology applications might affect adult co-participants’ social interaction and communication with children with ASD, and in turn, how the children respond to their co-participants in these different contexts. It is this gap in the literature that this exploratory study addresses.

The goal of the study was to investigate how educators engage with children with ASD at four different technology-based workstations in a technology-enhanced LE with multiple technologies. The LE consisted of four different technology-based workstations. Each station had a computer and specifically designed software to enhance children’s skills such as storytelling, visual perception, and motor skills. The contents of the software were modifiable according to the needs and preferences of the children. In addition, the LE was designed to support children’s active role, creativity and strengths. The purpose was to examine whether the guidance that the educators provide to the children, and how the children respond to it, is influenced by the four different technology applications in the LE. As teacher-directed activities are commonly used methods with children with ASD (Cadette, Wilson, Brady, Dukes, & Bennett, 2016; Flores & Ganz, 2009) in both Finnish and Chinese schools (Hu & Fan, 2016; Lerkkanen, 2013; Sahlberg, 2011), the study focused on interactional situations in which the educators verbally guided the activities of the children with ASD in a technology-enhanced LE with multiple technologies. As no existing study with similar design was found, the exploratory study presented in this article lays a basis for future studies. It is known that digitalization (McKnight et al., 2016) and the design and development of new technologies for children with ASD continue. It is also known that technologies can be used to support the learning of interaction skills of children with ASD. However, more research is needed on the triadic interaction between an educator, child and technology, and the impact of different technologies on it in order to take the best advantage of technologies in teaching children with ASD. The purpose of this study was to explore the triadic interaction in using technology to teach children with ASD.

Methods

Research questions

To explore how educators engage with children with ASD in a technology-enhanced LE with four workstations, we addressed the following research questions: 1. What type of verbal guidance educators use at the four workstations when interacting with children with ASD? 2. What type of responses children with ASD produce to answer educators’ verbal guidance at the four workstations? 3. What is the relationship between the different types of educators’ guidance and children’s responses? 4. Are there any differences in educators’ and children’s interactions while working at the four workstations? 5. Are there any differences between the Finnish and Chinese samples in children’s and educators’ interaction while working at the four workstations?

Research design

The study was exploratory by nature. An exploratory study aims at making sense of some social phenomenon rather than testing a hypothesis (Russell, 2009). In addition, exploratory research aims at scrutinizing new problems on which little or no previous research exists (Singh, 2007). As hardly any research existed on the impact of educators’ actions or environmental factors (i.e. the role of multiple technologies) on the process of the interaction between educators and children with ASD (Bunning et al., 2012; Jacklin & Farr, 2005; Tuononen, Kiiskinen, & Kärnä, 2014), it was significant to approach this scarcely investigated phenomenon from an exploratory perspective. By applying an exploratory design, the study provides novel insight into triadic interaction between educators, children, and technology. Moreover, the selected approach provides knowledge about a rarely studied topic how different technologies in a LE shape the interaction between the educators and children. This is a particularly important issue in relation to the education of children with ASD. By scrutinizing the triadic interaction, it is possible to reveal how technologies can be designed and used more efficiently to support the acquisition of interaction skills of children with ASD. The study was part of a larger study that investigated the interaction, communication, and creativity of children with ASD in a technology enhanced LE.

Participants

The participants included seven children (three from Finland, four from China), and seven educators (three from Finland, four from China). All the children were boys, aged six to 12. See Table 1 for information on the participating children.

Table 1. Information on the participating children.

	Child 1	Child 2	Child 3	Child 4	Child 5	Child 6	Child 7
Country	Finland	Finland	Finland	China	China	China	China
Age	6	8	11	7	8	9	12
Diagnosis^a	Moderate intellectual disability (F71.0), Childhood autism (F84.0)	Childhood autism (F84.0)	Unspecified intellectual disability (F79.0), Childhood autism (F84.0)	ASD (CARS-2 score: 43)	ASD (CARS-2 score: 45)	ASD (CARS-2 score: 52)	ASD (CARS-2 score: 49)
Characteristics	Delayed but sentence-length speech. Difficulties in understanding instructions. Perceptional, attentional, and motor difficulties. Engages in eye contact and interactions with others.	Delayed and occasionally repetitive speech. New situations and people difficult. Scarce eye contact. Easily distracted. Understands clear sentences and benefits from support and clear instructions.	Responsive but not initiative. Delayed and occasionally repetitive speech. Benefits from clear instructions and alternative communication methods. Needs prompting to act. Challenges in motor coordination.	Delayed repetitive speech. Restricted interest in balls. Needs prompting to act. Challenges in motor coordination.	Delayed expressive speech and occasionally repetitive speech. Restricted interest in balls. Needs prompting to act. Challenges in motor coordination.	Delayed repetitive speech. Needs prompting to act. Challenges in motor coordination, fine motor skills and gross motor skills.	Delayed receptive speech and no expressive speech. Needs prompting to act. Challenges in motor coordination.

^aThe Finnish diagnoses are according to the ICD-10 criteria.

The Finnish educators were all women. These educators had all formal education in working as a special needs assistant and their work experience varied from seven years to over 20 years of experience. The Chinese special educators included two women and one man. They had more than ten years of experience of working with children with ASD.

Setting

The LE consisted of four technology-based workstations. The pedagogical aspects of the technology-based workstations were carefully considered. The applications of the workstations were designed, for example, to support the children’s communication and the use of visual supports (Guldberg, 2010; Parsons et al., 2011), to support the children in using various senses (Bogdashina, 2003; Frith, 2003), to provide structure for activities, to foster children to make choices (Sze, 2009), and to support the emergence of the children’s strengths and creativity instead of focusing on difficulties (Happé & Frith, 2009). There were four technology-based workstations in the LE: symbol-matching, Lego-building, storytelling, and Kinect-playing (See Figure 1).

Figure 1. Symbol-matching, Lego-building, storytelling, and Kinect-playing.

At the symbol-matching workstation, the children were tasked with matching (e.g. the symbol for a cat and the sound of a cat: “meow,” matching pictures with similar objects, or matching a certain number of objects to corresponding numbers). The children chose the topic for the tasks and selected matching symbols with a computer mouse. Symbol-matching included tasks with different themes such as animals, emotions, shapes, vehicles. The matching tasks in China involved cultural adaptation of task content, for instance, the cartoon pictures and food items were adapted to the Chinese cultural background. Other symbol-matching tasks included colors, numbers, fruits, and stationery. The station was designed to increase the children’s vocabulary and conceptual knowledge as well as to strengthen their choice-making and matching skills.

At the Lego-building workstation, the children built a Lego Duplo or basic Lego construction from a model on the computer application. The children chose a task from two alternatives: 1) building from the picture of the whole model, or 2) step-by-step building of the model. In addition, the children chose between building a model according to a certain character (e.g. various animals) or a random, abstract model. The children adjusted the level of difficulty by changing the number of bricks used in the model in the application. The goal of the activities at the station was to enhance the children’s visual attention and matching skills.

At the storytelling workstation, the children created a story by using a picture-based computer application and a touch screen. Hand-drawn ready-made pictures with written one-word descriptions were categorized as people, creatures, places, and actions and were presented as visible main categories and subcategories. The children could also draw pictures of their own, save them in their own folder, and add them to stories. The children created stories by using a drag-and-drop function that allowed the pictures to be placed on a storyline. In addition, it was possible for the children to write the name of the story above the storyline, write text under the pictures, and record the story to be listened to later. The stories were saved to the story library, where the children could share their stories with others. In the story library, the children could review and listen to their own stories or review and listen to the stories created by other children. Finally, the children could also print out their stories and put together their own storybooks. The goals of the activities at this station were to enhance the children’s conceptual knowledge, sequence building, narrating, and creative thinking.

At the Kinect-playing workstation, the children played a catching game using Microsoft’s Xbox Kinect. The children had to catch objects (e.g. fish, birds, letters) by using their whole body, including their legs and hands. Playing the game was easy as it allowed a variety of movements as long as the player stayed within the play area. As the objects as well as the background of the game were changeable, the children could choose, for example, their favorite object from the game menu and play the game by catching the selected object. In addition, pictures drawn by the children could be used as objects or as the background of the game. The goals of the activities at this station were to enhance the children’s visual attention skills and their movement coordination.

The applications of the workstations were created in Finland. However, the language used in the applications at each workstation was translated into Chinese, and, thus, the same applications were used in China and Finland. As a user could change the content of the applications at the Kinect, storytelling, and symbol-matching workstations, these applications were modified according to the needs of the children participating in both countries. Researchers explained and showed the educators and children how to use the technologies, but they were allowed to be creative in their use and to interact as freely as they wished. In addition, the educators were asked to interact with the children the same way as they did in the classroom. The stations were designed to support the children’s independent work. However, as the children had multiple needs, the educators were advised to guide the children whenever needed. Thus, the goal of this arrangement was to keep the relationship between the children and the educators as natural as possible in order to investigate how the technologies used at the different stations might influence the interaction between educators and children.

The research project ran one-hour group sessions, called action groups, weekly, nine times each semester in Finland and China. At the beginning of each session, there was a short warm-up with greetings, and the researchers gave the children a pictured map of the workstations. Although the order of the workstations was predetermined, the children could choose from a variety of tasks or games to work with at each workstation. The children worked individually at each station for 10–15 min, and the educators were advised to help if needed (e.g. by setting the level of difficulty for a task or giving guidance when needed). The order of the workstations varied for each child every session.

In Finland, the technology-enhanced LE was set up in a spacious room (approximately 15 meters by six meters, and the Kinect room was approximately five meters by six meters) in the school building. The participating children and educators were able to see each other taking part in the activities in the LE. In China, the LE was set up in two adjacent rooms due to a lack of space in one room. The Lego-building, storytelling, and symbol-matching workstations were set up in the first room, and the Kinect workstation was set up in the other one. This arrangement guaranteed that children and educators had enough space to work in the LE and yet be able to interact with each other when necessary. Transitions from one room to another were easy as the rooms were located next to each other.

Data collection

Data included video-recorded interactions between children with ASD and their educators. These interactions took place during technology-enhanced action groups at two schools in Finland and China in which the educator-child pairs worked at the four different stations. Two digital video cameras were used at each workstation for comprehensive recording of the educators’ and children’s activities. The total amount of Finnish data analyzed was 13 h 45 min, and the Chinese data was 14 h 30 min.

Data analysis

A coding scheme was developed to analyze the video recordings. The scheme was first developed qualitatively by examining the video data for interesting interactional and communicational phenomena. Our interest was then narrowed down to the instructional interactions between the educators and children. The coding was not intended to be exhaustive as we were particularly interested in a few specific instructing actions of the educators and the children’s responses to these. First, we coded the frequency of educators’ initiating actions as 1) direct verbal instructions or 2) questions used to guide the children. Second, we coded the frequency of the children’s responsive actions as 1) verbal responses and 2) behavioral responses. These codes were interconnected, that is, the children’s responsive actions were coded solely in relation to the educators’ prompts. The coding thus began by identifying the educators’ initiating actions, which was followed by the identification of children’s possible responses to these. See Appendix for coding definitions. The coding was carried out by Master’s students in Special Education and supervised by the research team experienced in video analysis. 10% of both the Finnish and Chinese data were coded twice by two students for the purpose of triangulation. Potential bias during coding was controlled for by calculating inter-rater reliability. It was calculated separately for the Finnish and Chinese data using the intra-class correlation coefficient. Reliability was good for all the variables in both samples (see Table 2).

Table 2. Inter-rater reliability in the Finnish and Chinese samples.

	Finnish sample	Chinese sample
Educators’ direct instructions	.990	.994
Educators’ questions	.987	.995
Children’s verbal responses	.854	.996
Children’s behavioral responses	.980	.993

Note: In the Finnish sample, inter-rater reliability was calculated based on the data at the symbol-matching station due to limited availability of coders to code the entire data set including all the four workstations. The Chinese sample included all the four workstations.

Different methods were used to answer the research questions. To answer the first and second research questions, we drew on total and average frequencies of the educators’ guidance and children’s responses at the four workstations. To address the third question, we utilized Spearman’s rho to examine the correlations between the educators’ instructions and the children’s responses. To answer the fourth question, we compared the frequency of the educators’ initiating actions and the children’s responsive actions at each workstation using the Friedman test and the Wilcoxon Signed Ranks test for pair-wise comparisons. Finally, for the fifth research question, we used the Mann–Whitney U test. We did not apply Bonferroni correction due to the small sample size.

Results

Verbal guidance of educators at the four workstations

The first research question focused on the verbal guidance of educators at the four workstations. The average frequencies of the educators’ direct instructions and questions (see Table 3) per session show that both instruction types occurred at all four workstations. However, the frequency of instructions varied greatly between the children (e.g. compare the number of questions that Child 2 and Child 6 received on average at the storytelling station) and between the stations (e.g. compare the number of questions that Child 6 received at the symbol-matching station and the Kinect-playing station). The total frequencies (in Table 3) also suggest that question-type instructions were observed in higher frequency at the storytelling and symbol-matching stations, whereas direct instructions were more frequent at the Lego-building and Kinect-playing stations.

Table 3. Average frequencies of educators’ direct instructions and questions for the children at the four workstations.

	Storytelling		Symbol-matching		Lego-building		Kinect-playing
	DI	Q	DI	Q	DI	Q	DI	Q
Child 1	32.57	6.29	19.57	14.43	29.71	9.14	9	0.14
Child 2	6.13	5.13	9.5	9.5	16	1.63	6.5	0.13
Child 3	21.38	6.38	6.75	7	39.38	11.63	9.38	1.25
Child 4	15.7	26.8	16.8	73.5	28.5	24.6	24.8	2.9
Child 5	20.22	57.89	6.44	67.22	31.11	46.11	36.89	4.78
Child 6	11.11	59.67	6.11	25.78	14.89	28.22	22.11	3.22
Child 7	11.88	40.13	8.88	14.75	13.63	36.63	21.38	4.5
Total	118.99	202.29	74.05	212.18	173.22	157.96	130.06	16.92

Note: DI = Direct Instructions. Q = Questions. Data for Child 4 consists of 10 sessions, for Child 5 and Child 6 9 sessions, for Child 2, Child 3, and Child 7 8 sessions, and for Child 1 7 sessions.

Responses of children to educators’ verbal guidance at the four workstations

The second research question targeted the children’s responses to the educators’ verbal guidance. The frequency counts of the children’s verbal and behavioral responses (see Table 4) showed that the children used both two response types at all four workstations. Their responsiveness varied, however, between the workstations. According to the total frequencies, verbal responses were more frequent at the storytelling station whereas behavioral responses were observed in higher frequency at the other three stations.

Table 4. Average frequencies of children’s verbal and behavioral responses to educators’ verbal instructions at the four workstations.

	Storytelling		Symbol-matching		Lego-building		Kinect-playing
	VR	BR	VR	BR	VR	BR	VR	BR
Child 1	3.71	29.29	9.57	27.29	7.14	33.43	1.57	25.43
Child 2	3.63	6.75	9.13	10	16.88	0.88	0	7.5
Child 3⁠	3.38	23.5	1.63	13.38	5	46.13	0	11.88
Child 4	21.2	10.7	50.3	67.1	20	67.1	2.1	18.5
Child 5	32.33	8.22	39.56	73.56	36	64.22	4	19.33
Child 6	51.33	4.89	26.33	115.22	12.67	115.33	1.89	16.67
Child 7	34.38	5.25	10.25	96.88	35.75	55.88	4.75	19
Total	149.96	88.6	146.77	403.43	133.44	382.97	14.31	118.31

Note: VR = Verbal response; BR = Behavioral response. Data for Child 4 consists of 10 sessions, for Child 5 and Child 6 9 sessions, for Child 2, Child 3 and Child 7 8 sessions, and for Child 1 7 sessions.

The relationship between the educators’ guidance and children’s responses

The third research question examined the relationship between the educators’ guidance and the children’s responses. The analysis showed that only the educators’ question frequency had a significant correlation with the children’s frequency of verbal responses (rs = .903, p < .01).

Differences in educators’ direct instructions between the workstations

The fourth research question addressed the statistical differences between the educators’ guidance and the children’s responses at the different workstations. The results showed that the educators used direct instructions more at the storytelling and the Lego-building stations than at the symbol-matching station. These differences were statistically significant between the storytelling station and the symbol-matching station (Z = 1.992, p = .046), and between the Lego-building station and the symbol-matching station (Z = 2.366, p = .018). Other stations did not differ from each other statistically.

Differences in educators’ questions between the workstations

The educators used questions to instruct the children most at the storytelling, Lego-building, and symbol-matching stations. There were more questions at the storytelling (Z = 2.366, p = .018), Lego-building (Z = 2.366, p = .018), and symbol-matching (Z = 2.366, p = .018) stations than at the Kinect-playing station. Other stations did not differ from each other statistically.

Differences in children’s verbal responses between the workstations

The results showed that there were more verbal responses from the children at the storytelling (Z = 2.366, p = .018), Lego-building (Z = 2.366, p = .018), and symbol-matching (Z = 2.366, p = .018) stations than at the Kinect-playing station. Other stations did not differ from each other statistically.

Differences in children’s behavioral responses between the workstations

The results showed that there were more behavioral responses from the children at the Lego-building station than at either the Kinect station (Z = 2.366, p = .018) or the storytelling station (Z = 2.197, p = .028). Other stations did not differ from each other statistically.

Differences between the Finnish and Chinese samples

The fifth research question examined the differences between the Finnish and Chinese samples as regards the educators’ guidance and the children’s responses. There were no statistical differences between the samples in educators’ direct instructions, educators’ questions, children’s verbal responses, or children’s behavioral responses at any of the workstations.

Discussion

This exploratory study focused on examining interactions between educators and children with ASD and how these interactions were shaped by different technology-based workstations in a LE. Previous studies have suggested that different technology-based activities prompt the children to act in different ways (Vellonen, Kärnä, & Virnes, 2013; see also Tuononen, Laitila, & Kärnä, 2014). However, as these studies have focused mostly on children rather than children and their interactional partners together, little is known about the interactional contexts in which the children act in LEs with multiple technologies. In other words, this study gives insight to rarely investigated triadic interaction between technology, educators, and children with ASD. Moreover, the findings of this study are unique as they demonstrate how different technologies shape educators’ and children’s interaction in a LE.

The current study shows that the educators tend to instruct children differently at different workstations. Direct instructions were more frequent at the storytelling and Lego-building stations than at the symbol-matching station. Considering the design of the workstations, it can be suggested that the children were able to work at the symbol-matching station by following the detailed instructions provided by the computer application. However, the storytelling and Lego-building stations provided the children with multiple options for how to proceed. It thus became apparent that the educators’ active involvement and instruction were relevant and even required at these two stations.

The picture changed when educators’ questions were examined. The educators’ interaction with children at different stations was different compared to the findings on educators’ instruction at the stations; questions were posed more frequently at the storytelling, Lego-building, and symbol-matching stations than at the Kinect-playing station. This finding seems to fit particularly well with the design of the Kinect-playing station as an activity during which a child plays games relatively independently. Such an activity differs from the activities at the storytelling, Lego-building, and symbol-matching stations, which are more collaborative. This factor seems to encourage the educators to interact with the children beyond providing direct instructions (i.e. by posing questions). The workstations were also arranged differently physically: while Kinect games were played independently with other people present as an audience, working at the three other stations included sitting next to one’s co-participant and focusing on the task together, subsequently supporting the active role of the educators. Furthermore, the nature of the knowledge required of the children at the stations differed particularly between the symbol-matching, storytelling, and Kinect stations. While the symbol-matching and storytelling stations aimed at developing the children’s conceptual knowledge, the Kinect station focused on improving procedural knowledge. Thus, educators may have asked more questions at the symbol-matching and storytelling stations in order to elaborate on concepts, whereas direct instructions were sufficient to encourage the children at the Kinect station to work on a task using their bodies.

These findings are important as they suggest that the design of technologies can influence the frequency of educators’ interactions with children. It is known that educators’ supportive interaction with children with ASD facilitates their learning. Thus, effective teacher-led interventions and methods, for example to teach communication and interaction skills have been developed (Leaf et al., 2015; Verschuur, Huskens, Verhoeven, & Didden, 2017). Meanwhile, there have been a growing number of studies on technologies to facilitate the self-directed learning of children with ASD (Cannella-Malone, Brooks, & Tullis, 2013; Smith, Shepley, Alexander, Davis, & Ayres, 2015; Uzuegbunam, Wong, Cheung, & Ruble, 2015) in order to increase independent learning of children with ASD. These studies have however focused on the use of one technology and on learning a specific skill. Thus, the evidence of the advantages of self-directed learning by using technology is still scarce. The findings of this exploratory study indicate that it may be possible to predict some features of the interactions between educators and children with ASD when working in a LE with multiple technologies. The findings also resonate with previous studies that provide evidence on the influence of physical classroom environment (Delmolino & Harris, 2012; Martin, 2016; Scott, 2009) and virtual learning environment (Didehbani, Allen, Kandalaft, Krawczyk, & Chapman, 2016; Smith & Brown, 2000) on learning of children with ASD. In addition, specifically the use of technology has been found to enhance the learning of children with ASD (Grynszpan, Weiss, Perez-Diaz, & Gal, 2014; McKnight et al., 2016). Nevertheless, more research is needed to investigate how LEs with multiple technologies should be designed so that the potential of technology to support children’s independent acquisition of skills and children’s need for educator’s instruction could be optimized. The design of the LE in this study can be used as an example and inspiration for future studies.

As the study also found that the educators’ questions were related to the children’s verbal responses, it was not surprising to find that the children responded verbally more frequently at the storytelling, Lego-building, and symbol-matching stations than at the Kinect station. This might be due to the fact that the story content, Lego toys, and symbol characteristics were familiar to the children, which could have activated them to get more involved in the activities with their educators. In addition, as these workstations were less interactive than the Kinect workstation, the educators’ role as an interactive partner was emphasized. On the other hand, the children responded behaviorally more frequently at the Lego-building station than at the Kinect and storytelling stations. This demonstrated that the skills and knowledge required for each workstation varied and this variation influenced the triadic interaction given the fact that comparing to other workstations, the Lego-building station provided a collaborative activity during which the children were able to respond by performing task-relevant actions (e.g. picking up a Lego brick selected by the educator). The previous research supports these finding as it is known that the teaching methodologies and approaches that teachers use with children with ASD depend on the task nature and the teaching goal of the activities that the children are expected to complete (Kossyvaki et al., 2016; Schreibman et al., 2015).

Interestingly, although the educators posed more direct instructions than questions at the Kinect station, it seems that the children struggled in responding to such initiations. It may be that, when the educators gave direct instructions at the Lego station (e.g. find a white Lego), it was easy for the children to follow them. On the other hand, when the educators gave direct instructions at the Kinect station (i.e. a highly interactive environment compared with the Lego screen), the children might have had problems responding to the instructions. Support for these results can be found from previous studies on the influence of users’ characteristics and technical capabilities of the technical system on the facilitation of social interaction. For example, Ke et al. (2015) found in their study on adult facilitators’ role in a virtual-reality-based social interaction with children with ASD that the design of the virtual reality facilitated dynamic role switching between facilitators and the target children with ASD. Consequently, children switched social roles with facilitators (e.g. playing mentor or mentee, party host or guest), which awarded children with an ownership in social interactions, hence motivated their social participation. However, as the Ke et al. (2015) study focused on interaction in virtual reality, the finding does not indicate how similar role switching could be facilitated by technologies in real life. The results of this exploratory study however suggest that LEs with multiple technologies could facilitate the versatile interactions between educators and children with ASD, particularly if both the educators’ and children’s’ needs as well as the capabilities and limitations of the technologies are evenhandedly taken into account when designing the LE. However, more research is needed on how technology could be utilized in LEs to facilitate children’s flexible role switching and interaction with educators.

The differences in educators’ verbal guidance and children’s responses between the stations showed how the design and features of the technology-based solutions affected interactions around their use. The findings resonate with the results of previous studies (Guldberg, Parsons, Porayska-Pomsta, & Keay-Bright, 2017; McKnight et al., 2016) that indicate that technology influences teachers’ pedagogical activities in the classroom. As Jacklin and Farr (2005) and Bunning et al. (2010) have noted, children’s (inter)actions cannot be examined independently of the interactional context provided by their co-participants, in this case, educators. Our findings add to the previous research by showing that the complexity of technology-based solutions can affect children and educators in various ways, which Jacklin and Farr (2005) have described as a triangle-like relationship between a child, an educator, and technology. This study has demonstrated how a technology-enhanced LE that is designed to support collaboration between a child and their co-participant (e.g. the storytelling, Lego-building, and symbol-matching workstations) can result in a positive, engaging relationship, in line with the study by Jacklin and Farr (2005). Thus, technology-enhanced LEs have the potential to be effective in supporting communication and interaction between educators and children; yet, their design and features should be carefully considered prior to drawing any conclusions about the benefits of technology-based solutions in general. Our analysis also shows that this potential is not only limited to a certain context of use; there were no differences between the Finnish and Chinese samples in the study.

Despite this, the study also found that some children received more instructions at certain workstations than other children. Previously, Tuononen, Kiiskinen, & Kärnä (2014) have observed how the Lego-building activity supported social interactions for some children but seemed to hinder them for others. Had this previous study considered the educators’ actions, the mediating role of the co-participants would perhaps have become visible. Thus, despite working on the same technology-based tasks in similar settings, it was found that the children’s specific interactional contexts differed, making the activity itself different for each child. This aspect needs more attention in future studies when LEs with multiple technologies are investigated.

The findings of this exploratory study can be used in both the research and development of technology-enhanced LEs in the future. The implications for future practices should focus on developing different workstations in the classroom itself to better serve the complex education needs of children with ASD. Findings provide a preliminary scheme and procedures to develop technology-based workstations and LEs; more work should, therefore, be undertaken on designing and improving the content and goals of the workstations. Another important consideration should be taken concerning the role of human mediation (e.g. teachers and assistants) in technology use. Therefore, social interaction around technology use should be examined more. Concerning the line of research, it is imperative that future researchers systematically test the content, interactive methods, and generalization effects of workstations. Future research should expand to a comprehensive examination of the triangle-like relationship between the educators, the children, and the technology-based workstation.

Due to the scope of this study and the number of available research participants, we were limited in implementing the technology-enhanced LE to seven participants from two schools in Finland and China. It is necessary to replicate the study with more participants in different settings to establish and generalize findings concerning a technology-enhanced environment. The lack of empirical analysis of other variables, including social interaction with educators, matching accuracy, imitation, and initiation, also limited our interpretation of the function of the technology-enhanced LE with children with ASD. Further studies should focus on other variables related to LEs to better develop and modify the function of different workstations. Finally, the training content of the matching and storytelling workstations needs to be adapted when it is applied to different cultural settings.

Despite these limitations, the results of the current study have important implications for technology-enhanced LEs for children with ASD in Finland, China, and beyond. Our findings suggest that a responsive, educator-assisted, technology-enhanced LE is promising in supporting the verbal development and social interaction of children displaying profound verbal limitations. It is also possible to train educators to implement technology-based workstations in school settings. Not only is this important in order to promote responsiveness and engagement, but it is also necessary in a low-resource setting, such as in China, where professionals are sparse, so as to provide quality services for children with ASD.

Acknowledgements

The authors are grateful to the research participants. They thank Master’s students Aino Tölli and Nelli Nevalainen for their help with the coding of the Finnish data and Wenjing Fan and Xianmei Lei for their support with the coding of the Chinese data.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributor

Eija Kärnä works as a professor of Special Education at the Faculty of philosophy, University of Eastern Finland. Professor Kärnä has been involved as a principle investigator in many national and international development and research projects. The projects have been funded by organizations such as the Ministry of Foreign Affairs of Finland, the Academy of Finland, Finnish Slot Machine Association, and Regional Council of North Karelia. All development and research projects have been multidisciplinary. Her research interests are inclusive learning environments, technology for individuals with special needs, communication and interaction of individuals with severe developmental disabilities and ASD, and digital literacy for various age groups.

Katja Dindar holds a PhD in Psychology and is a licensed psychologist. Her research interests include multimodal interaction and communication, especially in individuals with autism spectrum disorder. She is about to start her work as a postdoctoral researcher at the Faculty of Humanities at the University of Oulu, Finland.

Xiaoyi Hu works as an associate professor of Department of Special Education, Faculty of Education, Beijing Normal University, and the director of Education Research Center for Children with Autism, Faculty of Education, Beijing Normal University. She is the Board Certified Behavior Analyst (BCBA) and the coordinator for BCBA program at the Beijing Normal University. Her interests focus on education for children with autism, and the family support to children with disabilities.

ORCID

Eija Kärnä http://orcid.org/0000-0001-8107-3604

Xiaoyi Hu http://orcid.org/0000-0002-6936-4799

Additional information

Funding

This research was supported by 2018 Comprehensive Discipline Construction Fund, Faculty of Education, Beijing Normal University. The research was also funded by the Academy of Finland (grant number 140450). Katja Dindar was also supported by the Alfred Kordelin Foundation and the Olvi Foundation.