International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Child-Computer Interaction

journal homepage: www.elsevier.com/locate/ijcci

Case report

Autism and tablet computers in Turkey: Teaching picture sequencingskills via a web-based iPad application

Q1 Ceymi Doenyas a,∗, Emre Şimdi b, Ezgi Çagla Özcan b, Zehra Çataltepe b,Binyamin Birkan c,d

a Princeton University, NJ, United Statesb Istanbul Technical University, Istanbul, Turkeyc TOHUM Autism Foundation, Istanbul, Turkeyd Istanbul Aydın University, Istanbul, Turkey

a r t i c l e i n f o

Article history:Received 12 July 2013Received in revised form6 April 2014Accepted 25 April 2014

Keywords:AutismTurkeyASDABAiPadApplicationSequencingCard gameTechnologyInteractionWeb-basedHTML5

a b s t r a c t

We conducted the first study on Turkish children with autism and tablet computers, with a web-basediPad application designed especially for them. We performed a pilot study on three Turkish boys ofdifferent ages with autism to observe their reactions to the tablet application and its effectiveness inteaching the sequencing skill, which is part of their educational curriculum. Our application had a testingsession with no prompts or rewards and a teaching session with prompts, rewards, and demonstration ofcorrect responses. First, our participants played the testing session to determine their baseline sequencingabilities. Next, they played the teaching session. Finally, they played the testing session again to see ifthey were now able to sequence the cards on their own. Through this application, the 11-year-old boy’ssequencing skills improved without external help, via only the prompts and reinforcements of the iPadapplication. The applicationwas not enough to teach sequencing to the 4-year-old, who required externalhelp, and it was too simple for the 15-year-old, who did not use any prompts and quickly became bored.Based on our findings, we discuss how to improve similar sequencing applications and offer suggestionsfor designing iPad applications for individuals with autism.

© 2014 Published by Elsevier B.V.

1. Introduction1

Individuals with autism are impaired in social interaction and2

communication, and have repetitive, restricted, and stereotypedQ23

behavior patterns and interests [1]. Early diagnosis and effective4

interventions through structured programs and educational meth-5

ods result in significant progress in autism [2]. Autism is universal.6

Individuals with autism show similar behavioral impairments and7

share similar learning characteristics all around the world. Infor-8

mation about new methods in the education of individuals with9

autism from one country can help those in another and is invalu-10

able for international generalization of new technologies that are11

available in today’s world.12

The Turkish government organized a project named Fatih Pro-13

jesi to deliver tablet computers to public schools and purchased the14

∗ Corresponding author.E-mail address: ceymidoenyas@gmail.com (C. Doenyas).

tablets in November of 2013 (relevant press release can be found 15

in http://fatihprojesi.meb.gov.tr/tr/haberincele.php?id=108). For 16

students with autism who attend public schools to benefit from 17

this project as well, there should be Turkish tablet applications de- 18

signed specially for the needs of these individuals. Moreover, the 19

United States Embassy in Ankara is conducting a project to develop 20

a Turkish iPad application to be used as a voice output commu- 21

nication aid, i.e. voice what the children with autism who have 22

problems with speaking wish to say. Yet, there is no information 23

currently available in Turkey on how Turkish children with autism 24

will interact with tablet applications designed for them. So, we 25

have performed a pilot study to code and test a tablet application in 26

Turkish for a skill that is in the educational curriculum of children 27

with autism. The purpose of this study was to observe the reaction 28

and interest of Turkish children with autism to educational iPad 29

applications and see whether they will be able to learn from such 30

an application. 31

This is the first study that has designed a Turkish application 32

for children with autism and has tested it on Turkish children with 33

http://dx.doi.org/10.1016/j.ijcci.2014.04.0022212-8689/© 2014 Published by Elsevier B.V.

2 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

autism.We carried out this pilot, data-based study since in Turkey,1

we currently do not have available educators for a period of time2

that is required for a single-subject experimental design and the3

time constraints on the curriculum of the participating children4

with autismdo not allow for such an extensive design. Our purpose5

was to open the way for the coding of such applications in Turkey6

and provide information to help design controlled experimental7

studies in the future when the resources will be available.8

This study is important in two ways: 1. to provide useful9

information about how Turkish children with autism react to the10

tablet application, if they are able to learn from it, and how the11

future applications should be developed, and 2. to lead the way12

for future studies that will allow cross-cultural comparison of the13

interaction of individuals with autism and tablet computers in14

Turkey and other countries.15

Why iPads? There are several iPad applications that can be used16

by both individuals with and without autism, some flashcards17

and images that could be used within Applied Behavioral Analysis18

(ABA) therapy (http://www.autismspeaks.org/autism-apps), and19

several iPad applications that are intended for individuals with20

autism. Because the use of mobile technology by individuals with21

autism is still quite new, experts call for more in depth research22

in this field (Yee 2012) and the purpose of this paper is to show23

how Turkish children with autism interact with tablet computers24

and whether they are able to learn from a tablet application. We25

selected a skill that our participants needed to learn and aimed26

to teach it through the tablet computer technology. We have27

conducted our tests in Tohum School for Autism in Istanbul, the28

only school for individuals with autism in Turkey that applies ABA29

principles.30

Previous experiments show that teaching concepts to children31

with autism through technological devices has many advantages32

over the traditional methods. Children with autism were more33

attentive, more motivated, and learned more vocabulary with34

educational software programs of a computer than with the35

behavioral program [3]. Adding computerized games to the regular36

language therapy sessions produced the following results: children37

were more attentive, had more communicative initiatives and38

more eye contact,weremore interactive, usedmore verbalizations,39

and made more action requests [4]. Given these advantages of40

technology, the portable and engaging nature of iPads and their41

easy-to-use touchscreens that provide instant gratification for42

children [5], we used iPads as our teaching medium.43

Why sequencing? The reason we chose to design an application44

for the skill of sequencing is because sequencing events in a story45

is an important expressive language skill that is frequently absent46

in children with autism. It is one of the basic skills underlying47

communication, reading, and speaking, but there is no available48

evidence showing how the development of sequencing skills in49

children with autism can be assisted by an iPad application.50

A lack of sequencing story events skills limits the use of lan-51

guage that promotes interactions, which results in a further de-52

crease in opportunities to engage in a meaningful communication53

experience. For example, when an adult initiates a conversation54

with a child (e.g., tell me what you did today), the child may not55

respond if he/she does not know how to sequence in his mind and56

explain the events he/she was engaged in during the day in the or-57

der they happened.58

Studies focusing on teaching sequencing story events skills to59

children with autism have reported favorable language develop-60

ment gains. Increase in sequencing story events skills has resulted61

in increases in social interactions, positive affect, and spontaneous62

speech [6], as well as improvements in expressive language and63

social-communicative behaviors [7]. Since children with autism64

are already at a communication disadvantage, these findings have65

led to the belief that sequencing story events should be a core66

component of any intervention program for young children with 67

autism [8]. 68

Learning how to sequence story events helps children form 69

connections between different stages of an event. For individuals 70

with autism, it is hard to assemble the pieces from each stage 71

and see the event as a whole. Sequencing game cards are used in 72

the education of individuals with autism since they help children 73

analyze pictures, put the events in the correct order, and predict 74

the consequences of actions. If the individual does not understand 75

the sequence of events, then he/she cannot narrate the event. 76

Sequencing activities also help children with autism acquire 77

the prerequisite skills for reading and writing. In order to sort the 78

images, children should be able to think of a story in pieces and 79

distinguish the differences between its stages. The ability to notice 80

the differences and understandwhat the images represent is one of 81

the prerequisite skills for reading. In order to read, children should 82

be able to differentiate the letters, i.e. realize that a is different 83

from b. The skill of telling apart different photographs depicting 84

distinct stages of an event prepares the student to differentiate 85

letters and words later on, which are more abstract than pictures. 86

Thus, the skills of reading and distinguishing different words that 87

are written or spoken build on the skill of assigning meaning and 88

distinguishing different stages of an event, which is acquired by 89

sequencing story event photographs [8]. 90

Applied behavioral analysis (ABA). Applied Behavioral Analy- 91

sis (ABA) is a scientific method that investigates environmental 92

variables influencing socially important behaviors and uses those 93

findings to implement interventions thatwill improve such behav- 94

iors [9]. ABA principles posit that immediate consequences have 95

the largest effects, and that reinforcement is key in behavioral 96

change programs [9]. These ideas are employed in ABA-based be- 97

havioral modification programs for individuals with autism. 98

The first phase of ABA starts with evaluating the individual’s 99

behaviors and skills. In this stage, target behaviors need to be 100

analyzed in detail. Skills are broken down into simple steps and 101

every session consists of discrete trials that aim to teach the indi- 102

vidual that specific skill. In the teaching session, first an instruc- 103

tion is presented to the individual in a clear and consistent tone. 104

After waiting for a response for a while, if the correct response is 105

not given, the individual is directed to the desired response by pro- 106

viding prompts. ABA, being a faultless teaching approach, aims to 107

prevent individuals from giving erroneous answers, so prompting 108

the individual to perform the desired behavior is an important part 109

of this method. These prompts can be provided as visual cues, such 110

as pointing out the correct answer or fading out thewrong options. 111

Prompts need to be repeated until the desired behavior is per- 112

formed and reinforcement should immediately follow the correct 113

response tomake the child understand the connection between the 114

correct answer and the reward. These steps are repeated by fading 115

the prompts and the reinforcers gradually as the child begins to 116

perform better [10]. In our application, we have used the prompt 117

and prompt-fading procedures of ABA and we have designed the 118

testing and teaching versions of the application according to the 119

ABA steps described above. 120

Relevant iPad applications. Below is a list and short description 121

of the existing iPad applications that are in some way related to 122

the sequencing game we set out to design for Turkish children 123

with autism. After analyzing what is missing in those applications 124

with regard to the ABA principles, we designed our application 125

specifically for the education of children with autism. 126

• ‘‘Speech with Milo - Sequencing’’ intends to build sequencing 127

and narrative skills for individuals who need special education. 128

In the application, the player gets three different sequencing 129

cards and needs to put them in the correct places. The 130

application has a rich settings page that allows users to toggle 131

on and off the success and unsuccessful sounds, background 132

C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx 3

music, andwritten and spokenwords. If thewritten and spoken1

words are turned on, the application gives a sentence that2

explains the event, such as ‘‘driving a car’’. In the beginning, the3

application gives the prompt ‘‘put the cards in order’’ in both4

written and spoken form. The figures on the cards are cartoon5

characters and the background space of the cards is very busy,6

which makes it hard to focus on and distinguish the actions in7

the frames. If the user chooses to get a hint, then each card is8

given with a phrasal explanation written below, but since there9

is no sound option for the card explanations, this prompting10

strategy is not the best one to help the children who cannot11

read. If the child tries to drag the card to the wrong place, he is12

not allowed to place it there, which prevents him from making13

a mistake, as is recommended in ABA. Yet, there are no hints14

indicating the card’s correct location to help the child get to15

the correct response. When the card is placed in the correct16

place, there is a clapping sound, which can serve as positive17

reinforcement for the correct behavior. Lastly, the application18

does not record the data of the user, which would have been19

beneficial to keep track of the improvements.20

• ‘‘Tell a Story with Tommy: Community Sequences’’ is mainly21

focused on the development of story telling and sequencing22

skills for all children. It has seven story options: go home, the23

car, library, pool, school, visit, and grocery. At the beginning of24

the game, the story is told with images. There are four images25

with cartoon characters that appear on the screen one at a time,26

accompanied by a few sentences explaining each image, which27

are written on the screen and are also narrated. The sentences28

and the narration can be turned on and off. After that, four29

different cards are given to the child in a random order and the30

child is expected to sequence them. The child is not allowed31

to perform a wrong move and the narrator’s voice encourages32

the child to try again (‘‘try again’’, ‘‘why don’t you try again’’,33

‘‘okay, not quite’’, ‘‘give it another shot’’, ‘‘one more try’’). The34

voice appreciates the childwhen she performs the correctmove35

(‘‘nice job, ‘‘right on’’, ‘‘way to go’’, ‘‘you’ve got it’’). When the36

child is able to complete the sequencing of all four cards, there37

appears a dancing kid on the screen as a reward. There are38

verbal reinforcers in the game, which have a lot of variation, but39

there is a lack of visual prompts that guide the player towards40

the correct move. This application does not track the user’s41

performance either.42

• ‘‘Making Sequences’’ allows the selection of the game that the43

user wants to play (going to the beach, sending a letter, blowing44

up a balloon etc.) as well as the creation of a new personal45

one by adding the user’s own images and sounds. Since actual46

photography with simple background and real people is used47

in this application, frames are suitable for children with autism.48

There is a ‘‘watch’’ option to listen to the stories before starting49

the ‘‘place in order’’ section. The main consideration is that,50

during the game, the user needs to slide the cards over each51

other in order to put them into the correct places. Numbers are52

placed above the photographs as a cue for the sequencing order.53

As a result, a player who knows the order of the numbers can54

sequence the actions correctlywithout paying attention towhat55

is depicted in the cards. In addition, until the ‘‘Done’’ button is56

hit, there are no clues as to whether the performed sequence is57

right or not. These features can be confusing for children with58

autism. During the game, when the child selects the ‘‘listen to59

story’’ option, the cards are narrated in the order that the child60

places them in, not necessarily in the correct order, which can61

cause the children with autism to learn an incorrect sequence62

as a correct one.63

• ‘‘iSequences’’ is especially designed for children with autism64

or children with other special educational needs. It aims to65

teach autonomous habits (such as brushing teeth, washing66

hands or getting dressed), leisure activities (such as going to 67

the movies or doing sports), daily activities (such as doing 68

laundry or going to school) and the emotions following the 69

sequenced events (such as happiness or surprise). There are a 70

hundred different sequencing games and each game includes 71

three or four cards. The application has many setting options 72

to customize the game, including five different languages. 73

Cartoon characters are used in this application. When the child 74

completes the sequencing successfully, the game rewards her 75

with an applause and a reward picture. There is a different 76

reward picture at the end of each sequence, a variety that 77

engages the player. This application is well designed in terms of 78

the diversity of images and activities, which is likely to prevent 79

boredom and enable interested interactionwith the application 80

for a long time. There are no prompts to guide the child to the 81

right action. The applicationwaits until the last card is placed to 82

indicate whether the sequencing is correct or not, and lets the 83

child place the cards in incorrect locations, so there is a lack of 84

immediate feedback to the user. 85

• ‘‘Sequencing Tasks - Life Skills’’ is another sequencing applica- 86

tion that uses simple, actual photographs, which helps the child 87

to focus on the actions in the frames. Before the sequencing, the 88

whole activity is described with a video. Then, each frame that 89

is used in the sequencing is narrated individually. The player is 90

not allowed to put the card in an incorrect place but there is no 91

cue to lead him to the card’s correct location. When the card 92

is placed in the correct location, the application rewards the 93

player with a ring sound and by saying ‘‘wonderful’’, ‘‘perfect’’, 94

‘‘excellent’’, and ‘‘that’s right’’. When the player tries to put a 95

card to an incorrect location, the application says ‘‘this doesn’t 96

go here, try again’’ and moves the card back. When each card 97

is touched, the activity depicted in it is narrated, which may be 98

a helpful cue. There is a ‘‘Statistics’’ option, which shows the 99

activity history that includes the date and time the application 100

was played, name of the activity, and the total number of incor- 101

rect placements during each session. This is useful to track the 102

improvement of the player. 103

Evaluation of the existing sequencing applications.After analyzing 104

the properties of the existing applications listed above, we decided 105

on the following guidelines for our application: 106

• It seems that most of these applications and other educational 107

applications for children with special needs do not have a real- 108

time cue system that teaches the child how to play the game. 109

Rather, the cue is given beforehand and a correct move is ex- 110

pected from the child or if no cue is given, the child is not 111

allowed to make an incorrect move. Prompting is very im- 112

portant for learning and it is advised that prompting should 113

be increased or decreased depending on the child’s perfor- 114

mance [11]. Thus, we decided to implement a real-time cue sys- 115

tem in our game, which adjusts the level of the cues according 116

to the performance of the child. 117

• Some applications use cartoon characters and some use real 118

photographs. A decade of experiencewith Turkish childrenwith 119

autism and a total of 23 years of experience with individu- 120

als with autism show that they learn more easily and better 121

from real characters compared to cartoon characters, which 122

seem to be too complicated for the children to make con- 123

nections with real life. This is also remarked by autism ed- 124

ucators on online sources, such as The Autism Helper, who 125

say that cartoon depictions can sometimes be too abstract for 126

children with autism and advise the use of real photographs 127

(http://theautismhelper.com/real-photo-pecs). This is the rea- 128

son behind our choice to use pictures of human models in our 129

application. We used two different characters instead of the 130

same person performing all the actions in order to better enable 131

4 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

Table 1Advantages and disadvantages of HTML5, IOS, and Android application platforms.

HTML5 IOS ANDROID

Advantages

• Application works on every platform • High speed native application • High speed native application• Promising Future Technology • Does not require internet

connection/web server• Does not require internet connection/webserver

• Higher speed than earlier web technologies • Attractive User Interface • Cost friendly to learn, code and maintain• Open sourced, available libraries on internet(Jquery, etc.)

• Attractive User Interface

• Easier and cost friendly to learn, code andmaintain

Disadvantages

• Relatively less speed comparing to the nativeapplications

• Need to get MAC for coding • Application only works in Android platform

• May require internet connection • Need to get iPad for testing• Not stable, frequently being updated • Need to pay in order to become official

developer• Application only works on Appleplatform

generalization across different people. Yet, we are aware that1

further tests should be performed on the comparative effective-2

ness of teaching from real vs. cartoon characters, which will be3

the focus of a future study on Turkish children with autism.4

• Wedidnot put anynumbers or letters on the pictures to prevent5

the children from paying attention solely to those cues and ig-6

noring the event depicted in the photographs that are supposed7

to be sequenced.8

• Moreover, all of the applications described above allow the9

player to start sequencing from any card they want. Based on10

long years of experience with children with autism in Tohum11

School, the educators and the principal recommended that our12

application should only allow the player to start sequencing13

from the first card. This is important because when typically14

developing individuals are engaged in a dialog in real life, they15

start explaining an event from its beginning, not from itsmiddle16

or end. This makes the story easier to understand and it helps17

the storyteller remember the continuation of the story, since18

the beginning of a story serves as a cue to what comes next.19

We are designing this application to improve real life commu-20

nication skills of individuals with autism and teaching them to21

begin with the first event in the sequence is an important step22

in helping them acquire the skill of talking about their daily life23

events to others.24

• In addition, none of these games except one keeps track of the25

player’s performance. In schools for children with autism, for26

every task in the curriculum, the instructors collect data prior27

to, during, and after learning to see whether the student was28

able to learn successfully. In order to see the development of29

children’s abilities, we implemented a performance tracking30

and data recording system to collect the results of this sequenc-31

ing game. Different from the data collection in ‘‘Sequencing32

Tasks—Life Skills’’ application, our application tracks the per-33

formance for each card and the overall performance.34

• We have decided to code our application as an educational35

game, defined as having a balance of educational and entertain-36

ment value (Prensky, 2001, as cited in [12]). Educational video37

games comprise rewards and interactive cues that offer feed-38

back and induce learning [13], qualities we have included in our39

application.We decided that our application should have a test-40

ing session to measure the skill level prior to and after learning,41

and a teaching session to teach the sequencing skill. The testing42

and teaching sessions are parts of our application, which is an43

educational game.44

Choosing the platform of HTML5. We could have coded our45

application for IOS, an operating system that runs on Apple46

products, or for Android, a mobile operating system that runs47

on devices other than Apple products. Or, we could have chosen48

to code a Web-Based application, HTML5, which is described in 49

detail by Vaughan-Nichols [14]. Afterweighing the advantages and 50

the disadvantages of these options, we decided to use HTML5. 51

The most important reason for our choice was the flexibility that 52

allowed the application to be available on all platforms, such as 53

PCs, different brands of mobile phones, and tablets. In Table 1, the 54

advantages and disadvantages of HTML5 and the other platforms 55

are listed. 56

2. Material and methods 57

2.1. Study design 58

This was a pilot study to observe the suitability of an Applied 59

Behavioral Analysis based iPad application for different aged 60

Turkish children with autism. We also aimed to see whether the 61

prompts and cues of the application would be enough to teach 62

the skill of sequencing without any external prompting from the 63

teachers. We used an ABA design over three participants; A was 64

our baseline condition where we used the testing session of the 65

iPad application that included no prompts, B was our intervention 66

condition where we used the teaching session of the application 67

that has prompts and reinforcers and used teaching with external 68

assistance to one participant who required it, and A was our 69

baseline condition that we used again post-intervention to see the 70

effectiveness of the teaching, where we used the testing session 71

that had no prompts. 72

We chose 3 Turkish boys of different ages with autism who 73

were ready to learn the skill of card sequencing. They were 4, 11, 74

and 15 years old. As opposed to regular participant pools where 75

an experimental and a control group can be established, with 76

children with autism we cannot form homogeneous experimental 77

and control groups because even when the children have the 78

same diagnosis of autism, the same gender, and the same age, 79

their academic incompetencies and the degree to which they are 80

affected by autism highly differ. So, we tested the application on 81

each child, and compared their performance after the learning 82

to their performance prior to it, instead of comparing their 83

performance to a control group of children with autism who did 84

not use the application or who used paper-based methods to learn 85

sequencing. 86

2.2. Participants 87

The inclusion criteria for our participants were the following: 88

• Owning a tablet computer 89

• Possession of the following prerequisite skills: (1) understand- 90

ing and following simple instructions, (2) adequate level of fine

C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx 5

Fig. 1. Preferences, where the tester selects the teaching or testing session.

motor skills, (3) picture identification skills, and (4) adequate1

on-task engagement during daily activities2

• Different age groups3

• Approval of the school and their parents4

• Need to learn the sequencing skill5

The participants were diagnosed with autism according to DSM6

IV criteria, which was the available diagnosing instrument at the7

time the testing took place. The diagnosis was given through the8

direct observations of child psychiatrists and behavioral analysts.9

Wehave received parental consent for each participant prior to the10

study.11

Cihan is a 15 years and 9month old boy diagnosed with autism.12

He has been participating in the educational curriculum of TohumQ313

for 7 years. He started the school section of Tohum when he was14

10 years old. Cihan has reading and writing skills, and his skill15

of understanding what he reads is improving. He is working on16

mathematical problems that require addition and subtraction. In17

terms of his expressive language, he is able to hold short conver-18

sations when he is informed about their topic beforehand. He is19

able to record his daily activities in his phone in order, and follow20

his schedule independently. He is also attending some classes in a21

music-focused high school. In Turkey, the translated and adapted22

version of the Wechsler Intelligence Scale for Children-Revised23

(WISC-R), [15] is used to measure the intelligence of children and24

this test was administered to measure Cihan’s IQ. Cihan’s verbal IQ25

is 50, his performance IQ is 96, and his total IQ is 70.26

Oğuz is an 11 years and 6month old boy diagnosedwith autism.27

He has been a student in the school section of Tohum for 6 years.28

Oğuz has reading and writing skills, and he is improving himself in29

terms of understanding what he reads. He is working on addition30

problems inmathematics.With the help of pre-determined scripts,31

he is able to hold short conversations about certain topics. Oğuz32

is able to follow his activities in order from a written activity33

schedule. Oğuz is attending a general educational school as part34

of an inclusion program. Oğuz’s IQ was measured using the WISC-35

R intelligence test [15]. Oğuz’s performance IQ is 42 and his verbal36

IQ and total IQ could not be computed because he was not able to37

provide sufficient responses.38

Yavuz is a 4 years and 9 month old boy diagnosed with autism.39

He has been a student in Tohum for 2 years. He is practicing40

following prompts that describe two actions. He is working on41

line drawing as preparation for acquiring writing skills. For fluent42

language skills, he is studying color recognition, and for expressive43

language skills, he is working on the skill of initiating interactions.44

He is able to independently follow activity schedules that have45

pictures. Denver II Developmental Screening Test was performed46

on Yavuz and his test results are abnormal. He has received a total 47

of 11G and 7U points in the areas of social, fine and gross motor, 48

and language skills. G denotes the areas that are lagging behind 49

in development when compared to younger individuals, and U 50

denotes areas that are lacking in performance when compared to 51

peers. 52

These participants were selected because theywere all ready to 53

learn the skill of sequencing. All three participants have the ability 54

of differentiating pictures and objects. Yavuz and Oğuz have not 55

practiced sequencing actions in traditional paper methods before. 56

Cihan had some prior practice ordering photographs. Although 57

none of them had an iPad at school, Cihan and Oğuz have been 58

using iTouchs, so they have prior experience with touch screen. All 59

of them received iPad donations three weeks before testing began 60

and started using iPads for their rewards and free times. All of them 61

have the ability of holding, grabbing, and carrying objects from 62

one place to another in real life. Cihan and Oğuz are able to tap 63

on, slide, and move pictures on touch screen. Yavuz just started 64

solving puzzles and playing instruments on his iPad, and his skills 65

of holding and dragging items on touch screen are improving. 66

2.3. The iPad application 67

2.3.1. User interface 68

In our sequencing card game application, after verifying the 69

user and selecting the student, the ‘‘Preferences’’ page appears 70

on the screen. There are two sessions: ‘‘Testing’’ and ‘‘Teaching’’ 71

(Fig. 1). After choosing the session, the child starts to play the se- 72

quencing game (Fig. 2). 73

There are 5 different sequencing stories: brushing teeth,making 74

a sandwich, getting on and riding a bike, taking a jacket out of the 75

closet and putting it on, and pouring orange juice into a cup and 76

drinking it. The child plays each activity two times in a random 77

order in each session. The child is required to put the cards in the 78

correct order by pulling them down. 79

In the testing session, the child gets a chance to play the game 80

on his ownwithout any prompts or reinforcers because the testing 81

session aims tomeasure the child’s independent performance. Still, 82

the child is not allowed to place the card in an incorrect location so 83

that he will not assume that it is the correct sequence and learn 84

it as such. If the child tries to place a card in an incorrect location, 85

then the card bounces back to its starting location and does not 86

get placed in the incorrect location. If the child cannot put the 87

card to the correct location within a certain amount of time, then 88

the card is placed into the correct location automatically without 89

any prompts and the child is considered to be unsuccessful in its 90

6 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

Fig. 2. During the sessions, the child is expected to pull the cards down in the correct order.

placement. There are no reinforcers when the child is able to place1

the cards in their correct location. If the child is able to put all three2

cards in their correct places, then he gets a feedback in the form of3

a video showing the sequence of actions accompanied by different4

theme songs from cartoons and musicals such as Disney movies5

and Grease. Video modeling was shown to be effective in teaching6

certain tasks to students with disabilities [16]. The skill that we7

aim to teach is sequencing and it is important for the child to see8

his end product when he sequences the cards correctly, just as he9

would see what he has produced if we were using paper-based10

methods. Thus, he does not just look at the three steps separately,11

but understands how they come together to produce the whole12

event.13

In the teaching session, the child plays the game with prompts14

and reinforcers. If he cannot perform the desired action, then the15

card to be moved next is highlighted and starts blinking to give a16

cue. If the child cannot make the correct response within 6 s after17

this cue, then the same cue repeats. If the child still cannot make18

the correct response, then all the other cards fade. After 6 more19

seconds of no correct response, the card is placed in its location20

automatically. If the child places the card in the correct location at21

any time before it is placed automatically, then there is a ‘‘Bravo!’’22

sound accompanied by clapping in order to reinforce the child’s23

correct behavior. The game does not let the child put a card in24

an incorrect location. When each sequencing game is finished, the25

child watches the video evenwhen he is not able to place the cards26

correctly because seeing the complete event will help his learning.27

Werecord the cardmoves produced by the child and their time-out28

counts.29

2.3.2. System architecture30

Different programming languages were used for the develop-31

ment of the application. The User Interface was coded by HTML32

technology. With the new features of HTML5, by using CSS3, we33

have given a child-friendly user interface to the application. More-34

over, using Jquery let us have impressive graphics, such as iPhone35

styled buttons. In addition to the HTML, CSS3 and Jquery technolo-36

gies, we usedMySQL for retrieving data from the database and PHP37

for manipulating MySQL. In order to provide a smooth user expe-38

rience [17], we implemented the AJAX technology (Fig. 3).39

2.3.3. Performance recording40

We have different performance recording parameters for each41

session of the game. In the teaching session, we use a real-time42

cue system (Fig. 4). After the instruction of the game, for every 6 s43

that the child cannot perform the desired action, a time-out occurs.44

Fig. 3. System architecture.

After the first and the second time-outs, the card to be sequenced 45

is highlighted with a red frame and it starts blinking. If the child 46

still cannot put the card to the right place, by the time the third 47

time-out occurs, then all other cards fade out. In case of a fourth 48

time-out, the card is placed in its correct location automatically. 49

According to the performance of the child during the game, 50

counting the time-outs, we developed a point-based performance 51

recording system: 52

• If the child placed the card before the 1st time-out = 5 points 53

• If the child placed the card after the 1st time-out = 4 points 54

• If the child placed the card after the 2nd time-out = 3 points 55

• If the child placed the card after the 3rd time-out = 2 points 56

• If the child fails to place the card after the 4th time-out = 1 57

point 58

In the testing session (Fig. 5), we check whether the child can 59

correctly sequence the cards or not. The same time-out system 60

exists in this session; but only one time-out is applicable for each 61

card. When the time-out occurs, instead of a cue, the card is 62

automatically sequenced. If the child performs the correct action, 63

he is given 5 points, and if he cannot, he gets 1 point. 64

2.3.4. Data collection 65

We collected data over a 2-week period in Tohum School for 66

children with autism. Behavioral analyst and autism professor Dr. 67

Binyamin Birkan, who is the principal of Tohum School, trained 68

the teachers to guide their participants to sit at a table and 69

attend to the iPad application. CeymiDoenyas andBinyaminBirkan 70

observed and supervised the tests. Ezgi Özcan, Emre Şimdi, and 71

Zehra Çataltepe collected and interpreted the data. 72

No outside prompts or help was given to the participants in 73

order to measure how effective the prompts and the rewards of 74

C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx 7

Fig. 4. Teaching session performance recording.

Fig. 5. Testing session performance recording.

the applicationwere in teaching the sequencing skill. However, the1

youngest participant Yavuz (4) had a very hard time understanding2

the prompts and what the application asked him to do, so after3

a few trials his teacher started guiding his hand and giving him4

external rewards for his correct actions to help him understand5

what he was supposed to be doing.6

With each participant, we aimed to do one initial testing session7

to determine their prior abilities of sequencing, three teaching8

sessions to teach them the skill, and one final testing session to9

measure how effective the teaching had been. We did this setup10

with Oğuz (11), but as the testing progressed, we had to modify11

this setup for Cihan (15) and Yavuz (4) for different reasons.12

Cihan (15) was very successful with the task, and he looked13

like he was starting to get bored when he was doing the second14

teaching session, so instead of giving him three teaching sessions,15

we gave him two, and then the testing session. Our reasoning16

behind this decision was if we gave him another teaching session17

after he looked bored, he may not have paid attention during the18

testing session that came after the third session, and this could19

have been a handicap in measuring his real ability of sequencing20

the cards in the testing session.21

After Yavuz (4)’s second teaching session, it was obvious that22

he could not understand the instructions and the cues. He would23

touch the cards, but he would not place them in the boxes below.24

With the suggestion of Dr. Binyamin Birkan, we decided to add25

external help from his teacher to assist Yavuz in understanding26

what he was asked to do and include three more teaching27

sessions with occasional external assistance. In the beginning of28

the following three teaching sessions and occasionally within each29

session as Yavuz seemed to require assistance, his teacher pointed 30

his hand towards the cards to get him to attend to them and helped 31

him pull the card to the correct location. The external prompts 32

provided to Yavuz were faded using graduated guidance; when 33

he needed assistance, his teacher guided his hand, and guidance 34

was stopped for the steps he was able to perform independently. 35

For the responses he was able to make independently, he was 36

given tangible rewards, such as food. We had Yavuz do one testing 37

session by himself, two teaching sessions with no external help, 38

three teaching sessions with partial external guidance, and a final 39

testing session by himself. 40

3. Results 41

The performance of each child is depicted in the following 42

graphic illustrations. In the testing session, if the child is able to 43

place a card in its correct location, he gets 5 points, if he is not 44

able to do so, he gets 1 point. In the teaching session, depending 45

on howmany cues the child requires before being able to place the 46

card in its correct location, he gets the following points: 5 points 47

for success without requiring any cues, 4 points for success after 48

1 cue, 3 points for success after 2 cues, 2 points for success after 49

3 cues, and 1 point when the application automatically places the 50

card because the child was unable to do so until the 4th time-out. 51

The y-axis shows the point that the child got from the sequenc- 52

ing game. The X-axis indicates the game session (testing or teach- 53

ing). Figs. 6–8 are the performance graphs of the participants for 54

each card (1st, 2nd, and 3rd cards). Figs. 9–11 show the average 55

performance graphs for each child, where we calculated the arith- 56

metic mean of the points collected from all three cards in each ses- 57

sion. The means and ranges for each testing and teaching session 58

are also calculated and listed below the graphs. 59

Yavuz’s graphs show that he was not very successful with the 60

task (Fig. 6). For the first card in each sequence (Fig. 6(a)), he scored 61

an average of 1 in the initial testing session, which indicates that 62

he was unable to place any of the first cards in their correct place 63

when he was first presented with the application. In the teaching 64

sessions without external help, Yavuz scored an average of 1.52 65

out of 5 and his scores ranged between 1–4, which means that he 66

did not score any fives. When he received occasional external help 67

in the teaching sessions, Yavuz scored an average of 3.36 out of 5, 68

and his scores ranged between 1–5. This increase in his average 69

performance did not continue in the second testing session, where 70

his average was 1. 71

8 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

(a) Yavuz card 1. (b) Yavuz card 2.

(c) Yavuz card 3.

Fig. 6. Yavuz’s performance for each card.

For the second card in each sequence (Fig. 6(b)), Yavuz1

performed an average of 1.4 in the initial testing, which is slightly2

higher than the average of his testing performance for the first card3

of each set. In teaching sessionswithno external help, Yavuz scored4

an average of 2.5, with a range of 1–5. In teaching sessions with5

occasional external help, Yavuz scored an average of 4.06, with a6

range of 2–5, meaning that he did not have any scores of 1 that7

indicate failure to place in the correct location. For the final testing,8

Yavuz scored an average of 2.1 for the second cards,which is higher9

than his performance of 1.4 in the initial testing.10

For the third card in each sequence (Fig. 6(c)), Yavuz scored11

an average of 1.76 in the initial testing, which is his highest12

performance in the initial testing among all three cards. In the13

teaching sessions without external help, Yavuz scored an average14

of 3.55 for the third card, with a range of 1–5. In the teaching15

sessions with occasional external help, Yavuz scored an average16

of 4.45 for the third card, with a range of 1–5. In the final testing,17

Yavuz scored an average of 3.6 for the third card,which is two times18

as high as his average performance in the initial testing.19

Fig. 7 illustrates Oguz’s performance for each card. For the first20

card (Fig. 7(a)), in the initial testing session, Oguz got an average of21

1.88. In the teaching session, Oguz scored an average of 4.75 out of22

5, and this increase in performance extended to the testing after23

teaching, where Oguz scored an average of 4.71. For the second24

card (Fig. 7(b)), Oguz performed relatively better on the initial25

testing, with an average of 3.22. His performance got up to 4.9626

with a range between 4 and 5 in the teaching session, and this27

improvement extended to the testing after the teaching, where28

Oguz performed an average of 4.69. For the third card (Fig. 7(c)),29

Oguz started with a high score of 4.55 in the initial testing, which30

increased to 4.96 in the teaching session, and reached 4.71 in the31

testing session after teaching.32

Fig. 8 shows that Cihan performed well for each card on both 33

testing and teaching sessions. Except for some mistakes he made 34

at the beginning of the game, his performance was almost perfect. 35

For the first card (Fig. 8(a)), Cihan started with an average of 4.2 36

out of 5 in the testing session. His performance was perfect in the 37

teaching session, with a score of 5 out of 5. In the testing after the 38

teaching session, his performance increased to 4.6 from the initial 39

4.2. For the second card (Fig. 8(b)), Cihan’s scores started with 4.6 40

in the initial testing, increased to 5 in the teaching session, and this 41

increasewasmaintained in the final testing,where his averagewas 42

also 5. For the third card (Fig. 8(c)), Cihan received the perfect score 43

of 5 for the initial testing, teaching, and the final testing sessions. 44

When we average the children’s performance in each game 45

across all three cards, we get the average performance graphs 46

depicted in Figs. 9–11. Yavuz scored an average of 1.28 in the 47

initial testing, which increased to 2.26 in the final testing after 48

learning. He scored an average of 2.50 in the teaching session with 49

no assistance, and an average of 4.00 in the teaching session with 50

external assistance. Oguz’s initial testing score of 3.22 increased to 51

4.71 in the final testing, and he scored an average of 4.88 in the 52

teaching session. Cihan’s average score of 4.60 in the initial testing 53

increased to 4.86 in the final testing, and his average score in the 54

teaching session was 5.00. 55

HTML5 related issues. Although using HTML5 provided us great 56

with flexibility for our application, we experienced some problems 57

both in the developing and testing stages. Each device and browser 58

needed special care, such as scaling the game screen or putting 59

different sound tracks due to their supporting capabilities. We 60

weremainly using IOS and even the new IOS 6 versionwas causing 61

fatal errors that were not caused by IOS 5. Thus, even though we 62

adjusted the code many times, we still had minor problems. 63

C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx 9

(a) Oguz card 1. (b) Oguz card 2.

(c) Oguz card 3.

Fig. 7. Oguz’s performance for each card.

(a) Cihan card 1. (b) Cihan card 2.

(c) Cihan card 3.

Fig. 8. Cihan’s performance for each card.

10 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

Fig. 9. Yavuz’s average performance.

Fig. 10. Oğuz’s average performance.

Considering that our target groupwas special users, special care1

had to be taken in order to assure the feasibility and efficiency of2

their iPad usage. Yavuz was pushing the screen, instead of touch-3

ing, andwas usingmore than one finger to slide a single card. These4

unexpected actionsmade the application crash and sometimes the5

browser tried to copy the cards to memory while not letting the6

user play.7

4. Discussion8

Although this is a quasi-experimental study, it is the first ever9

study on the interaction of Turkish childrenwith autism and tablet10

computers. We believe that the observations and the following11

suggestions offer valuable guidance for future applications devel-12

oped for Turkish children with autism, including those by the US13

Embassy, and for future experimental studies that will investigate14

the effectiveness of such applications in teaching basic skills to15

Turkish childrenwith autism. Thediscussion that follows should be16

regarded as the first documented observation of the interaction of17

Turkish children with autism and tablet computers and viewed as18

informative case studies that pioneer research in this unexplored19

field.20

Our pilot study aimed to determine whether an internet-based21

iPad application developed using Applied Behavioral Analysis22

methods, which are widely employed worldwide in the education23

of individualswith autism,was able to teach sequencing to Turkish24

childrenwith autism fromdifferent age groups. Our results suggest25

that different versions of this sequencing applicationwith different26

Fig. 11. Cihan’s average performance.

Table 2Average performance of each participant.

Participant Initial test Teaching Final testWithoutassistance

Withassistance

Yavuz 1.28 2.50 4.00 2.26Oguz 3.22 4.88 Not applicable 4.71Cihan 4.60 5.00 Not applicable 4.86

properties geared towards each age group would be more suitable 27

than a single application for every child regardless of their age. 28

The average results (Table 2) show that each participant 29

improved from the first testing to the final testing sessions. Yavuz’s 30

scores increased from an average of 1.28–2.26. Oguz’s scores 31

increased from an average of 3.22–4.71. Cihan’s scores increased 32

from an average of 4.60–4.86. The fact that all participants 33

acquired their highest average scores in the teaching session can 34

be attributed to the effectiveness of the cues in the web-based 35

application in guiding the participants towards the correct answer, 36

and the fact that Yavuz and Oguz’s averages in the teaching session 37

are not 5 suggests that solely iPad-generated cues need to be 38

improved in assuring that the children do not make any mistakes 39

in the teaching sessions. 40

To address this problem, in the future applications, the cues can 41

be made more individualized. By conducting short surveys with 42

the families and the educators, the prompts most effective with 43

the individual students can be determined. Then, the prompts of 44

the game can be personalized accordingly, such as making the 45

rewardsmore intense, or using the combination of visual and audio 46

prompts that seem to appeal most to the particular student. 47

All participants acquired their lowest average scores for the 48

testing of the first card, and their performance got better while 49

placing the second and the third cards. This may be due to having 50

increased practice with the cards or the first cards serving as cues 51

that make the placement of the second and third cards easier. 52

Yavuz (4)was not able to learn sequencing from the instructions 53

and the prompts of the iPad application itself, and needed external 54

guidance and reinforcers. Yet, his performance for the second and 55

the third card was better than his performance for the initial card, 56

which may be because after the first step, there is a decreased 57

possibility of making a mistake. Three possible explanations can 58

underlie why this application was not suitable for this 4-year old 59

boy with autism. 60

The first explanation concerns limitations in linguistic abilities. 61

Yavuz seemed not to be able to understand the instructions of 62

the application. Although he is able to understand and follow 63

simple instructions, we thought that maybe the initial instruction 64

C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx 11

of ‘‘Please put the cards in order’’ was too long for Yavuz to1

comprehend, so after the first two trials, his teacher started to look2

at him in the eye and say ‘‘Order!’’ to make him understand the3

task. But he was still unable to perform the task successfully.4

The second explanation is poor fine motor skills. Although5

he had the basic fine motor skills that are required to use an6

iPad and play this application, and he had started using the iPad7

prior to testing, Yavuz’s fine motor skills may not have been8

developed completely yet, and they have not been refined through9

the education in Tohum School because he has only been a student10

for 2 years. Although he had the necessary prerequisite fine motor11

skills, during the sessions he was on his own, he would tap on the12

images but had trouble dragging them on a delicate touch screen13

to the correct location. In future applications, a young children-14

friendly version of the sequencing game can place the pictures in15

the correct order after the child taps on them. To order the cards,16

the only thing the young child needs to do may be to tap on the17

cards in the order of the action they are depicting, and watch them18

be placed in order by the application itself. Another way to address19

this problem may be adding an image of a hand on the screen20

that carries the card from the top row to its correct location in the21

bottom row so that the younger child can understand what he or22

she is expected to do in this application.23

In order to test whether this second prediction is correct, we24

waited one more month for Yavuz to get used to dragging and25

moving objects on the iPad and improve his fine motor skills.26

When we observed that he was able to successfully place puzzle27

pieces on the iPad without pressing too hard, we had him play28

our sequencing game one more time. Yet, despite acquiring these29

iPad skills, he was still unable to understand and perform our30

sequencing game.31

The third explanation for Yavuz’s inability to learn from the32

game by itself is his lower engagement compared to other par-33

ticipants. During daily activities, Yavuz was on task and engaged.34

We could not have known his on-task engagement for this iPad35

application because he had just started using the device. Engage-36

ment is defined as the student being in the place he/she is ex-37

pected to be, and (a) looking attentively, manipulating, or using38

the educational or leisure material appropriately, (b) attending to39

and looking at his/her teacher, the material presented, or other40

students participating in the education, (c) responding to his/her41

teacher’s prompts, (d) interactingwith his/her teacher or other stu-42

dents appropriately, and (e) transitioning to the other activities or43

activity spaces. If the student is engaged in an activity as defined44

by these criteria, but is also engaging in inappropriate behaviors,45

then he/she is not considered to be engaged [18]. In order to in-46

crease Yavuz’s engagement, we may need to apply differential re-47

inforcement, such that when he is engaged we will give him the48

most effective reinforcers, andwhen he is not engaged,wewill not.49

Yavuz’s case is a good reminder for educators of autism that some50

students, evenwhen they have the prerequisite skills to learn a cer-51

tain skill, may need extra help depending on the nature of the new52

skill, which underlines the importance of constant observation of53

the students’ responses and making necessary modifications.54

Cihan andOguz have been participating in the special education55

of Tohum School for 7 and 6 years respectively, but Yavuz has56

only been participating for 2 years. Individuals with autism have57

low engagement and the special curriculum in Tohum and other58

schools for autism are known to increase their engagement with59

the tasks presented to them. Since Yavuz is a relatively new60

student, although he had the required level of engagement to learn61

the sequencing task, his engagement is not as high as the other62

participants, and this may have affected his low performance in63

this application, since he did not attend to the prompts and cues of64

the application as well as Cihan and Oguz did.65

A future version of the sequencing game geared towards66

younger children with autism can have three stages. In the first67

stage, the empty boxes in the bottom row can have faded versions 68

of the photos on the top row so that the younger children can 69

match the photoswith the faded ones on the bottom. In ABA, when 70

teaching the childrenwith autismhow to follow activity schedules, 71

the first step is to teach them to match a picture in the activity 72

schedule with the same picture on the folder that contains the 73

materials for the selected activity. Since matching is the one of 74

the first skills to be taught in ABA, this first stage that requires 75

matching pictures to their faded versionsmay be useful in teaching 76

younger children how to sequence. In the second stage, the boxes 77

can be empty but the numbers 1 2 3 that denote the order of the 78

boxes may be large enough to fill the box so that they will be 79

easier to read. Yet, no numbers will be on the cards to be placed, 80

so the children will not be just matching the numbers; they will be 81

attending to the events depicted in the cards. And the third stage 82

can be the version that we used in this pilot study. Also, in future 83

versions, the instructions for the younger children with autism 84

should be one or two words instead of a sentence to make them 85

easier to understand. 86

Oguz (11)was able to learn sequencing through the instructions 87

and prompts of the application itself without any external guid- 88

ance or reinforcers. He seemedhappywhile playing the sequencing 89

game. During all his sessions, Oguz was able to drag the pictures to 90

the bottom row without pressing too hard on the screen. He was 91

smiling and was attentive to the application. He was not looking 92

elsewhere, butwas focused on his iPad. These observations suggest 93

that the format of the application and its materials were suitable 94

for Oguz, and as can be seen from his data, this version of the game 95

seems to be suitable and successful in teaching Oguz (11) how to 96

sequence actions. 97

Cihan (15) quickly learned the game and did not make use of 98

most of the prompts and cues that were intended to direct the 99

player towards the correct card. This version was too easy for 100

Cihan (15), and he lost interest after playing the game a few times. 101

He was not very interested in the videos, and he tried to skip 102

them. A version of this sequencing game that is more suitable and 103

interesting for older children with autism like Cihanmay usemore 104

than three cards on each game in order to increase the difficulty 105

of the task and include more than five activities in total in order to 106

keep the player interested. 107

It is important to note that it may not be appropriate to make 108

inferences about and generalize for certain age groups from these 109

participants because autism affects each individual to a different 110

degree even when they are of the same age. Thus, the same aged Q4 111

children with autism can show differing performances in learning 112

tasks. 113

The next step after this applicationwill be asking the children to 114

verbally explain the event depicted in the cards. During the testing 115

of our pilot application, we were curious whether the participants 116

were placing the cards randomly which happened to be in the 117

correct order by chance, or whether they have memorized the 118

order, or they actually understood what the cards were depicting. 119

We asked Oguz and Cihan to explain what they saw in the cards. 120

After his third teaching session, we askedOguz to play the game 121

again, but this time to explain what was happening in each card as 122

he was placing them. At first, he was explaining the events with 123

sentences that were semantically incorrect, such as ‘‘wearing the 124

closet’’. When he formed such incorrect sentences, his instructor 125

corrected him. When the cards for the same activity appeared 126

again, Oguz was able to describe the actions with semantically 127

correct sentences. This observation suggests that a further version 128

of this game that requires explaining the events after sequencing 129

them may be suitable for individuals who are similar to Oguz. 130

After he finished playing the teaching session, we asked 131

Cihan (15) to explain what he saw in the sequencing game. He 132

remembered all five activities and said: ‘‘I drank orange juice, I 133

12 C. Doenyas et al. / International Journal of Child-Computer Interaction xx (xxxx) xxx–xxx

brushed my teeth, I put on my coat, I rode a bike, I ate my bread’’.1

Here, Cihan is explaining the actions he saw someone else do in the2

iPad as if he did them, which is an example of pronoun reversal,3

a common linguistic deficit reported in autism [19]. When asked4

what else the application had, he said ‘‘Grease music’’, which was5

playing in the background of one of the videos.6

What we learned from this pilot study inspired us to improve7

sequencing applications for children with autism in the following8

ways:9

1. Identify the level of the participants in terms of different skills,10

such as iPad usage and vocabulary level.11

2. Make different versions geared towards different age and skill12

groups.13

3. Add a speaking section where the child explains the activity14

verbally after putting the cards in the correct order. This15

speaking section can be used as a prompt for children with16

autism to initiate a conversation with another individual, since17

social interaction and communication are major problems in18

autism.19

4. Add a generalization step where the child explains the activity20

to individuals other than his or her own instructor, in order to21

generalize the ability of recounting the events depicted in the22

cards to other people in the child’s life.23

What should coders designing applications for individuals with24

autism watch out for?25

• In our experiment, childrenwith autism tried to slide andmove26

the entire screen of the iPad (we are not talking about the slide27

motion to unlock the screen). They placed one hand palm down28

on the iPad, and tried to slide the screen with the other one,29

similar to a DJ spinning a CD, which caused our application to30

give errors. In the future, a code should be added to prevent the31

sliding of the entire screen.32

• Sometimes while the child was dragging the card, he accidently33

put another finger on another part of the screen. This caused34

the cards to stay in between the two placed fingers instead of35

moving to any of the boxes, and led to unexpected errors.36

• Sometimes the children hit the NEXT button multiple times37

back to back, and it caused our application to give errors.38

Future application developers should be aware of this tendency39

of individuals with autism to press on buttons many times40

back to back, and design the sensitivity of their applications41

accordingly.42

• Sometimes when the children tapped on the screen, the iPad43

interpreted it as a select and copy command. It created a box44

with the copy option on top, and this prevented the child from45

continuing to play the game. We are currently not aware of46

any solutions to this problem but nonetheless it is an issue that47

should be noted.48

• When the children got bored of the activity or wanted to49

play their favorite game on the iPad, they frequently used the50

middle button to exit the application. In the future educational51

applications that will be developed, that button can be made52

inactive until the end of the session or if that is not possible, can53

be blockedwith an externalmaterial to ensure the continuity of54

teaching through the application. With the new IOS 7 that was55

not available during our testing, there is an easy way to do this.56

Go to Settings → General → Accessibility → Guided Access

(turn it on and enter a passcode). Go to your application, press 57

the middle button three times and enable Guided Access. This 58

will disable the middle button and prevent the children from 59

pressing it to exit the application in the middle of a session. In 60

order to end the Guided Access, press the middle button three 61

times, enter your passcode, and select End the Guided Access. 62

We sincerely hope that these observations will be helpful for 63

coders who wish to design tablet applications for children with 64

autism, and especially for Turkish children with autism who may 65

then have a chance to benefit from both the Turkish government’s 66

and the US Embassy’s tablet computer projects in Turkey. 67

Acknowledgments 68

We would like to thank the following educators in Tohum 69

School for Autism in Istanbul for their invaluable support and 70

help during our experiment: Zeynep Kad, Nurdan Durmaz, Deniz 71

Yılmaz, and Sandra Pasensya. 72

References 73

[1] Diagnostic and Statistical Manual of Mental Disorders, 4th edn., text rev.,American Psychiatric Association, Washington, DC, 2000.

74

[2] R.L. Simpson, ABA and students with autism spectrum disorders: issues andconsiderations for effective practice, Focus Autism Other Dev. Disabil. 16 (2)(2001) 68–71.

Q5 75

[3] M. Moore, S. Calvert, Brief report: vocabulary acquisition for children withautism: teacher or computer instruction, J. Autism Dev. Disord. 30 (4) (2000)359–362.

76

[4] F.D. Fernandes, T.H. Santos, C.A. Amato, D.R. Molini-Avejonas, Computerizedresources in language therapy with children of the autistic spectrum, Pro.Fono. 22 (4) (2010) 415–420.

77

[5] N. Shah, Special education pupils find learning tool in iPad applications,Education Week 30 (22) (2011) 1–4.

78

[6] C. Whalen, L. Schreibman, Joint attention training for children with autismusing behavior modification procedures, J. Child Psychol. Psychiatry 44 (2003)456–468.

79

[7] E.A. Jones, E.G. Carr, K.M. Feeley, Multiple effects of joint attentionintervention, Focus on Autism andOther Developmental Disabilities 19 (2006)13–26.

80

[8] L.E. McClannahan, P.J. Krantz, Teaching Conversation to Childrenwith Autism:Scripts and Script Fading, Woodbine House, Bethesda, MD, 2005.

81

[9] J.O. Cooper, T.E. Heron, W.L. Heward, Applied Behavior Analysis, second ed.,Pearson Prentice Hall, Upper Saddle River, NJ, 2007.

82

[10] B. Birkan, Otizm spektrum bozuklugu olan çocuklar, in: S. Vuran (Ed.), ÖzelEgitim, Maya Akademi Yayın Egitim Danışmanlık, Ankara, Turkey, 2013.

83

[11] S. Venkatesh, S. Greenhill, D. Phung, B. Adams, T. Duong, Pervasivemultimediafor autism intervention, Pervasive and Mobile Computing 8 (6) (2012)863–882.

84

[12] P. Moreno-Ger, et al., Educational game design for online education, Comput. 85

Hum. Behav. (2008) http://dx.doi.org/10.1016/j.chb.2008.03.012. 86

[13] M.J. Dondlinger, Educational video game design: a review of literature, J. Appl.Educ. Technol. 4 (1) (2007) 21–31.

87

[14] S.J. Vaughan-Nichols, Will HTML 5 restandardize the web? Computer 43 (4)(2010) 13–15.

88

[15] D. Wechsler, 1995, Manual for the Wechsler Intelligence Scale for Children— 89

Revised. (I. Savaşır & N. Şahin, Trans.). New York, NY: The Psychological 90

Corporation. (Original work published in 1974). 91

[16] D.L. Hammond, A.D. Whatley, K.M. Ayres, D.L. Gast, Effectiveness of videomodeling to teach iPod use to students withmoderate intellectual disabilities,Education andTraining inAutismandDevelopmental Disabilities 45 (4) (2010)525–538.

92

[17] L.D. Paulson, Building rich web applications with ajax, Computer 38 (10)(2005) 14–17.

93

[18] G.S. MacDuff, P.J. Krantz, L.E. McClanahan, Teaching children with autismto use photographic activity schedules: maintenance and generalization ofcomplex response chains, J. Appl. Behav. Anal. 26 (1993) 89–97.

94

[19] K.M. Wilkinson, Profiles of language and communication skills in autism,Ment. Retard. Dev. D. R. 5 (1998) 73–79.

95