Cae D 17 00325

Computers & Education
Manuscript Draft
Manuscript Number: CAE-D-17-00325
Title: Engagement and collaboration of students with autism during

elementary computer science instruction
Article Type: Research Paper
Keywords: Elementary computer science;

students with Autism Spectrum Disorder;
engagement;
collaboration;
Collaborative-Computing Observation Instrument (C-COI)
Abstract: Elementary schools have increasingly added computer science

(CS) and programming activities into their curricula. This focus on CS in
K-5 education means that there is an increasing range of young learners
engaged in CS activities in schools. The limited studies in CS education
that focus on academically diverse learners have shown that CS activities
can be both engaging and challenging, especially for students with
disabilities who struggle with complex problem solving and social
interactions. It is, therefore, important to gain further understanding
into the engagement and learning of these students during CS instruction,
particularly for students who may need specialized supports and
interventions so they can have positive computing experiences. This paper
describes findings from a study of three elementary students with autism
spectrum disorder (ASD) during CS instruction. Results indicated that the
students experienced multiple challenges with completing the computing
activities as well as with collaborative problem-solving. These
challenges all influenced their level of engagement during CS activities.
*Manuscript -- nothing identifying the author should be listed here
Click here to view linked References
1 Running Head: ENGAGEMENT OF STUDENTS WITH AUTISM IN CS EDUCATION

2
3
4
5
6
7
8
9
10 A Descriptive Analysis of Engagement and Collaboration of Students with Autism during
11
12 Elementary Computer Science Instruction: A Mixed Methods Study
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
2
1
2 ENGAGEMENT OF STUDENTS WITH AUTISM IN CS EDUCATION
3
4 Abstract
5
6
7 Elementary schools have increasingly added computer science (CS) and programming activities
8
9 into their curricula. This focus on CS in K-5 education means that there is an increasing range of
10
11
12 young learners engaged in CS activities in schools. The limited studies in CS education that
13
14 focus on academically diverse learners have shown that CS activities can be both engaging and
15
16
17 challenging, especially for students with disabilities who struggle with complex problem solving
18
19 and social interactions. It is, therefore, important to gain further understanding into the
20
21 engagement and learning of these students during CS instruction, particularly for students who
22
23
24 may need specialized supports and interventions so they can have positive computing
25
26 experiences. This paper describes findings from a study of three elementary students with autism
27
28
29 spectrum disorder (ASD) during CS instruction. Results indicated that the students experienced
30
31 multiple challenges with completing the computing activities as well as with collaborative
32
33
34 problem-solving. These challenges all influenced their level of engagement during CS activities.
35
36 Keywords. Elementary computer science, students with Autism Spectrum Disorder,
37
38
39
engagement, collaboration, Collaborative-Computing Observation Instrument (C-COI)
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
3
1
3
4 1. Introduction
5
6
7 Computational thinking (CT) is the term often used in K-12 computer science (CS)
8
9 education to describe the set of cognitive skills and approaches necessary for using computers to
10
11
12 solve complex problems (Wing, 2006). As part of the Computer Science Teachers Association’s
13
14 CT Task Force, Lee and Martin clarified the definition of CT as a thought process involving “the
15
16
17 human ability to formulate problems so that their solutions can be represented as computational
18
19 steps or algorithms to be carried out by a computer” (Lee, 2016, p. 3). CT is not new in K-12
20
21 education. In fact, in 1980, Papert described how young learners could enhance their thinking
22
23
24 through LOGO programming in Mindstorms (Papert, 1980). Despite the long history of CT in K-
25
26 12 education, CS and CT have only recently been discussed in terms of wide-scale instructional
27
28
29 implementation. Within the context of science, technology, engineering, and mathematics
30
31 (STEM) education in the U.S., this wide-scale adoption of CS in K-12 schools is exemplified by
32
33
34 CS for All initiatives, such as the 2016 CS for All White House Initiative
35
36 (https://www.whitehouse.gov/blog/2016/01/30/computer-science-all). Other initiatives include
37
38
39
the development of a K-12 Computer Science Framework (https://k12cs.org/) and the Hour of
40
41 Code (https://hourofcode.com/us) in which millions of students received exposure to CS through
42
43 an hour of computing. Additionally, this growing interest in integrating CS and CT within K-12
44
45
46 curriculum has resulted in schools adopting CS/CT instructional activities using computing
47
48 software, curricula, and instructional modules targeting K-12 students.
49
50
51 1.1 Students with Disabilities and CS/CT Education
52
53 In the United States, approximately 6.5 million students (13% of all students) in public
54
55
56
schools receive special education services due to a disability (U.S. Department of Education,
57
58 2015). Importantly, 95% of these students attend regular schools, and most them spend most of
59
60
61
62
63
64
65
4
1
3
4 their time in general education classrooms alongside their non-disabled peers (U.S. Department
5
6
7 of Education, 2015). This means that students with disabilities are likely participating in K-12
8
9 CS/CT instruction. Regardless of increased interest in a range of CS/CT experiences for students
10
11
12 in K-12 (Grover, Pea, & Cooper, 2015; Lee, Martin, & Apone, 2014), only limited research has
13
14 been done on developing CS/CT instructional supports for students with disabilities (e.g.,
15
16
17 Chiang, Thorpe, & Lubke, 1984; Ratcliff & Anderson, 2011; Snodgrass, Israel, & Reese, 2016).
18
19 In fact, although initiatives such as the Google and Gallup studies are beginning to examine
20
21 underrepresented students’ participation in CS education (see Diversity Gaps in Computer
22
23
24 Science: Exploring the underrepresentation of girls, Blacks, and Hispanics, 2016), there currently
25
26 is not a focus on examining the extent to which students with disabilities are included in such
27
28
29 experiences. Even though there is growing attention to the participation of students with
30
31 disabilities in CS instruction (e.g., the AccessCS10K Center at the University of Washington),
32
33
34 there is still little research discussing how students with cognitive or social-behavioral
35
36 disabilities are being included in K-12 CS initiatives or the effectiveness of CS curricula and
37
38
39
software on these learners. In one of only a few published studies of students with disabilities,
40
41 Ratcliff and Anderson (2011) found that students with disabilities benefitted from CS education
42
43 activities that used the LOGO programming environment, especially if the students were given
44
45
46 supports such as modeling, explicit instruction, opportunities for unplugged activities, and
47
48 physical activities to support more abstract programming tasks. In another study, Snodgrass,
49
50
51 Israel, and Reese (2016) studied the participation of two young students with disabilities during
52
53 computing instruction. The resulting data revealed that the students’ engagement in computing
54
55
56
tasks increased when teachers applied the individualized, student-specific supports that were
57
58 developed, used, and found to be effective in other content areas (e.g., prompting procedures or
59
60
61
62
63
64
65
5
1
3
4 behavioral interventions). The authors stated, however, that there is a need for more research to
5
6
7 better understand the challenges that students with disabilities face during computing instruction
8
9 to tailor strategies to increase both access and engagement in computing CS/CT activities.
10
11
12 1.2 Collaborative Problem Solving and CS/CT
13
14 Within K-12 CS/CT, there is an emphasis on student collaboration. For example, the K-
15
16
17 12 Computer Science Framework (2016), which is the set of guidelines designed to inform
18
19 standards, curricula, and course pathways in CS/CT, described collaboration as a core
20
21 computational practice. According to the K-12 CS Framework, this practice involves cultivating
22
23
24 working relationships, using and establishing team norms, soliciting feedback, and using
25
26 technological tools that support collaborative computing (K-12 CS Framework, 2016).
27
28
29 Previous research suggests that teachers often create situations wherein students
30
31 collaborate during CS activities (Israel, Pearson, Tapia, Wherfel, & Reese, 2015). Denner,
32
33
34 Werner, Campe, and Oritz (2014) found positive effects when middle school students engaged in
35
36 pair programming, a type of collaborative computing process wherein two people simultaneously
37
38
39
work on one computer to complete a programming task. Israel and colleagues (2015) found that
40
41 teachers often encouraged student collaboration during elementary CS instruction through either
42
43 teacher-facilitated collaboration such as formalized peer tutoring to help students who needed
44
45
46 additional support or naturally occurring collaborations in which students independently sought
47
48 partners with whom to collaborate to complete CS activities together.
49
50
51 Hesse, Care, Buder, Sassenberg, and Griffin (2015) explained that collaborative problem-
52
53 solving within computer-supported collaborative learning (CSCL) is complex and relies on
54
55
56
sophisticated social and cognitive skills. Consequently, within CSCL studies, there is a strand of
57
58 research focused on scaffolding and encouraging student collaboration by scripting collaborative
59
60
61
62
63
64
65
6
1
3
4 conversations. Although not a major focus in K-12 CS education, these scripted conversations
5
6
7 are used to support both learning and social interactions within the broader CSCL field (e.g.,
8
9 Koller, Fisher, & Hesse, 2016; Webb et al., 2009). In fact, in a meta-analysis of conversation
10
11
12 scripting within CSCL activities, Vogel, Wecker, Kollar, and Fisher (2016) found that the use of
13
14 scripted conversations provided students with guidance in how to interact with each other as well
15
16
17 as increase learning. Given the potential impact of scripted conversations on collaborative
18
19 interactions within the greater CSCL literature, conversation scripting may be an extension that
20
21 could be impactful within K-12 CS instruction, especially for students with disabilities who
22
23
24 struggle with meaningful collaboration within more open collaborative experiences.
25
26 1.3 Students with Autism and Collaborative Computing
27
28
29 Autism spectrum disorder (ASD) is considered a neurobiological disorder that impairs
30
31 social interactions, communication, and behavior (Diagnostic and Statistical Manual; 5th ed.;
32
33
34 DSM-V). ASD results in a highly variable set of functional and cognitive profiles, which means
35
36 that no person with ASD will have the same issues with communication, social interactions, or
37
38
39
executive functioning (Center for Disease Control and Prevention [CDC], 2012). When
40
41 considering social interactions, even students with ASD who have the ability to use age-
42
43 appropriate language often have challenges associated with social interactions and
44
45
46 communication. Consequently, many students with ASD have difficulty collaborating with their
47
48 peers (Koegel, Vernon, Koegel, Koegel, & Paullin, 2012). McGee, Feldman, and Morrior
49
50
51 (1997), for example, compared social behaviors between typically developing children and
52
53 children with ASD. Their study found that students with ASD were less likely to be near other
54
55
56
students and had fewer social initiations with peers.
57
58
59
60
61
62
63
64
65
7
1
3
4 When examining cognitive skills, it becomes apparent that many students with ASD also
5
6
7 show variability in their mathematics and literacy abilities, such as over- and underachievement
8
9 in these areas (Charman, Jones, Simonoff, & Happe, 2011). A large percentage of students with
10
11
12 ASD exhibit below average achievement across academic domains (Wei, Christiano, Yu,
13
14 Wagner, & Spiker, 2014). Given Hesse and colleagues’ (2015) explanation that collaboration
15
16
17 within computer-supported activities often requires students to engage in complex social and
18
19 cognitive interactions, it is likely that students with ASD may struggle during collaborative
20
21 problem solving within the context of CS/CT.
22
23
24 1.4 Purpose of the study.
25
26 Because of the emphasis within the CS education community on authentic participation
27
28
29 and collaboration, it is imperative to begin to study how students with ASD and other disabilities
30
31 interact with peers and engage in computing instruction. This focused attention on students with
32
33
34 disabilities is critical so that instructional supports can be developed to help this group of
35
36 learners meaningfully participate in CS activities. The purpose of this study, therefore, was to
37
38
39
examine how students with ASD behaved during CS/CT instruction. The research question for
40
41 this study was: After being taught a collaborative script called the Collaborative Discussion
42
43 Framework (CDF; Park & Lash, 2014) to help facilitate peer collaboration, what individual and
44
45
46 collaborative computing behaviors do students with ASD exhibit during computing instruction?
47
48 The initial hypothesis was that the introduction of the CDF into the computing instruction would
49
50
51 promote collaborative problem solving and engagement as defined in the methods section.
52
53 2. Materials and Methods
54
55
56
Given the paucity of research on students with disabilities in CS/CT, it was important to
57
58 use an exploratory methodological approach. This study, therefore, made use of instrumental
59
60
61
62
63
64
65
8
1
3
4 case studies (Stake, 1994) of three students identified as having ASD at a mid-sized elementary
5
6
7 school in the Midwest that included CS and CT as part of its general curriculum. These three
8
9 case studies were considered distinct in that each student had unique strengths and challenges
10
11
12 related to communication, socialization, and computing. As Stake (1994) explained, instrumental
13
14 case studies provide insight into an issue or to refine a theory. Since the purpose of this study
15
16
17 was to understand how students with ASD functioned and collaborated within CS/CT
18
19 instruction, this approach was implemented. Additionally, information from each of the three
20
21 individual case studies was also examined collectively to gain a broader understanding of how
22
23
24 the students with ASD functioned within elementary computing classes. Stake (1994) explained
25
26 that collective case studies are used to “inquire into the phenomenon, population, or general
27
28
29 condition” and these cases were chosen to “lead to better understanding, perhaps better
30
31 theorizing, about a still larger collection of cases” (p. 237). Thus, examining the three cases both
32
33
34 independently and collectively allowed for exploration of the unique information from each of
35
36 these as well as commonalities across the cases.
37
38
39
2.1 Computing Curriculum
40
41 The computing curriculum used in this study was Code.org Code Studio (http://code.org),
42
43 a widely used computing program for K-8 students. Code Studio utilizes a block-based, visually-
44
45
46 intuitive programming language wherein students drag and connect blocks of commands to
47
48 complete increasingly complex computing puzzles until they master skills that allow them to
49
50
51 freely create programs within the Code Studio Play Lab. Code Studio also includes a series of
52
53 unplugged activities, lessons that teach computing concepts without the use of technology
54
55
56
(Prottsman, 2014). Content taught in Code.org Code Studio includes sequencing, looping, nested
57
58 looping, conditionals, functions, and debugging. For example, sequencing is introduced early in
59
60
61
62
63
64
65
9
1
3
4 each of the three Code Studio levels. Sequencing puzzles start participants off with simple
5
6
7 directional tasks that progressively increase in difficulty. Concepts are often introduced in
8
9 multiple ways and emphasize that different algorithms can be written to accomplish the same
10
11
12 task, such as during the early Stage 3 Maze puzzles. Algorithmic formulation and sequencing are
13
14 furthered developed in a different context during the Stage 4. By Stage 6, loops are introduced as
15
16
17 part of Maze puzzles that emphasize how different algorithms can accomplish the same task with
18
19 differences in the level of efficiency.
20
21
2.2 Participants
22
23
24 Participants for this study were recruited from a Midwestern elementary school that had a
25
26 CS for All initiative wherein all students in grades K-5 were exposed to CS/CT during weekly
27
28
29 computing times (approximately 45 minutes per week) as well as within activities that integrated
30
31 CS/CT into the content areas. This school had a wide range of socioeconomic and cultural
32
33
34 diversity. At the time of this study, 353 students were enrolled in grades K-5. Within this student
35
36 population, 19% of the students received special education services for documented disabilities
37
38
39
and 74% of the students were classified as living in low-income households. The racial/ethnic
40
41 composition of the school was 41% African-American and 41% Caucasian with the remaining
42
43 18% split between Latino/a, Asian, and Multiracial.
44
45
46 After approval from the Institutional Review Board, three students were purposefully
47
48 selected to participate in this study based on the following inclusionary criteria: (1) parents
49
50
51 provided consent for participation in this study, (2) students received special education services
52
53 for a documented disability of ASD, and (3) students had basic computer skills. Basic computer
54
55
56
skills involved being able to turn on the computer and log into Code.org Code Studio with
57
58 limited assistance. It also involved being able to use the mouse or trackpad to navigate Code
59
60
61
62
63
64
65
10
1
3
4 Studio software. Additionally, the research team purposefully selected three students with ASD
5
6
7 who had differing challenges during CS/CT instruction in order to gain enough diversity across
8
9 context (Stake, 2006). In each case study, the students and their peers participated in weekly
10
11
12 CS/CT instruction during which they worked through modules within Code.org’s Code Studio.
13
14 Data collection took place between November 2015 through January 2016.
15
16
17 2.2.1 Bradley. Bradley was a Caucasian male in third grade. He spent most the school
18
19 day in the general education classroom. He received special education services for a primary
20
21 disability of ASD and a secondary disability of attention deficit hyperactivity disorder (ADHD).
22
23
24 He did not require the support of an instructional aide to participate in his general education
25
26 classes. According to Bradley’s teacher, he could decode words and do arithmetic calculations,
27
28
29 but he had difficulty with complex problem solving, text comprehension, and application and
30
31 generalization of skills from one context to another. For example, he could decode grade-level
32
33
34 text but had a hard time comprehending that text, and he could quickly calculate arithmetic
35
36 problems but had difficulty when problems were presented as word problems. His teacher
37
38
39
explained, “I think comprehension is really tough, but he can read anything that you put in front
40
41 of him…In math, he’s really good at his basic math facts, but I think application of those is hard
42
43 for him” (classroom teacher, interview transcript lines 25-29). His teacher further stated that
44
45
46 Bradley was highly motivated by learning new concepts, completing tasks, and persisting
47
48 through difficulties that he faced. She explained, “He definitely perseveres a lot through the
49
50
51 things he’s stuck on, but he also gets frustrated. Once he gets it, he’s super excited” (classroom
52
53 teacher, interview transcript lines 22-23).
54
55
56
2.2.2 Alex. Alex was a Caucasian male in third grade. He was Bradley’s twin brother.
57
58 He spent most the day in general education and received special education services for the
59
60
61
62
63
64
65
11
1
3
4 primary disability of ASD. He did not receive support from an instructional assistant. Like his
5
6
7 brother, Alex had basic computer skills, strong decoding and arithmetic skills, and difficulty with
8
9 reading comprehension and mathematical application. His teacher explained that Alex is
10
11
12 “definitely more introverted but he’s also excited when he learns something new” (classroom
13
14 teacher, interview transcript lines 33-34). Because both brothers were in the same classroom,
15
16
17 Alex often followed Bradley around the classroom. The classroom teacher stated that Alex was
18
19 motivated to learn new things and felt proud of himself when he accomplished a task. However,
20
21 he also got frustrated quickly when he was stuck on unfamiliar tasks and verbally expressed his
22
23
24 frustrations. When frustrated, he typically stopped working rather than asking for help. Alex’s
25
26 teacher stated that he enjoyed being on the computer, especially to play video games.
27
28
29 2.2.3 Demetrius. Demetrius was an African-American male in fifth grade. He received
30
31 special education services for a primary disability of ASD and secondary disability of intellectual
32
33
34 disability. Demetrius spent most of his school day in general education with one-to-one support
35
36 by an instructional assistant. During class, Demetrius was rarely seen engaging in social
37
38
39
interactions with his peers. The technology coach who assisted in computing in Demetrius’s
40
41 class described him as having very limited verbal communication although he saw Demetrius’s
42
43 social interactions as improving over the past two years. According to his classroom teacher,
44
45
46 Demetrius was well below the instructional level of his peers in all content areas, and typically
47
48 performed at the first-grade level in reading and mathematics. Both his classroom teacher and
49
50
51 technology coach indicated that Demetrius enjoyed using the computer, especially to play
52
53 arithmetic games. The technology coach further explained, “He is perfectly capable of doing
54
55
56
basic level computer IT [information technology] things. He can log in, enter his password, use
57
58 the trackpad...He can do all those things” (Technology coach, interview transcript lines 15-17).
59
60
61
62
63
64
65
12
1
3
4 2.3 Instructional Strategy: The Collaborative Discussion Framework Conversation Script
5
6
7 To facilitate collaborative problem solving during CS/CT instruction, students in this
8
9 study (both participating students and their peers) were taught to use a scripted conversation
10
11
12 called the Collaborative Discussion Framework (CDF; Park & Lash, 2014). The CDF focused on
13
14 adaptive help seeking and collaborative interactions by encouraging students to have more robust
15
16
17 conversations when discussing their CS/CT activities. This strategy was implemented because
18
19 teachers reported that many students, particularly those with disabilities, often exhibited low
20
21 persistence when confronted with a difficult task and would ask their peers or teachers to solve
22
23
24 their computing problems for them when they were stuck. To address these challenges, the
25
26 students were taught to use the CDF to ask for help and explain their difficulties rather than
27
28
29 simply verbalizing frustration such as saying “This is hard. I can’t do it.” The CDF script
30
31 includes four guiding questions: (1) What are you trying to do? (2) What have you tried already?
32
33
34 (3) What else do you think you can try? (4) What would happen if…?
35
36 The students were explicitly taught to use the CDF through (a) teacher modeling of what
37
38
39
to do when stuck on a computing task, (b) student practice with teacher feedback and
40
41 encouragement, and (c) visual prompts of the CDF by placing posters of the CDF in the
42
43 classrooms and referring to these when prompting the students to use the CDF (see Figure 1).
44
45
46 The three students in this study were encouraged to use the CDF whenever they needed help or
47
48 when they helped their peers. In this way, the whole-class intervention could support the needs
49
50
51 of the students with disabilities without these students feeling singled out for using this strategy.
52
53
54
55
56
57
58
59
60
61
62
63
64
65
13
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 1. Classroom poster of the CDF script (Park & Lash, 2014) *Used with permission.
27
28 2.4 Data Collection and Analysis
29
30
31 Two primary forms of data collection were used in this study to gain an understanding of
32
33 students’ engagement and interaction during computing instruction. Classroom observations
34
35
36 were used to understand the participants’ engagement during the classroom computing activities.
37
38 Video screen capture data was then used to examine the students’ computing behaviors.
39
40
41 2.4.1 Classroom Observation Data Collection and Analysis. Four researchers observed
42
43 the participants during computing instruction. These researchers had experience in both
44
45
elementary schools and special education settings. A structured observation protocol was used
46
47
48 that replicated procedures outlined in Snodgrass and colleagues (2016) wherein an interval-based
49
50 observation process called momentary time sampling was used (Lewis, Scott, Wehby, & Willis,
51
52
53 2014). The observer recorded the level of engagement every minute of the students’ classroom
54
55 participation. The observers also noted if and how the participating students asked for help and
56
57
58 whether they used the CDF with peers and adults. Lastly, the observers recorded general
59
60
61
62
63
64
65
14
1
3
4 observations about what other students in the class were doing to compare the targeted students
5
6
7 to their peers. Prior to beginning classroom observations, the researchers discussed the
8
9 observation protocol and calibrated coding for interrater reliability purposes.
10
11
12 It was important to observe each participant multiple times in order to gain an
13
14 understanding of how they interacted with peers and teachers as well as with the computing
15
16
17 tasks. Consequently, Bradley and Demetrius were each observed during four computing sessions
18
19 and Alex was observed during five computing sessions. Teachers confirmed that these
20
21 observations served as typical instructional situations and that the students’ behaviors during
22
23
24 these observations were consistent with typical computing behaviors that the teachers observed.
25
26 2.4.2 Evaluating students’ academic engagement. Researchers have defined academic
27
28
29 engagement as a complex, multidimensional relationship between persistence, self-regulation,
30
31 and working towards goals (Christenson, Reschly, & Wylie, 2012). It involves interactions
32
33
34 between the student’s classroom setting, behaviors within those classroom settings, and
35
36 connections with adults and peers (Appleton, Christenson, & Furlong, 2008). Therefore,
37
38
39
measuring academic engagement can be quite challenging.
40
41 Because engagement is complex and difficult to evaluate, it was evaluated in two ways:
42
43 (a) level of engagement (from disengaged to engaged using the CDF strategy that was taught in
44
45
46 class), and (b) with whom or what the student was engaged (i.e., with the task on the computer,
47
48 with peers, and/or with the adults). When students were unengaged, the observers described the
49
50
51 unengaged behaviors. Definitions of engagement were as follows: (a) unengaged behaviors
52
53 included not looking at the computer, looking at the computer but working on a non-computing
54
55
56
task, and/or talking with others on topics not related to CS, (b) engaged behaviors included
57
58 looking at the computer screen when the computer screen included CS activities and actively
59
60
61
62
63
64
65
15
1
3
4 using the trackpad or mouse to do the CS activity, and/or talking with peers or adults about CS
5
6
7 activities, and (c) CDF engagement included all aspects of engaged behaviors as well as
8
9 interactions when the students initiated interactions and/or used the CDF conversation script with
10
11
12 peers or teachers to solve the problem. It is important to note that CDF engagement included
13
14 times when students used the CDF for problem-solving regardless of whether the problem was
15
16
17 fully solved and that students could be engaged with the computer and peer/adult at the same
18
19 time. The total engagement time was then tallied in the categories of (a) engagement with peers,
20
21 (b) engagement with adults, and (c) engagement with the computer/CS activity. Lastly, the total
22
23
24 amount of each engagement type was divided by the total time to compute the proportion of each
25
26 engagement level during observations. Thus, the level of engagement was evaluated both
27
28
29 quantitatively (i.e., amount of time engaged during CS activities) as well as qualitatively (i.e.,
30
31 field notes describing what the student was doing during the observation).
32
33
34 2.4.3 Video Screen Recording Data Collection and Analysis. Video screen recording
35
36 software was used to capture the students’ on-screen behaviors and record their voices. This data
37
38
39
was collected using Screencastify, a video screen capture program for Google Chrome
40
41 (http://www.screencastify.com) that captured on-screen actions that occurred during computing
42
43 instruction and conversations between peers and adults. Data from these video screen captures
44
45
46 were analyzed using the Collaborative Computing Observation Instrument (C-COI), which
47
48 allows researchers to analyze on-screen individual and collaborative behaviors of students during
49
50
51 computing activities (see Israel et al., 2016). This instrument was piloted and validated over the
52
53 course of three years and has a high level of inter-rater agreement.
54
55
56
The C-COI includes codes used to analyze video screen recordings of the three students.
57
58 These codes included 16 broad categories referred to as nodes in the instrument and associated
59
60
61
62
63
64
65
16
1
3
4 sub-nodes that provided additional details about the students’ independent and collaborative
5
6
7 behaviors while they engaged in CS/CT activities. Within the C-COI, the nodes describe (a) what
8
9 the student does during the computing activity (i.e., works independently or collaboratively), (b)
10
11
12 whether or not the student encounters a problem, (c) if the student was involved in collaborative
13
14 problem solving or socialization with peers or adults, and (d) how peers or adults respond to the
15
16
17 student (Israel et al., 2016). This data was then used to generate directed graphs to visually
18
19 represent the data. Figure 2 provides a screen capture of the C-COI interface. Additionally, the
20
21 researchers took extensive field notes including transcriptions of conversations and creative
22
23
24 suggestions given for solving problems to accompany the codes.
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46 Figure 2. Screenshot of the C-COI Interface.
47
48
49 Lastly, the codes from the C-COI could be combined into directed graphs that could be
50
51 visually analyzed for trends (Tatsuoka, 1986). These directed graphs visually represent each
52
53
node and subnode that were coded within the C-COI. Examining trends of the directed graphs
54
55
56
57
58
59
60
61
62
63
64
65
17
1
3
4 allowed the research team to closely examine the student processes of solving problems related
5
6
7 to coding. This added analysis was then triangulated with the classroom observation data.
8
9 Figure 3 provides an example of a directed graph showcasing nodes and subnodes from
10
11
12 the C-COI analysis across three different 20-minute observations for Demetrius, one student with
13
14 Autism Spectrum Disorder from this study. Each of the nodes is represented as a circled number,
15
16
17 and the subnode is represented by a letter on the directed graph arrows. A great deal of
18
19 information could be gleaned from visually examining these directed graphs. For example,
20
21 Figure 3 shows the directed graph for C-COI analysis of one of the students in this study
22
23
24 (Demetrius’s) video screen capture recordings while he engaged in computing. All paths began
25
26 at node 0, which had 3 distinct subnodes (0A-collaboration, 0B-independent work, 0C-other).
27
28
29 The thicker the line, the more instances that path of subnodes occurred. For example, the thickest
30
31 trendline was from 0B to node 15, showing that Demetrius primarily worked independently
32
33
34 without collaborating with a peer or adult (N=26). This directed graph also showed that the
35
36 student did engage in some collaborations (0A) to (1D), which were always initiated by an adult
37
38
39
(N=19). There were no instances of Demetrius initiating a peer (1A) or initiating an adult (1B).
40
41 Finally, the presence of subnode 0C indicates occasions wherein the student engaged in
42
43 behaviors other than working independently or collaboratively (N=8). In this example, during
44
45
46 some points of instruction the classroom was noisy that made it difficult to distinguish who was
47
48 talking. For a more detailed explanation of the C-COI, see Israel and colleagues (2016).
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
18
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28 Figure 3. Example of a Directed Graph Comprised of C-COI Nodes and Subnodes.
29
30 2.4.4 C-COI Interrater Agreement Procedures. Researchers who analyzed video
31
32
33 screen capture data underwent extensive training in using the C-COI. This involved watching
34
35 videos and coding them together to operationalize definitions followed by independently coding
36
37
38 the same videos and checking for consistency. This process was used with videos that were not
39
40 part of this study. Once the researchers calibrated coding procedures, 20% of the videos in this
41
42
43
study were coded to establish interrater agreement across the three students. The reliability
44
45 procedures used in this study followed recommendations by the C-COI developers and
46
47 consequently consisted of two levels (see Israel et al., 2016). The first phase consisted of broad
48
49
50 agreement related to the nodes, specifically whether an interaction or independent problem-
51
52 solving event occurred. To establish a high level of interrater consistency, the coders needed to
53
54
55 reach 100% agreement in this first level prior to moving onto the second level of interrater
56
57 agreement procedures. Once this first level of agreement was established, the coders began the
58
59
60
61
62
63
64
65
19
1
3
4 second phase, in which agreement was established at the subnode level. For the second phase,
5
6
7 there was 80% agreement between the researchers on the mutually- analyzed videos.
8
9 3. Results
10
11
12 3.1 Bradley
13
14 Bradley was observed during four computing sessions in which he worked through
15
16
17 computing puzzles in Code.org Code Studio. He was observed for a total of 103 minutes.
18
19 Bradley could begin computing activities alongside his peers. When CS/CT instruction began, he
20
21 retrieved his computer, logged into his student account, typed the URL for Code.org Code Studio
22
23
24 and entered his credentials into the site without supports. Bradley’s teacher stated that during
25
26 coding, he enjoyed his accomplishments, such as figuring out the correct set of blocks in Code
27
28
29 Studio that allowed him to transition to a higher puzzle or level. If Bradley was stuck, he
30
31 verbally expressed frustration and either went to his teacher or peers for help.
32
33
34 He appeared to be engaged in the computing task for 67% of the time across all four
35
36 observation sessions. When looking at the data, he had a high percentage of engagement with the
37
38
39
computing task as well as with peers on three of the four observations. The exception was on the
40
41 third observation, where he appeared distracted. On that day, he appeared to have a cold, as he
42
43 sniffed, used tissue, and then threw it away several times. Also, he was disengaged from the
44
45
46 computing activity to attend to another task in the room until his teacher redirected him back to
47
48 the computing task. Towards the end of class, however, he became interested in the computing
49
50
51 work of a friend and began discussing his friend’s work for a short time. Overall, Bradley was
52
53 engaged with his peers 42% of the observation time, of which he used the CDF script for most
54
55
56
interactions in three of the four observations. On the last observation, Bradley worked
57
58 independently for most of the session. At one point, he became excited about his work. He
59
60
61
62
63
64
65
20
1
3
4 expressed to his teacher that he won two trophies within Code.org Code Studio and asked her to
5
6
7 see his work. He and his teacher discussed the blocks he used. On two other occasions, he
8
9 indirectly sought help by looking at his peers and stating, “Nooooo. This is confusing!” and
10
11
12 “This is extremely hard. I can’t do it.” He did not use the CDF conversation script, but rather
13
14 asked his peers how to solve the computing puzzles. Figure 4 represents Bradley’s engagement
15
16
17 across sessions as related to engagement with peers, adults, and the computer.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Figure 4. Bradley’s Engagement during CS/CT Activities.
47
48 When engaged in the computing tasks, he worked individually or with a peer on
49
50 attempting to complete the Code Studio puzzles. It was common for Bradley to immediately ask
51
52
53 for help from a peer after starting a new puzzle. For example, on 11/2/2015, he asked his friend,
54
55 “Ronnie, can I have some help? Ronnie, you are not helping me!” On 11/12/2016, when Bradley
56
57
58 asked for help, his peer, Samantha, asked him to demonstrate the actions he wanted to perform
59
60
61
62
63
64
65
21
1
3
4 on the computer with his body. At that point, Bradley moved to the carpet and attempted to
5
6
7 perform the task physically by walking in a square, as that was the step that he could not figure
8
9 out in Code Studio (see Figure 5 for a screen capture of Bradley’s project). He sat down at his
10
11
12 computer once again and attempted to program the actions he physically performed on the
13
14 carpet. After a couple of minutes, he stated, “No one is helping me. This is just not fair.” At that
15
16
17 point, Bradley’s teacher asked whether he attempted to solve the computing task with the CDF to
18
19 which he replied that he did. She brought her chair to his desk and gave him verbal prompting to
20
21 help him problem solve the task.
22
23
24 During CDF engagement, Bradley often asked for help from his peers, but he answered
25
26 their CDF questions without specificity. For example, on 11/19/2015, when a peer asked “What
27
28
29 have you tried already?” Bradley answered by saying, “I don’t even know.” Although Bradley
30
31 could use the CDF, he did so more as a strict script than as a method to prompt authentic
32
33
34 conversations. In another example from 11/19/2015, Bradley took his computer to join Samantha
35
36 who was helping a friend. He used the questions in the CDF but appeared to do so more as a cue
37
38
39
card rather than to begin authentic interactions. That is, he asked Samantha each of the four
40
41 questions in the CDF in order without follow up questions and without using her answers to
42
43 prompt additional conversation. When unengaged, Bradley either chatted with friends about off-
44
45
46 task topics or expressed frustration.
47
48 During the observations, Bradley rarely interacted with adults (approximately 14.5% of
49
50
51 the time). Bradley did not ask the teacher for help unless she approached him first to check his
52
53 work. In fact, Ms. Post explained that when Bradley experienced an unfamiliar task, he often
54
55
56
“just shuts down” (Classroom teacher, interview transcript line 167). She stated that she believes
57
58 Bradley is showing slow improvement in this area. Many of the interactions between Bradley
59
60
61
62
63
64
65
22
1
3
4 and Ms. Post involved her coaching him to problem solve in a step-by-step manner. Figure 5
5
6
7 provides a picture of the program that Bradley was attempting to complete.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 5. Bradley's Computing Task.
30
31 Below was a typical interaction in which Ms. Post approached Bradley and attempted to help
32
33
34 him with his CS activity.
35
36 Ms. Post: “What did you do?”
37
38
39
Bradley: “I went move forward, move forward, left, move forward.”
40
41 Ms. Post: “Yeah. Did it work?”
42
43 Bradley: “No, wait. See?”
44
45
46 Ms. Post: What else do you have to do?
47
48 Bradley: I think I got to go right [pauses for two seconds] two times. [Bradley then dragged the
49
50
51 turn right by 90 degrees block.] I think it would work.
52
53 Ms. Post: What’s gonna happen when he turns right?
54
55
56
Bradley: No. He has to turn right two times.
57
58 Ms. Post: Which way is right?
59
60
61
62
63
64
65
23
1
3
4 Bradley: Right there.
5
6
7 Ms. Post: This way?
8
9 Bradley: Yes
10
11
12 Ms. Post: Look at this. If you go this way, what way is that? This is my what hand?
13
14 Bradley: [Pause for three seconds] The left
15
16
17 Ms. Post: Ooohhhhh!
18
19 Bradley: No, see. That is right.
20
21 Ms. Post: So which way does he have to turn?
22
23
24 Bradley: Left
25
26 They continued to work together. Although the problem was not solved completely, Bradley
27
28
29 expressed excitement because he succeeded in moving his sprite.
30
31 Analysis of video screen capture data using the C-COI confirmed analysis from the
32
33
34 classroom observations. In one collaborative event that lasted for 11 minutes, for example, in
35
36 which Bradley worked through a Code.org Code Studio puzzle, Bradley had six different
37
38
39
communications that fell into three types of interactions: (1) Bradley addressed a peer twice
40
41 seeking help with a computing task (1A: student verbally addresses a peer), (2) he once was
42
43 approached by his teacher and communicated with her about a problem he faced (1B: Student
44
45
46 verbally addresses an adult), and (3) on three occasions, he was approached by a peer without
47
48 initiating that interaction because the peer noticed that Bradley was struggling (1C: A peer
49
50
51 initiates the student to begin an interaction). These interactions related to problem-solving within
52
53 the context of CS/CT. Unlike his peers, who had more varied types of interactions that included a
54
55
56
combination of problem solving, socialization, and interactions wherein students expressed their
57
58
59
60
61
62
63
64
65
24
1
3
4 excitement about their CS/CT work as well as the work of their peers, most of Bradley’s
5
6
7 conversations focused on attempting to solve a challenge he faced during the computing task.
8
9 During these interactions, Bradley did not explicitly state his difficulty (2B: Student
10
11
12 expresses need for help but not explicit to the problem or topic). Consequently, his peers had to
13
14 ask probing questions to ascertain how to help him. Interactions typically centered around
15
16
17 problems that Bradley encountered (7A), but these interactions did not result in the problem or
18
19 issue being solved (9A). In fact, at the end of the 11-minute interaction, Bradley abandoned the
20
21 task and began working on another task. A partial transcript of a conversation between Bradley
22
23
24 and his peers is provided below:
25
26 Peer: What are you trying to do? That?
27
28
29 Bradley: Yes, I am trying to go up, so do you put a circle, do you… [student is interrupted
30
31 by his peer]
32
33
34 Peer: What have you tried already?
35
36 Bradley: Yeah, that worked, this worked.
37
38
39
Peer: Let me see.
40
41 They pressed Run, and a message appeared: “You are using all necessary types of blocks to
42
43 complete this puzzle.
44
45
46 Peer: See, that does not work because [pauses for three seconds], what else you think you
47
48 can try?
49
50
51 Bradley: I can’t do this. Oh gosh!
52
53 Peer: What would happen if you put a right or left?
54
55
56
Bradley: Let us see if left works okay? [Bradley drags the turn left by 90 degrees block into
57
58 the coding interface.]
59
60
61
62
63
64
65
25
1
3
4 Peer: Right there
5
6
7 Bradley: Right there?
8
9 Peer: On the top
10
11
12 Bradley: Left on the top? [Bradley dragged the turn left by 90 degrees] Do I put it on the
13
14 top?
15
16
17 Peer: Yeah, put it down. Put it down.
18
19 [Bradley places the turn left by 90 degrees block above the two moves forwards blocks.
20
21 Then, he pressed reset and ran the code. A message appears: “You are using all of the
22
23
24 necessary types of blocks, but try using more of these types of blocks to complete this
25
26 puzzle.”]
27
28
29 Bradley: Seeeeee! You have to have two move forwards.
30
31 Peer: You put the left at the bottom. [Bradley places the left by 90 degrees block at the
32
33
34 bottom of the two moves forwards blocks. He pressed reset and runs the code. A message
35
36 appears: “You are using all of the necessary types of blocks, but try using more of these
37
38
39
types of blocks to complete this puzzle”.
40
41 Bradley: There is not left!
42
43 Peer: No. You see, you turn left and then move forward.
44
45
46 Bradley: And then what left and then? [The peer stopped talking to Bradley and the
47
48 problem was not solved.]
49
50
51 This conversation illustrated the frustration Bradley experienced as he tried to program the
52
53 house in the puzzle. He could draw the bottom line of the house and turn the sprite to the
54
55
56
left, but he could not figure out how to move the sprite up to build the side of the house.
57
58
59
60
61
62
63
64
65
26
1
3
4 In this and other interactions between Bradley and his peers, the peer typically facilitated
5
6
7 the use of the CDF rather than Bradley. In these conversations, the peer attempted to coach
8
9 Bradley, who was having a difficult time understanding the purpose of the coding blocks, how
10
11
12 the blocks could be used together to complete the task, as well as difficulties with sequencing
13
14 and directionality. As can be seen both in the observation notes and C-COI data analysis,
15
16
17 although Bradley was engaged in the computing activities for most of the observations, he rarely
18
19 experienced success in completing computing tasks, either independently or collaboratively. This
20
21 lack of success was seen even when the CDF was used by Bradley, his peers, and his teacher.
22
23
24 3.2 Alex
25
26 Alex was observed during five computing sessions in which he worked in the Code.org
27
28
29 Code Studio. Alex could retrieve his assigned computer, access the Code.org website, and enter
30
31 his username and password. Across all five observations, Alex was actively engaged with the
32
33
34 computer and computing tasks 80% of the time. Figure 6 represents his engagement patterns
35
36 across the sessions as related to his peers, adults, and the computing tasks. He engaged in
37
38
39
conversations about computing with his peers 31% of the time and with his teacher 16% of the
40
41 time. During these times of engagement, Alex typically did not succeed in finishing his projects
42
43 or solving the computing puzzles. Like Bradley, when working with a peer, Alex typically did
44
45
46 not describe the difficulty he faced, and the peer did not know how to help him.
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
27
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 Figure 6: Alex's Engagement during CS/CT Activities.
31
32
33 On multiple occasions, Alex asked for help before attempting to work independently on
34
35 the computing task. The peer would then begin the CDF discussion by asking, “What did you
36
37
38 try?” However, because Alex did not attempt to solve anything prior to asking for help, the
39
40 conversation did not last (Observation notes, 11/12/15; 1/28/16). For example, on 11/12/2015,
41
42
43 after Alex asked a peer for help, the peer responded by using the CDF and asked, “What did you
44
45 try?” Alex did not respond, so the teacher prompted Alex by asking, “What did you need to do?”
46
47
Alex responded to his peer and the teacher by saying, “I didn’t try anything.” At this point, his
48
49
50 peer left the conversation, but the teacher remained with Alex to offer additional prompting to
51
52 solve the computing puzzle. After the teacher left, Alex immediately sought help from another
53
54
55 peer on a new computing puzzle. The peer stated that he was familiar with the coding task. At
56
57 that point, the first peer that Alex interacted with on the previous problem joined the new peer to
58
59
60
61
62
63
64
65
28
1
3
4 help Alex. The two peers took control of the computer and mouse to try to solve the computing
5
6
7 puzzle. When they did not immediately solve the puzzle, Alex said, “What? No!” The peers
8
9 attempted again successfully. Alex then independently tried to solve the next puzzle for a few
10
11
12 seconds unsuccessfully and said, “I hate code. I hate it.” At that point, he pushed his chair away
13
14 from the computer and said, “I didn’t even get to play. I’m mad.” At other times, Alex’s peers
15
16
17 were more successful in helping him by offering step-by-step directions. For example, on
18
19 11/19/2015, when Alex asked for help, his peer offered help by telling him the specific directions
20
21 and blocks to use. Alex then followed these explicit directions alongside his peer.
22
23
24 Within the times that Alex was actively engaged in communicating with his peers, he
25
26 displayed CDF engagement 37% of the time. During these times, Alex did independently ask for
27
28
29 help when he got stuck and communicated with a peer about coding through the CDF questions.
30
31 Within those CDF engagement instances, Alex tended to solve or get close to figuring out a
32
33
34 computing difficulty he faced. Additionally, when he solved the problem or moved on to the next
35
36 level, he verbally expressed his excitement such as “Yes! I did it, I did it! I’m on level seven”
37
38
39
(Observation notes, 11/2/15) or boast to himself, “Wow, this is easy. Wow, really easy!”
40
41 (Observation notes, 11/19/15).
42
43 When Alex was unengaged with the computing task, it was often in response to
44
45
46 frustration with being unable to solve the problem. For example, on one occasion, Alex struggled
47
48 to figure out the correct code, and within two minutes of working, he got frustrated and made the
49
50
51 “awwwww” sound. He then stopped working and simply sat at his computer (Observation notes,
52
53 12/3/15). He sometimes expressed his frustration verbally. For example, when frustrated, he said,
54
55
56
“I hate code, I hate it!” and then stopped both coding and interacting with others (Observation
57
58
59
60
61
62
63
64
65
29
1
3
4 notes, 11/12/15). At this point, Alex sat without doing anything for several minutes until the
5
6
7 teacher came over to inquire about how he was doing.
8
9 A 10-minute event captured through the C-COI analysis exemplified how Alex interacted
10
11
12 with his peers. In this event, Alex started working independently three times (0B: Student works
13
14 independently). He twice interacted with peers by requesting help (1A: Student verbally
15
16
17 addresses peers). In both instances, Alex did not explicitly explain his difficulty (2B: Student
18
19 expressed a need for help, but it is not explicit to the problem), and during both instances, even
20
21 with the peer’s assistance, the problem was unsolved. After less than two minutes, Alex asked
22
23
24 for help from another peer. A new peer came to his desk and used the CDF to attempt to
25
26 ascertain Alex’s difficulty. Again, Alex did not explicitly explain his problem to the peer. This
27
28
29 time, however, Alex looked at his peer’s computer to see if his peer was on the same coding
30
31 puzzle, which provided some necessary information. After the peer left, Alex attempted again to
32
33
34 solve the computing puzzle independently and was successful. The following example was the
35
36 first interactive instance with the second peer.
37
38
39
Alex: Did you pass this level?
40
41 Peer: Ehhh, yes.
42
43 Alex: I do not know what it is, but I did this correct. See? [Alex presses the Run button to show
44
45
46 the peer what he did so far].
47
48 Peer: Yes, you did, but I know how to do it. You need right, I think [Alex follow’s his peer’s
49
50
51 directions] and push forward. [Alex did not follow the peer’s direction this time; instead, he
52
53 clicked Run]. I did this before. Trust me man. Check if it works.
54
55
56
[Alex clicks run. A message appeared: "You are using all of the necessary types of blocks, but
57
58 try using more of these types of blocks to complete this puzzle]
59
60
61
62
63
64
65
30
1
3
4 Alex: Ohhhh
5
6
7 Peer: Aye!
8
9 Alex: Ahhh [Alex's voice indicated frustration]. This is so hard! Can you show your computer?
10
11
12 [Alex moves the turn right by 45 degrees and move forward by 100 pixels blocks into the trash]
13
14 Peer: I am not at that level. I am on a different level, man. I cannot remember how to do this.
15
16
17 Alex: Yeah, me either.
18
19 Like Bradley, Alex spent most of the computing time on task and engaged. However, he
20
21 had difficulty understanding the purpose of each programming block and how these blocks
22
23
24 worked to create a program. Despite these challenges, he was motivated to learn and got close to
25
26 solving the computing problems through the interaction with his peers.
27
28
29 3.3 Demetrius
30
31 Demetrius was observed during four computing sessions in which he worked on
32
33
34 Code.org’s Code Studio puzzles. Demetrius was familiar with the routine of computing and with
35
36 the Code.org Code Studio. He could take his computer to his desk independently. To log into
37
38
39
Code Studio, Demetrius used a cue card with step-by-step directions to enter his student account
40
41 and type the URL for Code.org. With the use of this cue card, he could successfully log into
42
43 Code Studio.
44
45
46 He often engaged in self-talk during computing time. For example, on 11/5/2015, he
47
48 repeated the phrase, “Let’s get started!” during the class. Demetrius was engaged with the
49
50
51 computing puzzles 82% of the time, although this level of engagement often involved returning
52
53 to computing tasks he previously completed rather than working on new tasks or puzzles. For
54
55
56
example, the observation on 11/12/15 showcased high engagement; however, when looking at
57
58 what Demetrius was doing during that observation, he was observed rerunning computing
59
60
61
62
63
64
65
31
1
3
4 puzzles that he previously completed. Demetrius did not do any new programming on this day
5
6
7 (such as adding new programming blocks). Rather, although he was engaged with the computing
8
9 software, this engagement only involved running his previous code and watching the “Flurb”
10
11
12 (i.e., the creature) get to the treasure in each level.
13
14 During other observations in which the computing puzzles were not already completed, it
15
16
17 did not appear that Demetrius attempted to complete the computing tasks independently. For
18
19 example, on 11/19/15, Demetrius was playing a Curious George video game on the computer
20
21 when his teacher informed the class that they were going to start coding time. He switched to
22
23
24 Code Studio and began to ask for help from one of the research assistants in the classroom as his
25
26 instructional aide was not yet in the classroom. The research assistant offered cues and verbal
27
28
29 prompts until the puzzle was solved. At that point, Demetrius turned off Screencastify and
30
31 switched back to the Curious George video game. The research assistant decided not to restart
32
33
34 the video screen recording as Demetrius was becoming agitated. Demetrius consequently played
35
36 this game until it was time to move onto another class activity.
37
38
39
Figure 8 showcases Demetrius’s engagement patterns across the sessions as related to
40
41 peers, his teachers, and the computing tasks. Demetrius did not interact with his peers during any
42
43 of the observations, and all interactions observed were initiated by the instructional aide.
44
45
46 According to the technology coach, his lack of collaboration was primarily due to his lack of
47
48 functional communication.
49
50
51 Demetrius’s computing behavior is best characterized as a cycle of working
52
53 independently for a few seconds on a computing puzzle he has previously completed or working
54
55
56
unsuccessfully on a new puzzle for a couple of seconds followed by skipping to another puzzle
57
58
59
60
61
62
63
64
65
32
1
3
4 and repeating the cycle. The amount of time he persisted independently on any given puzzle was
5
6
7 low (Median=21 seconds, Range= 12 - 89 seconds).
8
9 In some instances, the instructional aide took over Demetrius's computer and mouse to
10
11
12 get him through a difficulty because she was unsuccessful in helping him through verbal
13
14 prompting and support. For example, on 11/12/2015, Demetrius remained on the same
15
16
17 computing puzzle for over 20 minutes with little success. During this time, the instructional aide
18
19 sat near Demetrius during the entire observation. He worked on the computing puzzle for a few
20
21 seconds and then become distracted by something else. The instructional aide then attempted to
22
23
24 redirect him to the task. Observation notes indicated that rather than using a repeat block, which
25
26 was necessary to complete the task, Demetrius continued to add additional move forward blocks
27
28
29 for his code. The instructional aide provided verbal prompting to attempt to help Demetrius
30
31 understand the code. After approximately 20 minutes, the instructional aide took over the
32
33
34 computer and mouse for two minutes. She then handed the computer back to him and pointed
35
36 towards blocks that were needed to solve the computing puzzle successfully.
37
38
39
Demetrius did not independently attempt to interact with his peers, teacher or
40
41 instructional aide. If interactions did occur, these interactions almost always occurred because
42
43 the aide or teacher initiated those interactions (Observation notes, 11/5/15; 11/12/15; 11/19/15;
44
45
46 12/3/15). For example, on 12/3/2015, within the 16-minute observation, Demetrius did not
47
48 interact with his peers even though his peers were near him physically. This was consistent with
49
50
51 other observations. He interacted with the instructional aide for 11 minutes. These interactions
52
53 involved the aide assisting him in problem-solving when he was stuck.
54
55
56
57
58
59
60
61
62
63
64
65
33
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32 Figure 8: Demetrius's Engagement during CS/CT Activities.
33
34 In another example, on 12/3/2015, in five of the six interactions where the instructional
35
36
37
aide initiated an interaction with Demetrius around a computing problem, he did not respond to
38
39 this attempted interaction. In one instance, an interaction between Demetrius and the aide
40
41 occurred (6B: Adult and student interact), but the interaction led to the TA clearly stating that
42
43
44 she did not know how to help (8L: Adult and student discuss the problem/difficulty and then the
45
46 adult directs the student to another person).
47
48
49 When examining Demetrius’s interactions with the C-COI, it became clear that his most
50
51 common interactive behaviors were initiated by the instructional aide (1D: Student is initiated by
52
53
54
an adult) while he was working independently (4B: Adult offers support to the student who is
55
56 working independently on a problem or topic). The instructional aide usually began the
57
58
59
60
61
62
63
64
65
34
1
3
4 interactions by asking Demetrius questions about his computing/programming activities (5A:
5
6
7 Problem or topic is related to computing/programming). She asked him questions to attempt to
8
9 help him focus on the puzzle. These types of interactions typically ended in Demetrius ignoring
10
11
12 the instructional aide (6G: Student ignore or cannot hear student’s response).
13
14 For example, below is an example wherein the instructional aide attempted to interact with
15
16
17 Demetrius:
18
19 Instructional Aide: Do you know what that means? Every time you see a question mark you
20
21 have to add a number? Can you click in that? Click on that number?
22
23
24 Demetrius: [ignores and changes levels]
25
26 In both the classroom observations and CCOI analysis, although Demetrius was engaged
27
28
29 with computing, the quality of that engagement was low. When working independently, he
30
31 exhibited the repetitious behavior of starting a computing puzzle and immediately pressing run
32
33
34 without manipulating any blocks. After approximately 20-30 seconds, he would skip to the next
35
36 puzzle and repeat the same behavior. The CDF was not an effective intervention for Demetrius
37
38
39
as he needed more verbal and specific prompting on computing activities than his classmates.
40
41 4. Discussion
42
43 In this study, three students with ASD participated in class-wide CS/CT instruction using
44
45
46 a fairly structured computing curriculum that guided students through increasingly more
47
48 complex programming activities. The three students in this study had significant challenges both
49
50
51 with the content of those activities and with working collaboratively with their peers. Despite
52
53 these challenges, however, all three students did engage in the computing activities for much of
54
55
56
the instructional time, although the level of engagement they exhibited was not always high.
57
58 Three important findings emerged from this study that will be discussed in detail below. First,
59
60
61
62
63
64
65
35
1
3
4 the instructional challenges these students faced in CS/CT closely resembled academic
5
6
7 challenges that students with ASD face in areas such as mathematics. For example, the students
8
9 all had difficulty persisting and working through computing problems and often either avoided
10
11
12 challenging tasks or requested help prior to attempting to persevere. Second, although the
13
14 scripted questions of the CDF were taught to promote elaborated discussions and to guide
15
16
17 students’ interactions, the CDF on its own did not provide enough support to these learners.
18
19 Third, it did not appear that the students received sufficient supports to help them meaningfully
20
21 engage in the computing activities. That is, none of the observations made note of any
22
23
24 proactively planned individualization or scaffolding. Rather, when the students needed
25
26 immediate help, the teachers and peers provided that help.
27
28
29 4.1 CS/CT Challenges
30
31 In this study, although the students could all complete the steps of getting their computers
32
33
34 and signing into the computing software, both the observation and C-COI data revealed several
35
36 challenges including a lack of understanding of the programming blocks and/or how the blocks
37
38
39
could be used to solve the computing puzzles. Because the students in this study struggled with
40
41 completing the computing activities, they either gave up quickly or got frustrated. In many cases,
42
43 they immediately asked or received support from a peer or teacher.
44
45
46 4.1.1 Connection between mathematics and CS/CT challenges. There were several
47
48 parallels between the challenges that the students in this study faced with CS/CT and challenges
49
50
51 that many students with ASD face during mathematics instruction. The connection with
52
53 mathematics may provide a roadmap for developing instructional approaches in CS/CT that
54
55
56
could benefit students with ASD. For example, several studies in mathematics education research
57
58
59
60
61
62
63
64
65
36
1
3
4 point to promising instructional strategies developed and tested to increase math learning for
5
6
7 students with ASD (e.g. King, Lemons, & Davidson, 2016; Gevarter, et al., 2016).
8
9 It is important to remember that the mathematics achievement of students with ASD is
10
11
12 variable (King et al., 2016); this variability was also observed in this computing-focused study.
13
14 Despite these differences, the challenges that Bradley, Alex, and Demetrius faced in computing
15
16
17 were not unlike challenges that many students with ASD face during mathematics instruction.
18
19 For example, Bradley and Alex’s teacher explained that both students had strong recall of
20
21 mathematics facts, but they struggled with using these mathematics facts to solve problems. This
22
23
24 is consistent with findings from research on students with ASD showcasing the discrepancy
25
26 between arithmetic recall and problem-solving (Wei, Christiano, Yu, Wagner, & Spiker, 2014).
27
28
29 Future research should closely examine the relationship between challenges in computing
30
31 and challenges in mathematics. For example, Rockwell, Griffin, and Jones (2011) explained that
32
33
34 students with ASD struggle with mathematical problem solving because they need to use both
35
36 numeric and semantic knowledge. They investigated the use of schema-based instruction, which
37
38
39
helps students create mental problem-solution representations with visual representations of
40
41 problems alongside direct instruction. Rockwell et al. found that this instruction increased
42
43 mathematical problem-solving skills of students with ASD. These instructional approaches may
44
45
46 benefit students with ASD in computing instruction. In Bradley’s case, for example, offering him
47
48 a mechanism for visually representing the task and building a model prior to coding might have
49
50
51 helped him understand the next step in figuring out the problem presented in Figure 5.
52
53 Computing is a process that requires multiple steps, and those steps are often not clearly
54
55
56
defined. The students in this study expressed frustration about the computing activities they
57
58 could not solve independently. For these learners to be successful within CS/CT activities,
59
60
61
62
63
64
65
37
1
3
4 serious consideration must be given in how instruction can be delivered in a manner that
5
6
7 minimizes frustration for students who struggle with complex multi-step problems. Whether
8
9 intentionally or not, the adults and peers in this study provided explicit instruction to the students
10
11
12 help participate in the computing activities. This finding was consistent with results from Ratcliff
13
14 and Anderson (2011) in which students with disabilities engaged in computing activities when
15
16
17 provided additional supports that included explicit instruction.
18
19 Explicit instruction involves verbal prompting, modeling, and immediate feedback both
20
21 from teachers and peers, which can move students toward mastery by breaking down information
22
23
24 into meaningful chunks, so the students are not overwhelmed and immediately give up (Archer
25
26 & Hughes, 2011). The explicit instruction that the adults and peers used in this study was
27
28
29 implemented in reaction to the students’ difficulties and frustration rather than proactively to
30
31 mediate the difficulties. This explicit instruction involved the use of feedback to address errors
32
33
34 that the students made.
35
36 A complicating factor within computing instruction is that computing is typically
37
38
39
designed to be open-ended. Explicit instruction can seem like a contradictory instructional model
40
41 within computing. Therefore, many questions emerge about how to provide the right amount of
42
43 explicit instruction to students to require this type of instruction while maintaining the integrity
44
45
46 of the computing experience. Future research should therefore examine how to (a) proactively
47
48 build in some explicit instruction into the more open-ended computing tasks and (b) balance
49
50
51 explicit instruction within computing instruction in a manner that allows for student creativity
52
53 and choice alongside effective modeling and feedback.
54
55
56
Additionally, the current study had similar findings to those of Snodgrass et al. (2016) as
57
58 related to the need to examine the students’ individualized supports. Snodgrass and colleagues
59
60
61
62
63
64
65
38
1
3
4 found that prior to implementing each student’s individualized supports, the students exhibited
5
6
7 limited engagement in computing. Once the individualized supports that were used across the
8
9 school day were introduced into these students’ computing activities, their engagement
10
11
12 increased. Thus, they showcased the need to tailor instructional supports to the individuals with
13
14 disabilities in order to increase access and engagement within CS/CT. These findings were
15
16
17 unsurprising as there is a great deal of evidence that students with disabilities can be successful
18
19 in academic learning, including the STEM areas, once they are provided with individualized
20
21 supports. In the current study, the initial hypothesis was that the use of the CDF discussion
22
23
24 script would help the students collaboratively problem solve, but this strategy on its own was not
25
26 sufficient. It was not individualized and tailored to the unique needs of the students.
27
28
29 Additionally, the explicit instruction that was used occurred in reaction to difficulty rather than
30
31 proactively to attempt to remediate some of the challenges faced by the students. Future research
32
33
34 should extend both the current study and Snodgrass et al. (2016) to focus on how to translate
35
36 individualized supports that are successful in other instructional areas to CS/CT.
37
38
39
4.1.2 Locus of control and CS/CT engagement. For Bradley and Alex, the observation
40
41 and C-COI analysis revealed that their level of engagement was closely related to their level of
42
43 success with the computing tasks. When they made progress, the students were engaged. When
44
45
46 they were stuck, they often became frustrated and stopped working. For these students, the
47
48 connection between success or failure and engagement mirrors studies related to learned
49
50
51 helplessness, attribution theory and locus of control. Ability, effort, task difficulty, and luck have
52
53 been described as attributions for either perceived success or failure (Weiner, 1985). Within
54
55
56
these attributions, locus of control refers to the location of the cause of success or failure (e.g.,
57
58 internal or external to the individual) and controllability refers to the student being able to affect
59
60
61
62
63
64
65
39
1
3
4 the cause of success or failure such as the amount of preparation (Weiner, 1985; 2010). When
5
6
7 students have control of the outcome (e.g., they have strategies such as proactively planning their
8
9 steps to attempt to solve the problem), then they are more likely to persist in increasingly
10
11
12 difficult tasks. On the other hand, after repeated unsuccessful attempts of solving a task, students
13
14 can develop performance-avoidance tendencies and eventually learned helplessness manifests
15
16
17 (Jarvis & Seifert, 2002). These behaviors include giving up easily and asking for help the instant
18
19 a difficulty is experienced, which was observed in both Bradley’s and Alex’s data.
20
21 It was difficult to tell whether learned helplessness or attribution theory explained
22
23
24 Demetrius’ computing behaviors, as he enjoyed running programs that he previously completed
25
26 and was engaged with these activities. He did not attempt new computing puzzles unless the
27
28
29 instructional aide encouraged and scaffolded these experiences for him. When he was observed
30
31 on other computer-based activities, such as playing the Curious George video game, he did not
32
33
34 exhibit these same behaviors. Because of his limited functional verbal ability, however, it was
35
36 difficult to tell whether he chose to spend his time on already-completed puzzles because of
37
38
39
frustration with new tasks or because he simply liked to experience the completed puzzles. It
40
41 could also be that when Demetrius was unable to solve a computing puzzle, he might have
42
43 attributed his failure to lack of computing ability and thus gave up. However, his motivations and
44
45
46 emotions were masked by his inability to explain when and why he was frustrated.
47
48 Future research should focus on using non-verbal communication strategies to support
49
50
51 students with ASD that have limited verbal communication. For example, many individuals with
52
53 ASD prefer learning using visual materials rather than by listening (Cohen, 1998). Because of
54
55
56
these preferences, there is evidence suggesting that visual supports can lead to positive outcomes
57
58 to help students such as Demetrius express needs, challenges, and make requests of teachers and
59
60
61
62
63
64
65
40
1
3
4 peers. Examples of such supports include Picture Exchange Communication System (PECS;
5
6
7 Bondy & Frost, 1994) and social stories (Gray & Garand, 1993). The visual supports could be
8
9 pictures with words cuing the next step in a problem-solving strategy or a request for help. It can
10
11
12 also involve a cue to maintain attention, represent a concept in a concrete manner, or help the
13
14 student express thoughts (Rao & Gagie, 2006). These supports should be studied within the
15
16
17 context of computing in order to effectively make these strategies available to students with ASD
18
19 within computing activities and to see whether they can be used to increase meaningful
20
21 engagement for students such as Demetrius.
22
23
24 4.1.3 Peer Collaboration Challenges. Given the focus on collaboration within the CS
25
26 education community (K-12 CS Framework, 2016), it was critical to examine how the students
27
28
29 in this study interacted with their peers. Each of the students in this study had different levels of
30
31 social and communication strengths and deficits. This was consistent with the literature about
32
33
34 social and communication challenges of students with ASD (e.g., Koegel et al., 2012; McGee et
35
36 al., 1997). For example, although Bradley and Alex could verbally communicate with their peers
37
38
39
and teachers, the communication script of the CDF only afforded limited support. The CDF
40
41 provided a method for the students to ask for help. It also gave the peers an initial routine for
42
43 interacting with the students. However, these interactions did not prove to be particularly helpful
44
45
46 in either supporting the students in solving the computing tasks or promoting more authentic
47
48 interactions. For Demetrius, who had more limited verbal communication, the CDF was not
49
50
51 effective as he could not verbally ask for help or provide answers to the instructional aide if she
52
53 asked questions from the CDF. For him, it would be helpful to modify the CDF script so that the
54
55
56
instructional aide or peers could say things such as, “You tried to move the sprite left. I wonder
57
58 what would happen if we moved it right?” Although the students in this study received support
59
60
61
62
63
64
65
41
1
3
4 from peers, teachers, or the instructional aide, none of the students received enough support to
5
6
7 allow them to navigate the computing experiences successfully.
8
9 Findings related to the limited impact of the scripted communication of the CDF leads to
10
11
12 further questions about how to leverage the advantages of scripted conversations within
13
14 computing instruction, especially for students with disabilities. For example, it appeared that at
15
16
17 least for Bradley and Alex, there should have been a step prior to the CDF that could potentially
18
19 create the conditions for the CDF to be more effective. Because they immediately requested help,
20
21 the students in this study might benefit from strategies related to self-regulation to help them
22
23
24 moderate their frustration. Self-regulation refers to the process that allows students to engage in
25
26 goal-directed activities (Karoly, 1993; Schunk & Zimmerman, 1994). Self-regulation is critical
27
28
29 to success in CS/CT as it is involved in tasks such as planning, problem-solving, and choice
30
31 making. Additionally, several studies have tied self-regulation with join engagement (defined as
32
33
34 experiences wherein children actively engaged in an activity with another person) as well as
35
36 academic success (Jahromi, Bryce, & Swanson, 2013). Future research should closely study the
37
38
39
relationship between self-regulation, collaborative problem solving, and computing for students
40
41 with disabilities.
42
43 4.2 Limitations
44
45
46 The results of this study should be viewed within the context of several limitations. First,
47
48 although the participants in this study were recruited with specific criteria related to disability
49
50
51 diagnosis and basic computing ability, the limited number of participants and the variability
52
53 among students with ASD limits the generalizability of the findings. Replication studies will be
54
55
56
necessary to ascertain whether the challenges faced by the three students in this study are
57
58 consistent with other students with ASD. Second, this study only included four or five
59
60
61
62
63
64
65
42
1
3
4 observations for each of the participant’s computing and collaborative behaviors. Although this
5
6
7 number of observations was consistent with recommendations for C-COI use and most CSCL
8
9 studies examine what make single collaborative experiences successful or unsuccessful (Barron
10
11
12 et al., 2009), additional observations focused on the relationship between time, communication,
13
14 and learning could lead to a fuller understanding of the temporal dimensions of discourse and
15
16
17 learning (Mercer, 2008). Third, although the findings suggest a strong relationship between
18
19 challenges in CS/CT and challenges in mathematics, the current study did not include data
20
21 collection within mathematics instructional settings. Future studies should investigate the same
22
23
24 students across academic content areas to better understand the relationship between challenges
25
26 in CS/CT and other content areas. Lastly, although the study made use of the C-COI instrument,
27
28
29 a validated instrument with strong reliability measures, there might be aspects of student
30
31 engagement in computing that were not captured through this instrument alongside the
32
33
34 observations. For example, this study did not include student or parent interviews or other
35
36 measures of self-regulation or persistence. Future research should include additional measures to
37
38
39
triangulate observational and C-COI data.
40
41 5.0 Conclusions
42
43
44 This study provided an initial lens into the engagement of students with ASD in inclusive
45
46
47
computing instructional settings. Findings support the assertion that although the students in this
48
49 study and their peers were taught to use the CDF, a scripted conversation strategy to help
50
51 facilitate collaborative problem solving, and two of the three students in this study were able to
52
53
54 use the CDF, the students in this study needed additional support beyond the CDF to actively
55
56 engage in the CS/CT activities. These findings are important to consider to better understand
57
58
59
60
61
62
63
64
65
43
1
3
4 how all learners, including those with disabilities, can attain Papert’s (1980) hope that computing
5
6
7 can be a broad learning tool. The present study provides important directions for future research
8
9 on the integration of the CDF alongside strategies for metacognitive self-regulation and
10
11
12 additional explicit instruction. Findings also suggest a parallel between learning and engagement
13
14 within CS/CT and mathematics for students with ASD, which should be further explored in
15
16
17 future studies. Future studies should, therefore, also expand upon the design of this study by
18
19 examining student engagement and collaboration across both CS/CT and mathematics instruction
20
21
to compare challenges, supports, and instructional strategies across settings.
22
23
24 References
25
26
27 American Psychiatric Association. (2015). Diagnostic and statistical manual of mental disorders
28
29 (5th ed.) Washington, DC: Author.
30
31
32
Appleton, J. J., Christenson, S. L., & Furlong, M. J. (2008). Student engagement with school:
33
34 Critical conceptual and methodological issues of the construct. Psychology in the
35
36 Schools, 45(5), 369-386.
37
38
39 Archer, A. L., & Hughes, C. A. (2011). Explicit instruction: Effective and efficient teaching
40
41 (What works for special-needs learners). New York: The Guilford Press.
42
43
44 Barron, B., Martin, C. K., Mercier, E., Pea, R., Steinbock, D., Walter, S., ... & Tyson, K. (2009,
45
46 June). Repertoires of collaborative practice. In Proceedings of the 9th international
47
48
49 conference on Computer supported collaborative learning-Volume 2 (pp. 25-27).
50
51 International Society of the Learning Sciences.
52
53
54
55
56
57
58
59
60
61
62
63
64
65
44
1
3
4 Center for Disease Control and Prevention. (2012). Prevalence of autism spectrum disorders.
5
6
7 Morbidity and Mortality Weekly Report, 61(3), 1–24. Retrieved from
8
9 http://files.eric.ed.gov/fulltext/ED530639.pdf
10
11
12 Bondy, A. S., & Frost, L. A. (1994). The picture exchange communication system. Focus on
13
14 Autistic Behavior, 9(3), 1–19.
15
16
17 Charman, T., Jones, C. R., Pickles, A., Simonoff, E., Baird, G., & Happé, F. (2011). Defining the
18
19 cognitive phenotype of autism. Brain Research, 1380(22), 10-21.
20
21 Chiang, B., Thorpe, H. W., & Lubke, M. (1984). LD students tackle the LOGO language:
22
23
24 Strategies and implications. Journal of Learning Disabilities, 17(5), 303-304.
25
26 Christenson, S. L., Reschly, A. L., & Wylie, C. (Eds.). (2012). Handbook of Research on Student
27
28
29 Engagement. Springer Science & Business Media.
30
31 Cohen, S. (1998). Targeting autism: What we know, don’t know, and can do to help young
32
33
34 children with autism. Berkley, CA: University of California Press.
35
36 Denner, J., Werner, L., Campe, S., & Ortiz, E. (2014). Pair programming: Under what conditions
37
38
39
is it advantageous for middle school students? Journal of Research on Technology in
40
41 Education, 46(3), 277-296.
42
43 Dweck, C. S. (2006). Mindset: The new psychology of success. Random House.
44
45
46 Gevarter, C., Bryant, D. P., Bryant, B., Watkins, L., Zamora, C., & Sammarco, N. (2016).
47
48 Mathematics interventions for individuals with autism spectrum disorder: A systematic
49
50
51 review. Review Journal of Autism and Developmental Disorders, 1-15.
52
53 Google Inc. & Gallup Inc. (2016). Diversity Gaps in Computer Science: Exploring the
54
55
56
Underrepresentation of Girls, Blacks and Hispanics. Retrieved from
57
58 http://goo.gl/PG34aH.
59
60
61
62
63
64
65
45
1
3
4 Gray, C. A., & Garand, J. D. (1993). Social stories: Improving responses of students with autism
5
6
7 with accurate social information. Focus on Autistic Behavior, 8(1), 1-10.
8
9 Grover, S., Pea, R., & Cooper, S. (2015). Designing for deeper learning in a blended computer
10
11
12 science course for middle school students. Computer Science Education, 25(2), 199-237.
13
14 Hesse, F., Care, E., Buder, J., Sassenberg, K., & Griffin, P. (2015). A framework for teachable
15
16
17 collaborative problem solving skills. In Assessment and teaching of 21st century skills
18
19 (pp. 37-56). Springer Netherlands. doi: 10.1007/978-94-017-9395-7_2
20
21 Israel, M., Pearson, J. N., Tapia, T., Wherfel, Q. M., & Reese, G. (2015). Supporting all learners
22
23
24 in school-wide computational thinking: A cross-case qualitative analysis. Computers &
25
26 Education, 82, 263-279.
27
28
29 Israel, M., Wherfel, Q. M., Shehab, S., Ramos, E. A., Metzger, A., & Reese, G. C. (2016).
30
31 Assessing collaborative computing: Development of the Collaborative-Computing
32
33
34 Observation Instrument (C-COI). Computer Science Education, 26(2-3), 208-233.
35
36 Jahromi, L. B., Bryce, C. I., & Swanson, J. (2013). The importance of self-regulation for the
37
38
39
school and peer engagement of children with high-functioning autism. Research in
40
41 Autism Spectrum Disorders, 7, 235-246.
42
43 Jarvis, S., & Seifert, T. (2002). Work avoidance as a manifestation of hostility, helplessness or
44
45
46 boredom. Alberta Journal of Educational Research, 48, 174-187.
47
48 Karoly, P. (1993). Mechanisms of self-regulation: A synthesis view. Annual Review of
49
50
51 Psychology, 44, 23-52.
52
53 K-12 Computer Science Framework (2016). Retrieved from https://k12cs.org/
54
55
56
57
58
59
60
61
62
63
64
65
46
1
3
4 King, S. A., Lemons, C. J., & Davidson, K. A. (2016). Math interventions for students with
5
6
7 autism spectrum disorder: A best-evidence synthesis. Exceptional Children, 82(4), 443-
8
9 462.
10
11
12 Koegel, L. K., Vernon, T. W., Koegel, R. L., Koegel, B. L., & Paullin, A. W. (2012). Improving
13
14 social engagement and initiations between children with autism spectrum disorder and
15
16
17 their peers in inclusive settings. Journal of Positive Behavior Interventions, 14(4), 220-
18
19 227.
20
21 Kollar, I., Fischer, F., & Hesse, F. W. (2006). Collaboration scripts—a conceptual analysis.
22
23
24 Educational Psychology Review, 18(2), 159–185. doi:10.1007/s10648-006-9007-2
25
26 Lee, I. (2016). Reclaiming the voice of CT. CSTA Voice, 12(1), 3-4. Retrieved from
27
28
29 http://c.ymcdn.com/sites/www.csteachers.org/resource/resmgr/Voice/csta_voice_03_201
30
31 6.pdf
32
33
34 Lee, I., Martin, F., & Apone, K. (2014). Integrating computational thinking across the K-8
35
36 curriculum. ACM Inroads, 5(4), 64-71.
37
38
39
Lewis, T. J., Scott, T. M., Wehby, J. H., & Wills, H. P. (2014). Direct observation of teacher and
40
41 student behavior in school settings: Trends, issues and future directions. Behavioral
42
43 Disorders, 190-200.
44
45
46 McGee, G. G., Feldman, R. S., & Morrier, M. J. (1997). Benchmarks of social treatment for
47
48 children with autism. Journal of Autism and Developmental Disorders, 27(4), 353-364.
49
50
51 doi:10.1023/A:1025849220209
52
53 Mercer, N. (2008). The seeds of time: Why classroom dialogue needs a temporal analysis. The
54
55
56
Journal of the Learning Sciences, 17(1), 33-59.
57
58 Papert, S. (1980). Mindstorms: Children, computers, and powerful ideas. Basic Books, Inc.
59
60
61
62
63
64
65
47
1
3
4 Park, M., & Lash, T. (2014). The Collaborative Discussion Framework. Champaign Unit 4
5
6
7 School District.
8
9 Prottsman, K. (2014). Computer science for the elementary classroom. ACM Inroads, 5(4), 60-
10
11
12 63.
13
14 Rao, S. M., & Gagie, B. (2006). Learning through seeing and doing: Visual supports for children
15
16
17 with autism. Teaching Exceptional Children, 38(6), 26-33.
18
19 Ratcliff, C. C., & Anderson, S. E. (2011). Reviving the turtle: Exploring the use of Logo with
20
21 students with mild disabilities. Computers in the Schools, 28(3), 241-255.
22
23
24 Rockwell, S. B., Griffin, C. C., & Jones, H. A. (2011). Schema-based strategy instruction in
25
26 mathematics and word problem-solving performance of a student with autism. Focus on
27
28
29 Autism and Other Developmental Disabilities, 26, 87-95. doi:
30
31 10.1177/1088357611405039
32
33
34 Schunk, D. H., & Zimmerman, B. J. (Eds.). (1994). Self-regulation of learning and performance:
35
36 Issues and educational applications. Hillsdale, NJ: Erlbaum.
37
38
39
Snodgrass, M. R., Israel, M., & Reese, G. (2016). Instructional supports for students with
40
41 disabilities in K-5 computing: Findings from a cross-case analysis. Computers &
42
43 Education. 100, 1-17.
44
45
46 Stake, R. E. (1994). Case Studies in N. K. Denzin and Y. S. Lincoln (Eds.): Handbook of
47
48 qualitative research, (p.p. 236-247). Thousand Oaks: Sage.
49
50
51 Stake, R. E. (2006). Cross-case analysis. In R. Stake, Multiple case study analysis. (pp. 39-77).
52
53 New York: Guilford.
54
55
56
Tatsuoka, M. M. (1986). Graph theory and its applications in educational research: A review and
57
58 integration. Review of Educational Research, 56, 291-329.
59
60
61
62
63
64
65
48
1
3
4 U.S. Department of Education, Office of Special Education Programs, Individuals with
5
6
7 Disabilities Education Act (IDEA) database, retrieved from
8
9 http://www2.ed.gov/programs/osepidea/618-data/state-level-data-files/index.html#bcc.
10
11
12 Vogel, F., Wecker, C., Kollar, I., & Fischer, F. (2016). Socio-cognitive scaffolding with
13
14 computer-supported collaboration scripts: a Meta-analysis. Educational Psychology
15
16
17 Review, 1-35.
18
19 Webb, N. M., Franke, M. L., De, T., Chan, A. G., Freund, D., Shein, P., & Melkonian, D. K.
20
21 (2009). Explain to your partner: Teachers’ instructional practices and students’ dialogue
22
23
24 in small groups. Cambridge Journal of Education, 39(1), 49–70.
25
26 doi:10.1080/03057640802701986
27
28
29 Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
30
31 Psychological Review, 92, 548-573.
32
33
34 Weiner, B. (2010). The development of an attribution-based theory of motivation: A history of
35
36 ideas. Educational Psychologist, 45, 28-36. doi: 10.1080/00461520903433596
37
38
39
Wei, X., Christiano, E. R., Yu, J. W., Wagner, M., & Spiker, D. (2014). Reading and math
40
41 achievement profiles and longitudinal growth trajectories of children with an autism
42
43 spectrum disorder. Autism: The International Journal of Research and Practice, 19, 200-
44
45
46 210.
47
48 Wing, J. M. (2006). Computational thinking. Communications of the ACM, 49(3), 33-35.
49
50
51 Yin, R. K. (2009). Analyzing case study evidence: How to start your analysis, your analytic
52
53 choices, and how they work. In R. Yin, Case study research: Design and methods (4th
54
55
56
ed) (pp. 126-163). Thousand Oaks, CA: Sage.
57
58
59
60
61
62
63
64
65
*Highlights (for review)
Highlights
 Elementary students with Autism Spectrum Disorder engaged in computing instruction.

 Students learned a collaborative conversation script for collaborative problem solving.
 Students were engaged although frustration and limited success were present.
 Instructional strategies for students with disabilities in computing are needed.

Cae D 17 00325

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Cae D 17 00325

Enviado por

Direitos autorais:

Formatos disponíveis

Computers & Education

Manuscript Number: CAE-D-17-00325

Title: Engagement and collaboration of students with autism during

Article Type: Research Paper

Keywords: Elementary computer science;

Abstract: Elementary schools have increasingly added computer science

1 Running Head: ENGAGEMENT OF STUDENTS WITH AUTISM IN CS EDUCATION

 Elementary students with Autism Spectrum Disorder engaged in computing instruction.

Você também pode gostar