I NTERNATIONAL JOURNAL of S PECIAL E DUCATION

Robot-Mediated Interventions
to Enhance Communication and Social Abilities
of Children and Youth with Disabilities:
a Review of the Literature

Jiyeon Park1, Ye Eun Baek2, Bo Lyeong Lim2 and Heeseon Ko3

+LWHY[TLU[ VM :WLJPHS ,K JH[PVU , OH VTHUZ <UP]LYZP[ :LV S 2VYLH

:WLJPHS ,K JH[PVU 9LZLHYJO 0UZ[P[ [L , OH VTHUZ <UP]LYZP[ :LV S 2VYLH
.YHK H[L 7YVNYHT PU :WLJPHS ,K JH[PVU , OH VTHUZ <UP]LYZP[ :LV S 2VYLH

/6> 6 *0 , ABSTRACT
7HYR 1 )HLR @ , 3PT ) 3
2V / 9VIV 4LKPH LK Objectives: e current paper presents a review of the literature discussing
0U LY LU PVUZ V ,UOHUJL the trends in research that conducted robot-mediated interventions to en-
*VTT UPJH PVU HUK :VJPHS hance communication and social abilities of children and youth with disa-
(IPSP PLZ VM *OPSKYLU HUK @V O bilities. Method: irty-two dissertation or research articles published after
P O +PZHIPSP PLZ ( 9L PL 2000 were selected and divided into thematic categories for analysis, inclu-
VM OL 3P LYH YL 0U LYUH PVUHS ding participants, setting, research method, characteristics of robot-mediated
1V YUHS VM :WLJPHS ,K JH PVU interventions, roles of robots, and the results of studies. Results: First, the
research most frequently focused on children and youth with Autism Spec-
*699,:765 05. ( /69 trum Disorders. Second, most of the research was conducted in educational
@L , U )HLR settings. ird, most research used quantitative research design, in particular
I LUU NTHPS JVT a single-subject research design method was the most common. Fourth, in
60 most of the research humanoid robots assisted the intervention, acting as an
O WZ KVP VYN interaction partner of participants and as an instrumental medium to prompt
PQZL the participants to demonstrate desirable behavior. Finally, most studies repor-
ted positive e ects of robot-mediated intervention in the communication and
social abilities of children and youth with disabilities. Conclusion: e paper
provides the outcomes and limitations of robot-mediated interventions for
*67 90./ : ( ,4,5 children and youth with disabilities and the implications for future research.
*VW YPNO ( OVYZ
6WLU HJJLZZ W ISPJH PVU UKLY
OL LYTZ HUK JVUKP PVUZ
VM OL *YLH P L *VTTVUZ
( YPI PVU ** )@
SPJLUZL O W JYLH P LJVTTVUZ 2L VYK YVIV[ TLKPH[LK PU[LY]LU[PVU JOPSKYLU HUK V [O P[O KPZHIPSP[PLZ
VYN SPJLUZLZ I SP[LYH[ YL YL]PL O THUVPK JVTT UPJH[PVU ZVJPHS HIPSP[

PU[LYUH[PVUHSZWLK JVT
IJSE 9VIV P Y
INTRODUCTION
In the eld of special education, the use of robots began
in 1970 for the purpose of engineering to provide walking
assistance for people with visual impairments and those
with physical disabilities including the elderly. Among
the rst robots used were the walking aid robot „Sil-Bo”
developed in Korea, the robot suit „HAL” in Japan, „Ca-
re-O-bit” in Germany, and „Nursebot” in the US (Kim
& Song, 2012). In addition to engineering applications,
robots can also be used for the purpose of assisting the
education and therapy of children and youth with disa-
bilities or motivating and promoting their participation
(Kim, Lee, Shin, Kang, Kim, Choi, & Song, 2012).
Robots generally enable simple, consistent, and predic-
table interactions with users through systematic program-
ming, and they also promote interactions with the users
such as providing children and youth with immediate fe-
edback (Choi, 2015). e physical appearance of robots
induces children’s and youth’s interests and curiosity, and
their audiovisual stimuli can themselves serve as rewards.
Furthermore, the mobility and uidity of the robots ena-
ble active interaction with the users, such as touching and
manipulating body positions and thus, providing a sense
of vitality (Choi, 2015). ese characteristics of the robots
were found to have positive e ects on promoting social
interaction and improving communication skills of child-
ren and youth with disabilities (Kim, Lee, Chang, & Bae,
2011; Kim & Shin, 2014). Moreover, robots have positi-
vely impacted children’s and youth’s psychological aspects,
for example, children and youth who have interacted with
robots recognize them as friendly objects or experience de-
creased anxiety (Kim & Han, 2005).
Many domestic and international research teams have
developed robots for educational and therapeutic purpo-
ses of children and youth with disabilities (Choi, 2015).
e “KASPAR” of the AuRoRa Project, the “Bandit” ro-
bot, and the “Keepon,” were developed abroad; and a
humanoid robot named “Engkey,” initially developed for
English education, and intelligent robots such as “Irobi”
and “Pleo” were developed in Korea for language and
communication skills, eye contact and facial expression
reading skills, and social behavior of children and youth
with ASD (Kim, Lee, Shin, Kang, Kim, Choi, & Song,
2012; Kim & Song, 2012; Yun et al., 2015).
Children and youth with ASD, who have appeared
most often as participants in robot-related research, exhi-
PV V O J VTT PJ PV VJP S IPSP P VM OPS Y
bit limited, repetitive, and stereotyped behaviors, inte-
rests, and activities (Kim, Lee, Chang, & Bae, 2011).
Moreover, children and youth with ASD do not appear
to use or hardly use non-verbal communication beha-
vior, such as eye contact or facial expressions, and due
to lack of social communication skills—such as lack of
emotional sharing with others, imitation, and joint at-
tention skills—they can experience di culty in social
interactions (Lee & Park, 2011). Examining the research
that conducted robot-mediated interventions for child-
ren and youth with disabilities, it has been reported that
children and youth with ASD were not afraid of robots;
rather, they seemed to feel psychologically stable from
the simple and predictable behavior of the robot, as seen
in the observation of children interacting well and imi-
tating the behavior of the robot even when they were in
unfamiliar contexts (Dautenhahn & Billard, 2002).
Furthermore, several studies have reported that robots
appeared to play a role of social mediator in facilitating
communication and social interactions among children
and youth with ASD (Fachantidis et al., 2020; Kim, Lee,
Chang, & Bae, 2011). For instance, Fachantidis and his
colleagues (2020) conducted a robot-mediated coopera-
tive activity program for a small group of children and
youth with ASD and their peers without disabilities. It
was found that children and youth with ASD showed im-
provement in their social skills and communication skills
and a decrease in challenging behaviors, and also the peers
without disabilities who participated in the intervention
perceived children and youth with ASD more positively
than before. ese research studies show that robot-me-
diated intervention positively a ects initiative, autono-
my, preference exploration, and the formation of new
methods of interaction among children and youth with
ASD. It also reduces teachers’ e orts to conduct behavio-
ral modeling or to analyze the task when guiding children
and youth with ASD. Accordingly, the research indicated
that these interventions increased the concentration and
participation rate of children and youth with disabilities
and their peers who participated in the intervention by
stimulating their interests and curiosity (Charron, Lewis,
& Craig, 2017; Kim, Lee, Chang, & Bae, 2011).
As discussed above, the results of studies show that
children and youth with ASD are more engaged in inte-
ractions with automatic and uid robots than rigid, re-
petitive, and non-interactive toys and to date, numerous
studies on robots for children and youth with ASD have
been reported (Choi, 2015). However, considering the
+VP O[[WZ KVP VYN PQZL
1 7HYR , )HLR ) 3 3PT HUK / 2V
importance of inclusive education settings, in which chil-
dren and youth with disabilities and their peers without
disabilities are integrated, robot-mediated interventions
need to be applied to children and youth with diverse
disabilities. According to some previous studies, robots
have appeared to facilitate acquisition of communication
skills including articulation, voluntary conversation, and
verbal participation of children with disabilities (Silvera
-Tawil, Bradford, & Roberts-Yates, 2018). In addition,
there is an increasing interest in the eld of socially-as-
sistive robotics, which improves communication skills
and social abilities of children with disabilities through
practical social interactions with robots (Silvera-Tawil &
Roberts-Yates, 2018).
Children and youth having emotional and behavioral
problems, for example, can experience di culties in so-
cial interactions with people around them and engage in
behavior that is undesirable and inappropriate for the si-
tuation and their age due to various causes such as lack of
social skills and performance (Kim & Shin, 2014). With
regard to this, researchers have aimed to improve social
skills of children and youth with disabilities and provi-
de them with diverse learning opportunities by utilizing
various functions of robots such as expressing emotions
very similar to those of humans, having diverse and ac-
tive interactions through mobility and uidity, and pro-
viding immediate feedback according to the student’s
reaction to the robots (Kim & Shin, 2014; Shin & Kim,
2014). Moreover, a recent study examined the e ective-
ness of animal-type social robots (iJINI) on improving
verbal and social interactions of children and youth with
physical disabilities and the satisfaction of their parents
with regard to the robot (Song, Kim, & Kwak, 2020).
us, research using robots, which until recently was
only in its early stages, seems to be gradually diversifying
the area of robot use, re ecting over time the di erent
individual needs and requirements of users. In the past
research, robots have been mainly used in social tech-
nology–oriented programs, but they are used increasin-
gly more often in the areas of language, cognitive and
academic skills, and behavior (Kim & Shin, 2014). As
discussed above, research studies that conducted robot-
mediated interventions for children and youth with di-
sabilities are expanding further in terms of the diversity
of targeted children and youth and functions of robots.
e introduction and the uses of robot technology in
the educational eld is a strategy that can provide useful
support in the education and therapy of children and
PU[LYUH[PVUHSZWLK JVT
IJSE
youth with diverse and complex needs, such as the areas
of concentration, motivation, social abilities and repetiti-
ve learning (Choi, 2015).
In this study, we searched for and selected research
that employed robots as a medium for interventions tar-
geting communication and social abilities of children
and youth with disabilities for the past 20 years, and
we conducted a systematic analysis by categories. ro-
ugh this, we examined the outcomes and limitations of
robot-mediated interventions for children and youth
with disabilities and presented the implications for fu-
ture research. Moreover, this study endeavors to provide
robot developers with a basis for developing functions
that can be practically applied to interventions for chil-
dren and youth with disabilities and to provide educa-
tional experts, who develop content for robot-mediated
interventions, with a knowledge base to be considered
when they develop contents and strategies for those in-
terventions.
e research questions for this study are as follows:
rst, what has been the overall trend of research using
robots to improve communication and social abilities of
children and youth with disabilities in Korea and other co-
untries since 2000? Second, what is the trend in research
by category, especially with regard to the types and roles of
robots and the main subjects of the independent variables?
METHOD
1. Search strategy and the research selection process
e review is based on dissertations or research articles
that conducted robotic interventions for children and
youth with disabilities and were originally published
between January 2000 and September 2020. e elec-
tronic databases searched were ERIC, RISS, and Google
Scholar, and the keywords used to identify the articles
were (a) “disability*,” “autism,” “ASD,” “intellectual di-
sability*,” or “ADHD”; and (b) “robot,” “arti cial intel-
ligence,” and “AI.” e search on Google Scholar was
implemented as an initial step, and then the formal
search was done in the rest of the databases, including
ERIC and RISS. Looking at whether articles retrieved
on Google Scholar could be selected for analysis based
on the title and abstract of each article, it was found that
a large number of articles were not directly related to the
entered keywords and the selection criteria of this study.
Accordingly, a formal search of the remaining databa-
 IJSE 9VIV P Y PV V O J VTT PJ PV VJP S IPSP P VM OPS Y

e following inclusion criteria were em-

9LZLHYJO +H HIHZL
,90* 90:: .VVNSL :JOVSHY U
ployed: (a) dissertation or academic journal ar-
ticles that were published after January 2000;
(b) articles on experimental studies that direc-
*YP LYPH MVY ZLSLJ PVU *YP LYPH MVY L_JS ZPVU tly utilize robots for interventions, such as a
7 ISPZOPUN WL U 7 ISPZOPUN WL U
• Dissertation of degree • Research included in the academic robot leading or assisting the intervention; (c)
• Academic journal article JVUMLYLUJL OLZPZ IVVR articles on studies that include children and
7LLY YL PL LK • Research that cannot secure the
VYPNPUHS L_ youth with disabilities as participants; (d) ar-
ticles examining communication and/or social
abilities as a dependent variable. e following
Z
:LSLJ PVU VM 9LZLHYJO U
exclusion criteria were employed: (a) articles
published only at academic conferences and
*YP LYPH MVY ZLSLJ PVU *YP LYPH MVY L_JS ZPVU articles where the original text cannot be se-
9LZLHYJO TL OVK U 9LZLHYJO TL OVK U cured due to the lack of copyright consent; (b)
• Experimental research • Theoretical research on robots
OH ZLK YVIV Z JOHYHJ LYPZ PJZ KLZPNU M UJ PVU L J theoretical studies related to robots, research
HZ H TLKPH VY • Exploring research on robot utilization on the possibility of using robots, and research
• Analysis of research trend on robot
PSPaH PVU on trends in using robots; and (c) articles on
research using robots only for adults and chil-
dren and youth without disabilities. is pro-
UK
:LSLJ PVU VM 9LZLHYJO U
cess resulted in 32 research articles in total for
the systematic review (See Figure 1).
*YP LYPH MVY ZLSLJ PVU *YP LYPH MVY L_JS ZPVU
7HY PJPWHU Z U 7HY PJPWHU Z HUK JVW YPNO U
• Preschool, School ages • Adults (over the age of 18)
• Research including participants
2. Analysis methodology
(≤18)
• ASD, EBD, DD P OV KPZHIPSP PLZ VUS e selected 32 studies were analyzed by classi-
fying them into categories such as participants,
environment, types and roles of robots, types
YK
:LSLJ PVU VM 9LZLHYJO U of independent variables, dependent variables,
and overall research ndings. In order to un-
derstand the trends of participants, age, school
*YP LYPH MVY ZLSLJ PVU *YP LYPH MVY L_JS ZPVU level, disability types, and number of partici-
+LWLUKLU HYPHISL +LWLUKLU HYPHISL U
U • Other than communication and social pants in each study were examined. To exa-
• Language HIPSP PLZ ILOH PVY JVNUP P L ZRPSSZ mine the trends in the intervention settings,
HUK JVTT UPJH PVU VUS L J
• Social abilities
the study was coded as “schools”, “controlled
settings such as therapy rooms”, “mixed settin-
gs including more than one place”, “home”,
O
:LSLJ PVU VM 9LZLHYJO U and “settings not speci ed”. In addition, the
research method was analyzed by categories
-PN YL 7YV LK YL MVY ZLSL PUN YLZLHY O V IL HUHS aLK including a quantitative research, a qualitative
research, and a mixed research design. In order
to understand the current status of robot tech-
ses was conducted. A total of 1,384 articles were exported nology being used for communication and so-
from the electronic database search. After reviewing the cial intervention of children and youth with
article titles and abstracts, any research articles that were disabilities, the types of robots used in the in-
not related to disability and robots were excluded, and if tervention were coded as “a humanoid robot”
the dissertation was published in an academic journal, it and “a non-humanoid robot”, and the roles of
was considered appropriate for the study. is search pro- the robots in the intervention were coded as
cess yielded 162 articles. Overall, it was found that many “an intervention assistant” and “an interven-
studies were focused on communication or social abilities tion leader”. Types of interventions using ro-
of children and youth with disabilities. bots were examined according to language,

+VP O[[WZ KVP VYN PQZL

 1 7HYR , )HLR ) 3 3PT HUK / 2V
5 TILY VM HY PJSLZ
@LHY VM W ISPJH PVU
-PN YL
academic skills, and social emotional competence. e
dependent variables were coded as communication and
social abilities; and social abilities were in turn subdivi-
ded into social interaction, joint attention, social interac-
tion behavior in play, and others.
3. Inter-rater reliability
While there are literature review studies in which raters
independently analyze studies based on variables and cal-
culate its consistency, this study conducted an analysis
based on the previous study (Fettig & Barton, 2013) that
measured the inter-rater reliability in a unanimous man-
ner through consultation. e authors developed de -
nitions for each analysis variable and used a spreadsheet
to record information from the 32 studies. e following
variables were coded across 32 studies: “the participants”,
“the settings”, “the types(name) of the robot” and “its
roles in the intervention”, “independent variables”, “de-
pendent variables”, and “overall outcomes”. At least two
authors independently coded each study and compared
the spreadsheet with each other. Disagreements were
discussed until authors reached a consensus. However,
such disagreement only occurred twice in determining
whether the robot in two studies played an actual role
in the intervention. e overall agreement between the
two authors was calculated by dividing the number of
agreements by the number of agreements plus disa-
greements. e overall agreement was 99%.
RESULTS
Looking at the overall trend of 32 selected articles by
the year of publication, the largest number of articles on
PU[LYUH[PVUHSZWLK JVT
IJSE
5 TILY VM HY P SLZ I OL LHY VM W ISP H PVU
robot-mediated interventions was published in 2013,
followed by ve articles, each in 2014, 2015, and 2020.
None of the articles published between 2009 and 2011
met the selection criteria for this study, and over the rest
of the period, at least one article was published that met
the selection criteria for this study (See Figure 2). e
results of the analysis of 32 articles are summarized in
Appendix 1.
1. Analysis of study by category
1-1. Participants
According to the review of 32 articles, 15 studies were
conducted on preschoolers, accounting for 46.9% of the
total, whereas 11 studies on elementary school students
accounted for 34.4%. ere were ve studies (15.6%)
conducted on a mixture of preschoolers and elementary
school students, and one study (3.1%) was conducted on
middle school students.
In terms of the type of disability among the study
participants, ASD was the most common among 26 stu-
dies (81.3%), followed by EBD in three studies (9.3%).
In the others, there were two studies (6.3%) on deve-
lopmental delay (DD) and one (3.1%) on “minimally
verbal” children.
1-2. Setting
e research settings were analyzed by ve categories inc-
luding 1) educational settings such as general education
classrooms, special education classrooms, and daycare
centers, 2) controlled settings such as therapy rooms and
laboratories, 3) mixed settings including more than one
place such as natural settings and controlled settings, 4)
IJSE 9VIV P Y
home, and 5) settings not speci ed. Among 32 studies,
15 studies (46.9%) took place in the educational settings
and 11 studies (34.4%) in therapy rooms and laborato-
ries. Two studies (6.3%) were conducted in more than
one place including daycare centers and laboratories,
and only one study (3.1%) was conducted in the parti-
cipants’ home. Other than that, three studies (9.3%) did
not specify the research settings.
1-3. Research method
According to the result of analysis, the most common re-
search design was quantitative research, which was used
in 27 studies (accounting for 84.4% of the total), follo-
wed by four studies (12.5%) using a qualitative research
design. In addition, there was also one study (3.1%) using
a mixed research design that conducted both quantitati-
ve and qualitative research. In 27 quantitative research
studies, 21 studies examined the intervention e ect on
a single subject and six studies proving the intervention
e ect through comparison between the groups. Further-
more, among the experimental design methods used to
prove the e ect of robot-mediated intervention in single
-subject research studies, the reversal design and the multi
-baseline design across subjects proved the most popular,
followed by the pretest-posttest control group design.
2. Analysis of independent variables
2-1. Types of robots used for intervention
e robots used in the intervention could be classi ed into
(1) humanoid robots that have human-like facial and phy-
sical features and that can perform functions like humans,
such as interactions, and (2) non-humanoid robots, inclu-
ding robots with animal or character shapes. Seven types
of humanoid robots were used in 28 studies (87.5%), of
which two types of humanoid robots were used in the stu-
dies of Yun et al. (2015) and Lee et al. (2016). Six types of
non-humanoid robots were used in six studies (18.8%), of
which both a humanoid robot and a non-humanoid robot
were used in the study of Lee et al. (2016).
2-2. Types of robot-mediated interventions
e types of interventions using robots were categori-
zed into language, academic skills, and social emotional
competence, including joint attention, expression, imi-
tation. ere were seven studies using robots to develop
language and academic skills of children and youth with
disabilities, and 27 studies were focused on the social
emotional competence area (duplicate analyzed studies
included).
PV V O J VTT PJ PV VJP S IPSP P VM OPS Y
Among the 27 studies using robots in the social emotio-
nal competence area, nine studies focused on improving
joint attention skills. It can be inferred from this that there
has been an attempt to apply the method of using robots
in interventions to improve the skills that children and
youth with ASD, who have most often appeared in robot
-related studies, have the most di culty with.
2-3. Roles of robots in intervention
e roles of the robots used in the intervention was
categorized into leading intervention and assisting in-
tervention. In 15 studies (46.9% of the total), the role
of “leading intervention” was given for minimizing the
involvement of teachers in general and allowing the ro-
bot to lead the content for each session. Robots played
a role of presenting learning content and stimuli in the
process of conducting activities with children and youth
or while providing responses and feedback according
to the actions of children and youth. In particular, the
robot-driven instruction is programmed in advance to
output the input content or allow the researcher to re-
motely control the class while viewing the class situation
through the camera built into the robot to immediately
respond to children’s and youth’s behavior and provide
a feedback (Bekele et al., 2014; Charron et al., 2017;
Huskens et al., 2015; Pop et al., 2013; Yun et al., 2015;
Zheng et al., 2020).
However, when the robot was used as an assistant, it
was most often used as a social mediator or peer to the
participants to promote the target behaviors of children
and youth (eight of 17 studies). Subsequently, in another
eight studies, the teacher controlled the content loaded
into the robot in advance, allowing the robot to output
it at the appropriate time during class, or used the robot
as one of the classroom tools, such as for augmentative
and alternative communication (AAC) or as a student
modeling target. In addition, the robot presented quiz
questions related to classroom activities and played a role
in checking whether the children’s answers were provi-
ded correctly (Jeong, 2015).
3. Analysis of dependent variables
e dependent variables of 32 articles were categorized
into communication and social abilities. e results of
the analysis of dependent variables are shown in Table 1.
Ten studies examined the e ectiveness of robot-media-
ted intervention in improving communication skills—
such as self-directed questioning, sight word acquisition,
+VP O[[WZ KVP VYN PQZL
1 7HYR , )HLR ) 3 3PT HUK / 2V IJSE

;HISL (UHS ZPZ VM KLWLUKLU HYPHISLZ

LWLUKLU =HYPHISL *H LNVY 9LZLHY O U

*VTT UPJH PVU 3HUN HNL :OPU )HL L HS H / ZRLUZ L HS 2 VU L HS
HUK *VTT UPJH PVU 1LVU 3LL /HU 1LVUN :HHKH aP L HS
-HJOHU PKPZ L HS
:VJPHS (IPSP PLZ :VJPHS PU LYHJ PVU 9VIPUZ L HS + X L L L HS 7VW L HS )LRLSL L HS
@ U :PT L HS *OHYYVU L HS 2PT
:V L HS *VU P L HS +H PK L HS AOLUN L HS
1VPU H LU PVU 2PT :OPU )HL 2PT L HS 2PT L HS
PUP PH PUN :OPU L HS 3LL L HS
HUK YLZWVUKPUN ILOH PVY
:VJPHS PU LYHJ PVU )HL )HL L HS H )HL I 2PT L HS 3LL L HS
ILOH PVY PU WSH :V L HS
6 OLYZ :OPU / ZRLUZ L HS 7HYR L HS :PT L HS
-HJOHU PKPZ L HS 2VZ Y IPLJ L HS

9LZLHYJO HY[PJSL P[O K WSPJH[L HUHS ZPZ

and the ability to express phrases—of children and youth
with disabilities through interactions with robots.
Among the selected studies, 25 studies had social
abilities as a dependent variable, and these studies were
categorized based on which part of the social abilities was
targeted. Accordingly, social abilities were again classi ed
into four categories, including social interaction, joint at-
tention (initiating and responding behavior), social inte-
raction behavior in play, and others. Most of the studies,
including Duquette and her colleagues (2008), conduc-
ted robot-mediated interventions to improve the social
abilities of children and youth with disabilities. Among
them, most (12 studies) con rmed their e ectiveness in
social interactions, such as imitation, eye contact, and
turn-taking skills, followed by seven studies that focused
on improving initiating and responding behavior (joint
attention). Five studies examined participants’ social in-
teraction behavior in play, such as the selection of play
materials and diversity of play methods through interac-
ting activities with robots. Moreover, seven studies con-
rmed the e ectiveness of robot-mediated interventions
in other subcategories of social ability, such as caregiving
behavior, cooperative behavior, and task performance.
4. Analysis of study results
Most of the studies reported positive e ects of robot-me-
diated intervention in communication and social abilities.
In eight out of 10 studies that conducted robot-mediated
interventions related to communication, the intervention
improved children’s and youth’s expression behavior and
syntax expression, in particular it contributed to an in-
crease in average length of story, and also promoted the
acquisition of sight words, resulting in an increase in the
rate of reading words accurately and a decrease in the rate
of not being able to read words within a given period of
time. It also improved the vocalization of children and
youth with minimal speaking, increased communication
using AAC and other various communication functions,
and increased the number of sentences used.
Next, most of the studies conducted robot-mediated
intervention and con rmed the e ectiveness of utili-
zing robot on enhancing social abilities of children and
youth with disabilities. First, after participating in the
robot-mediated intervention, children and youth’s social
interactions - such as imitation, eye contact, turn-taking
performance, participation, selection, and so on - and
initiating and responding behaviors (joint attention) in-
creased signi cantly. e same was true for the use of va-
rious play methods and symbolic play-related behaviors.
Furthermore, in a study that examined the e ects of ro-
bot-mediated intervention by comparing the outcomes
between robot-mediated intervention and human-led
intervention, children’s joint attention and adaptive be-
havior signi cantly increased under the robot-mediated
intervention condition compared to human-led interven-
tion. Accordingly, it appeared that the robot-mediated in-
tervention was e ective for children’s social abilities, such
as showing very high satisfaction in the interaction with
robots (Kostrubiec et al., 2020; Pop et al., 2013).
However, some studies reported that there were no
intervention e ects found on children’s and youth’s co-
operative behavior, eye contact, symbolic play-related
behavior, and joint attention skills when working with
their siblings without disabilities (Huskens et al., 2015;

PU[LYUH[PVUHSZWLK JVT
IJSE 9VIV P Y
So et al., 2019; Yun et al., 2015; Zheng et al., 2020). In
some other studies, it was con rmed that robot-media-
ted intervention showed lower e ects, when compared
with the e ects of using non-robot media, such as hu-
mans or computers. Also, some researchers reported that
they did not see any signi cant e ect of robot-mediated
intervention by con rming that there was little di eren-
ce in the outcomes of the conditions for the dependent
variables (Jeong, 2015; Kwon & Kim, 2013).
DISCUSSION
In this section, we will discuss the implications of these
results for future robotic research.
First, it was found that robot-mediated intervention
for children and youth with disabilities was most applied
to elementary school students in terms of age and to chil-
dren and youth with ASD in terms of disability types.
Only one among 32 studies was conducted on secon-
dary students with disabilities. Since intervention using
robots is still in its infancy
It is considered more appropriate for elementary
school students who can cover simpler content than se-
condary students because. However, the fact that 81.3%
of the analyzed studies were conducted on children and
youth with ASD seems to be the result of the research
trends that continued to some extent as the previous re-
search using robots for the intervention in children and
youth with disabilities was also centered on children and
youth with ASD (Diehl, Schmitt, Villano, & Crowell,
2012). Furthermore, considering the characteristics of
ASD, which hinder social interactions, including eye
contact with others, the fact that robot-mediated in-
tervention is more advantageous for ASD than for any
other disability appears to have contributed to this rese-
arch trend. However, considering that technology, inc-
luding robots, has recently been used in interventions
among all age groups, it is desirable that the intervention
using robots is extended to children and youth of various
ages and types of disabilities (Chelvam, Zamin, & Sy-
lviaSteele, 2014; Katz, 2015).
Second, the educational settings and controlled labo-
ratory environments were the most frequent research set-
tings where robot was used for interventions. It is highly
desirable that robot-mediated interventions were con-
ducted in real-world educational settings. e fact that
PV V O J VTT PJ PV VJP S IPSP P VM OPS Y
more than a third of the selected studies conducted the
interventions in controlled settings raises concern. is
is not only because the laboratory is more advantageo-
us in creating space and building facilities necessary to
control and manipulate robots compared to the actual
training sites but also because the laboratory is relatively
easy to safely store robots and auxiliary equipment that
are expensive and require careful storage. However, con-
sidering that robots are already widely used in various
daily activities (e.g., robot cleaners) or human services
(e.g., robot nurses), more e orts are needed to utilize it
in natural environments. In order to take advantage of
modern technology, including robots, in the adult lives
that children and youth with disabilities will nd them-
selves in after high school, robotic intervention programs
should be devised in a natural environment, not just in a
de ned way for a limited time.
ird, studies that conducted robot-mediated in-
tervention used various research methods, but the most
frequently used design was a single-subject design. ere
were only six group comparison studies, and among them,
ve used a randomized controlled trial. is is because a
considerable number of analyzed studies have conducted
interventions in a one-on-one form with a small num-
ber of robots because it required considerable cost and
manpower to produce several robots at once and apply
them to a large number of children and youth. Conside-
ring that robot-mediated intervention in the special edu-
cation eld has not yet accumulated enough evidence to
become an evidence-based practice, this trend of research
design is understandable. Four out of 32 studies used qu-
alitative research methods, mainly to observe children’s
and youth’s reactions to robots. Although the number of
studies is not extensive, it is considered encouraging to at-
tempt to examine the e ects of robot-mediated interven-
tions in a qualitative manner, rather than relying solely
on quantitative measurements. Children’s and youth’s
reactions to the existence of robots themselves and their
responses to robot-mediated interventions require an in
depth qualitative consideration, and it is necessary to stu-
dy how to improve the social validity of robot-mediated
interventions by qualitatively exploring the perceptions
of children and youth, teachers, and parents about robots
used in educational practices through interviews or parti-
cipatory observations.
Fourth, most of the robots used for the intervention
were humanoid robots, and the representative areas were
joint attention and social interaction skills. In addition,
+VP O[[WZ KVP VYN PQZL
1 7HYR , )HLR ) 3 3PT HUK / 2V IJSE

robots led the intervention in about half of the analyzed
studies, and in the other half, robots assisted the teacher.
Joint attention and social interaction skills are represen-
tative di culties of children and youth with ASD, and
the most frequent interventions aimed at these areas are
likely linked to the fact that children and youth with ASD
were the largest number of participants in the analyzed
studies. Based on the literature gathered so far, it is dif-
cult to assess whether leading or assisting interventions
prove to be more e ective and e cient. Future research
will need to consider how to distribute roles between ro-
bots and teachers to maximize the educational e ect.
e ciency of technology so that it can have a direct e ect
on the motivation, learning, communication, and social
abilities of children and youth with disabilities. To this
end, transdisciplinary research that encompasses special
education, arti cial intelligence, and robotics, needs to
be conducted, and nancial aid should be provided to
enable the introduction of these advanced engineering
technologies to special education setting.

e limitations of the study are as follows: rstly, it

was di cult to control the quality level of the studies
because they were chosen based on the selection criteria
of research that conducted interventions using robots for
children and youth with disabilities. Some of the selected
studies did not provide speci c details, requiring analysis
as it contained limited information. If robots are more
actively used in the education and intervention of chil-
dren and youth with disabilities in the future and more
research is conducted, it is necessary to select studies ba-
sed on a certain level of qualitative indicators (e.g., Ger-
sten & Edyburn, 2007) and conduct a literature analysis.
Secondly, this study did not elaborate on the technical
aspects of robots because the purpose of this study was to
examine the use of robots in educational settings. Robots
do not think, judge, and intervene autonomously like
humans but, rather, have a built-in system and intervene
through external control, which includes a considerable
amount of robotic engineering procedures. Although this
study has not conducted such an engineering analysis, it
is necessary to analyze the engineering aspects of robots
used for intervention for children and youth with disabi-
lities through transdisciplinary research with robots and (*256>3, .,4,5
arti cial intelligence experts in the future. 5VUL
Despite the above mentioned limitations, the present 0:*36: 9, : ( ,4,5
study is meaningful since the author selected studies that No potential con ict of interest
used robots to improve communication and social abi- HZ YLWVY LK I OL H OVYZ
lities of children and youth with disabilities over the last FUNDING:
two decades to examine the e ects of robots in the in- ;OPZ VYR HZ Z WWVY LK
terventions described. Additionally, it provided a general I OL ;LJOUVSVN 0UUV H PVU 7YVNYHT
overview to understand the trends of the interventions. +L LSVWTLU VM H ZJYLLUPUN
Furthermore, this study examined in detail the types HUK YHPUPUN YVIV Z Z LT
of robots used for interventions and their roles. It is no PO WLYJLU ILOH PVYHS YLJVNUP PVU
longer new or unfamiliar to actively utilize arti cial in- IHZLK VU /90 MVY JOPSKYLU
telligence–based technologies, including robots, for in- P O KL LSVWTLU HS KPZVYKLY
tervention and education of children and youth with di- M UKLK I OL 4PUPZ Y VM ;YHKL
sabilities. Now, our mission is to maximize the access and 0UK Z Y ,ULYN 46;0, 2VYLH

PU[LYUH[PVUHSZWLK JVT
 (WWLUKP_ 6 LYHSS OHYH LYPZ P Z VM YLZLHY O ZPUN YVIV Z MVY OPSKYLU HUK V O P O KPZHIPSP PLZ
IJSE
7HY P PWHU Z
PHNUVZPZ
( OVY 9VIV 9LZLHY O 0UKLWLUKLU HYPHISL
5V (NL YHUNL Y :L PUN LWLUKLU HYPHISL 9LZ S Z
LHY 5HTL 9VSL TL OVK 5HTL VM 0U LY LU PVU
4V : OVVS
SL LS 5
9VIPUZ
/ THUVPK 85
L HS (:+ UZ , ,+ 9VSL VM H ZVJPHS TLKPH VY :VJPHS PU LYHJ PVU ZRPSSZ 0UJYLHZLK PU LYHJ PVU P O OL YVIV
9VIV H 0( ZPUNSL

UJYLHZLK ZOHYLK MVJ ZLK H LU PVU HUK PTP H PVU

+ X 0TP H PVU WSH :OHYLK -VJ ZLK ( LU PVU
PU YVIV JVUKP PVU
L L 3- (:+ 4VIPSL 9VIV 85 P O H YVIV PJ :OHYLK *VU LU PVUZ
;3 /PNOLY MVYTZ VM ZOHYLK JVU LU PVUZ HUK VM MHTPSPHY
L HS 7 ;P V 03 ZPUNSL HUK P O H O THU (IZLUJL VM :OHYPUN
HJ PVUZ PU O THU JVUKP PVU
TLKPH VY 6 OLY 7OLUVTLUVU
- No group di erence in absence of sharing
;LSLWYLZLUJL ,_WYLZZP L HJ P P PLZ *OPSK 9LZWVUZL 0UJYLHZLK H LU PVU PTP H PVU VM ZV UKZ
9VIV

2PT (:+
,+ 9VIV 83 OYV NO WYLZLU LK ( LU PVU 0TP H PVU HUK TV LTLU Z PUP PH PVU HUK L_WYLZZPVU
7
,UNRL 03 TH LYPHS PH YVIV 0UP PH P L ,_WYLZZPVU VM LTV PVUZ HUK MLLSPUNZ PU YVIV JVUKP PVU
P

+PUVZH Y ZOHWLK 0UJYLHZLK JHYLNP PUN ILOH PVYZ HUK LYIHS

:OPU (:+ 85 *HYLNP PUN ILOH PVYZ *HYLNP PUN )LOH PVY
;3 9VIV PJ (UPTHS PUP PH PVUZ V HYK YVIV VY YLNHYKPUN
7 , ZPUNSL ZPUN H ZVJPHS YVIV =LYIHS 0UP PH PVU
73,6 0( JHYLNP PUN ILOH PVY
Y

: K 0U LYHJ PVU : K 0UJYLHZLK L_WYLZZPVU VM PU LYLZ

;LSLWYLZLUJL 4P_LK : K 9LZWVUZL VM *OPSK
)HL (:+ P O H YVIV HUK LTV PVUZ HZ LSS HZ HJ P L L_WSVYH PVU
,+ 9VIV 83 85 : K 7SH )LOH PVY
7 : K 9VIV IHZLK : K 0UJYLHZLK WSH ILOH PVY HUK UVU WSH
PV

,UNRL 0( ZPUNSL 5VU WSH )LOH PVY

WSH PU LY LU PVU WYVNYHT ILOH PVY VM (:+ JOPSKYLU
V

)HL ;LSLWYLZLUJL
(:+ 4 ZPJHS HJ P P PLZ 0U LU PVU ,_WYLZZPVU 0UJYLHZLK ILOH PVY VM L_WYLZZPUN PU LU PVU
L HS ,+ 9VIV 83
7 P O YVIV Z ,_JOHUNL )LOH PVY HUK ILOH PVY VM L_JOHUNPUN PU LU PVU
O

H ,UNRL 0(
J

)HL ;LSLWYLZLUJL 0UJYLHZLK WSH ILOH PVY ZPUN WSH TH LYPHSZ

++ 9VIV IHZLK WSH
L HS ,+ 9VIV 83 7SH )LOH PVY HUK PTP H PVU VM WSH ILOH PVY 0UJYLHZLK PU LYLZ
7 U LY LU PVU WYVNYHT
I ,UNRL 03 PU WSH TH LYPHSZ HUK ILOH PVY VM WLLYZ K YPUN WSH
VTT

(:+ (+/+
/ THUVPK
/ ZRLUZ 56: 7++ 85 ()( IHZLK PU LY LU PVU 0UJYLHZLK U TILY VM ZLSM PUP PH LK X LZ PVUZ
;3 9VIV :LSM PUP PH LK X LZ PVUZ
56: ZPUNSL JVUK J LK I H YVIV MVY IV O L_WLYPTLU HS NYV WZ
5(6 03
PJ PV

,
:VJPHS )LOH PVY
0U LY LU PVU WYVNYHT 0UJYLHZLK ZVJPHS PUP PH P L YLZWVUZL ILOH PVY
2PT ;LSLWYLZLUJL 0UP PH P L 9LZWVUZL )LOH PVY

+VP O[[WZ KVP VYN

++ 85 ZPUN YVIV MVY ZVJPHS ZLSLJ PVU VM WSH TH LYPHSZ HUK PU LYLZ PU SLHYUPUN
L HS 4 9VIV 7SH )LOH PVY :LSLJ PVU VM
7 ZPUNSL ILOH PVYZ HUK WSH WSH TL OVK
,UNRL 0( 7SH 4H LYPHS 7SH 4L OVK
HJ P P PLZ +LJYLHZLK UVU WSH ILOH PVY
5VU WSH )LOH PVY
;OYLL ZL PUNZ YVIV
VJP S IPSP P

2 VU 0U LSSPNLU
(:+ 85 TLKPH LK PUZ Y J PVU 3HUN HNL ,U PYVUTLU - No di erences among three conditions
L HS ,+ 9VIV

PQZL
, ZPUNSL ZWLJPHS LK JH PVU =LYIHS 0U LYHJ PVU HZ MVY SHUN HNL LU PYVUTLU HUK LYIHS PU LYHJ PVUZ
0YVIP /VTL 0(
PUJS ZP L LK JH PVU
(:+ 4VUZ LY ZOHWLK
VM OPS Y

7VW 85 0UJYLHZLK PUKLWLUKLUJ VM (:+ JOPSKYLU

7 , 5: :VJPHS 9VIV :VJPHS Z VYPLZ :VJPHS 9LZWVUZL
NYV W PU L_OPIP PUN ZVJPHS JVTWL LUJL
,_ * 7YVIV 03
 (:+ (:+ NYV W Z PUJYLHZLK SL LS VM PU LYLZ PU YVIV YPHSZ
.HaL KPYLJ LK
7 / THUVPK *SVZLK SVVW HKHW P L )V O NYV WZ YLX PYLK H OPNOLY SL LS VM WYVTW PUN
1 7HYR

)LRLSL 85 H OL HKTPUPZ YH VY
* 5V ;3 9VIV YVIV TLKPH LK Z Z LT PU YVIV YPHSZ
NYV W 3L LS VM 7YVTW PUN 73
KPZHIPSP PLZ 5(6 03 1VPU H LU PVU WYVTW Z 0UJYLHZLK YH L VM Z JJLZZ PU O THU HKTPUPZ YH VY
;HYNL : JJLZZ /P -YLX LUJ
7 - Hit frequencies were not di erent across conditions
4PUPTHS 0U LSSPNLU /PNOLY MYLX LUJ VM JVTT UPJH PVU
1LVU 85 / THUVPK YVIV HZZPZ LK
ZWLHRPUN ;3 9VIV *VTT UPJH PVU ZRPSSZ 0UJYLHZLK JVTT UPJH PVU H LTW Z ZPUN
ZPUNSL ((* PU LY LU PVU
7 0YVIP /VTL 0( VJHSPaH PVU HUK ((*
0UJYLHZLK ZVJPHS PUP PH PUN ILOH PVY
2PT ;LSLWYLZLUJL :VJPHS ZRPSSZ WYVNYHT

PU[LYUH[PVUHSZWLK JVT
,)+ H YPZR 85 :VJPHS 0UP PH PUN )LOH PVY HUK KLJYLHZLK JOHSSLUNPUN ILOH PVY
L HS ,+ 9VIV ZPUN H LSLWYLZLUJL
, ZPUNSL 3PUN PZ PJ *OHYHJ LYPZ PJZ 7VZP P L YLZWVUZL WH LYUZ VM WLLYZ
,UNRL 0( YVIV
OV KPK UV WHY PJPWH L PU OL Z K
0U LSSPNLU 5 TILY VM VYKZ :WVRLU 5V JOHUNL PU OL U TILY VM VYKZ ZWVRLU
3LL (:+ 85 4 S PTLKPH HJ P P PSP
,+ 9VIV :LU LUJL 3LUN O 0UJYLHZLK ZLU LUJL SLUN O HUK WLYJLU HNL
, )HLR ) 3 3PT HUK / 2V

, ZPUNSL aPUN HU PU LSSPNLU YVIV

0YVIP /VTL 0( *VTT UPJH PVU - UJ PVU VM ZPUN JVTT UPJH PVU M UJ PVUZ
:OPU ;LSLWYLZLUJL 9VIV HZZPZ LK ZVJPHS 9LZWVUZP L )LOH PVY 0UJYLHZLK YLZWVUZL V LTWH O HUK HKQ Z TLU
,)+ 85
L HS ,+ 9VIV ZRPSSZ WYVNYHT JLU LYLK V ,TWH O HUK (KQ Z TLU H LTW Z HUK TVYL KP LYZL WL VM YLZWVUZP L
, ZPUNSL
,UNRL 0( VU PY L ( LTW Z ILOH PVY
( LYHNL : VY 3LUN O ( LYHNL
0U LSSPNLU : VY PU LY LU PVU 0UJYLHZLK H LYHNL SLUN O VM Z VYPLZ HUK WOYHZLZ
/HU (:+ 0+ 85 7OYHZL 3LUN O *VTWSL_
/ 9VIV ZPUN LK JH PVUHS and number of cohesive devices (diversi cation)
4 ZPUNSL :LU LUJL 9H PV
0YVIP 8 0( O THUVPK YVIV 5V JOHUNL PU OL YH PV VM JVTWSL_ ZLU LUJLZ
5 TILY VM *VOLZP L +L PJLZ
(:+ (+/+
5V JOHUNL PU PUP PH PVU VM PU LYHJ PVU HUK YLZWVUZL
/ ZRLUZ UZ / THUVPK 9VIV TLKPH LK *VSSHIVYH P L )LOH PVYZ
85 ILOH PVY
L HS :PISPUN V ;3 9VIV PU LY LU PVU IHZLK 0U LYHJ PVUZ 0UP PH PVUZ
NYV W UJYLHZLK WSH PUN VNL OLY ILOH PVY
KPZHIPSP PLZ 5(6 03 VU 3LNV OLYHW HUK 9LZWVUZLZ 7SH ;VNL OLY
6US VUL WHPY VM JOPSK ZPISPUN
,
/- (:+ 0UJYLHZLK U TILY VM LTV PVUHS VYKZ
0U LSSPNLU : VY PU LY LU PVU ,_WYLZZPVU VM ,TV PVUHS
1LVUN , 85 PU OL YVIV JVUKP PVU
4 9VIV ZPUN TLKPH =VJHI SHY =VJHI SHY
7* NYV W - No di erence between intervention media
0YVIP 8 0( O THUVPK YVIV Z 7* +P LYZP
9VIV PU VJHI SHY KP LYZP
.LULYHS LK JH PVU ; Z L HS H PVU 0U LY LU PVU
-HJ VYZ VM :VJPHS was e ective in three out of ve elements of social
7HYR ;LSLWYLZLUJL ;LSLWYLZLUJL *VTWL LUJL )HZLK competence and e ective on four out
,)+ 85
L HS ,+ 9VIV YVIV HZZPZ LK ZVJPHS VU OL , HS H PVU of ve students
, NYV W
,UNRL 0( ZRPSSZ WYVNYHT VM .LULYHS HUK :WLJPHS :WLJPHS LK JH PVU ; L HS H PVU 0U LY LU PVU
,K JH PVU ;LHJOLYZ was e ective in all elements of social competence
and e ective on all students
IJSE

0U LSSPNLU 9VIV
0UJYLHZLK SL LS VM PU LYLZ PU OL YVIV
@ U 0YVIP 8 9VIV HZZPZ LK
85 , L *VU HJ PU OL ILNPUUPUN I KLJYLHZLK P O YLWLH LK
L HS (:+ 7 5: *SPUPJHS ILOH PVYHS PU LY LU PVU
ZPUNSL 9LHKPUN -HJPHS ,_WYLZZPVUZ ZLZZPVUZ 0UJYLHZLK WLYJLU HNL VM JVYYLJ HUZ LYZ
(ZZPZ HU 9VIV Z Z LT
MVY YLHKPUN MHJPHS L_WYLZZPVUZ
*(96 03
(:+ 7 (UPTHS ZOHWLK
3LL 7SH JLU LYLK HJ P P PLZ 9LZWVUZL )L LLU
7LLY V 9VIV .LUPIV 0UJYLHZLK JVVWLYH P L YLSH PVUZOPW IL LLU JOPSKYLU
L HS ,+ 83 ZPUN HU PU LSSPNLU YVIV *OPSKYLU P O HUK P OV
KPZHIPSP PLZ / THUVPK P O KPZHIPSP PLZ HUK JOPSKYLU P OV KPZHIPSP PLZ
0U LYHJ PVU )LOH PVYHS +PZHIPSP PLZ
7 9VIV 0YVIP 8 0(
 7HY P PWHU Z
IJSE
PHNUVZPZ
( OVY 9VIV 9LZLHY O 0UKLWLUKLU HYPHISL
5V (NL YHUNL Y :L PUN LWLUKLU HYPHISL 9LZ S Z
LHY 5HTL 9VSL TL OVK 5HTL VM 0U LY LU PVU
4V : OVVS
SL LS 5

:PT 4VUZ LY ZOHWLK

(:+ 85 ;HZR 7LYMVYTHUJL
L HS ,+ :VJPHS 9VIV 4HRPUN MY P ZHSHK 0UJYLHZLK L L JVU HJ
UZ NYV W :VJPHS ILOH PVY 0U LYHJ PVU
7YVIV 0(
*OHYYVU / THUVPK 1VPU H LU PVU
(:+ 85 0UJYLHZLK PUP PH PUN HUK YLZWVUKPUN
L HS ,+ 9VIV PUP PH PUN YLZWVUKPUN 1VPU ( LU PVU :RPSSZ
, ZPUNSL ILOH PVY V QVPU H LU PVU
5(6 03 ILOH PVY
;LSLWYLZLUJL 4 S PTLKPH JVU LU Z )LOH PVY ,UNHNLTLU 0UJYLHZLK LUNHNLTLU PU MYLL WSH PTL
2PT (:+ 85
,+ 9VIV MVY ZLSM KL LYTPUH PVU :LSLJ PVU 0UJYLHZLK ILOH PVY VM ZLSLJ PUN HUK L_WYLZZPUN
7 ZPUNSL
9VIV QQHUN 03 HJ P P PLZ 7YLMLYLUJL L_WYLZZPVU WYLMLYLUJLZ PU JSHZZ
9VIV

:HHKH aP / THUVPK =PY HS ,U PYVUTLU 0UJYLHZLK WLYJLU HNL VM VYKZ YLHK JVYYLJ S
(:+ 85 VYKZ 9LHK *VYYLJ S
L HS ;3 9VIV HUK 7LKHNVNPJHS (NLU +LJYLHZLK YH L VM TPZYLHKPUN VY THRPUN LYYVYZ
P

UZ ZPUNSL ,YYVYZ 4HKL I OL 7HY PJPWHU

5(6 0( :PNO VYK 0UZ Y J PVU PU H NP LU VYK
,_ (:+
/ THUVPK 0UJYLHZLK PUP PH PUN ILOH PVY V QVPU H LU PVU
:V L HS ( L 7 85 9VIV IHZLK WSH 1VPU ( LU PVU :RPSSZ
Y

;3 9VIV 5V JOHUNLZ VU OL YLZWVUKPUN ILOH PVY

* (:+ NYV W KYHTH PU LY LU PVU : TIVSPJ 7SH
5(6 03 V QVPU H LU PVU HUK Z TIVSPJ WSH
( L 7
PV

*VU P / THUVPK
(:+ 85 0UJYLHZLK YLJVNUP PVU VM V OLYZ LTV PVUZ
L HS ;3 9VIV ;,(**/ A ect Recognition
V

UZ ZPUNSL HUK MLLSPUNZ

5(6 03
O

+H PK / THUVPK )V O JVUKP PVUZ PUJYLHZLK YU HRPUN WLYMVYTHUJL

(:+ 85 ; YU HRPUN 7LYMVYTHUJL
J

L HS ;3 9VIV ; YU HRPUN HZR 0UJYLHZLK YLJVNUP PVU VM YVIV WHY ULY HZ ILPUN
UZ ZPUNSL :VJPHS 0U LYHJ PVU
5(6 03 TVYL PU LYLZ PUN OHU OLPY O THU WHY ULY
-HJOHU
(:+ , 9VIV )PJ JSL :VJPHS :RPSSZ UJYLHZLK ZVJPHS ZRPSSZ JVTT UPJH PVU ZRPSSZ
VTT

PKPZ 85
5V KPZHIPSP PLZ ,+ + 3,.6 Y 3,.6 OLYHW WYVNYHT *VTT UPJH PVU :RPSSZ HUK WVZP P L H P KLZ VM WLLYZ P OV KPZHIPSP PLZ
L HS ZPUNSL
, obot”)/IL *OHSSLUNPUN )LOH PVY +LJYLHZLK JOHSSLUNPUN ILOH PVY
PJ PV

2VZ Y 0UJYLHZLK SL LS VM PU LYLZ HUK YLZWVUZL V Z WWVY

:WOLYL ZOHWLK 9LX LZ PUN
IPLJ (:+ UZ 85 :VJPHS :RPSSZ HUK YLPUMVYJLTLU VM YVIV
5: 9VIV HUK YU HRPUN
L HS ;9 NYV W 9VIV (JJLW HUJL 0UJYLHZLK WYV ZVJPHS ILOH PVY PU OL WYLZLUJL

+VP O[[WZ KVP VYN

UZ 0( PU LY LU PVU
VM H IHSS P L OL JVUKP PVU P OV YVIV
AOLUN (:+ ( L / THUVPK
85 9VIV TLKPH LK
L HS UZ ,_ ;3 9VIV 1VPU ( LU PVU :RPSSZ - No di erence between conditions
NYV W PU LY LU PVU
* 5(6 03
VJP S IPSP P

PQZL
5V[L 7 7YLZJOVVS , ,SLTLU[HY ZJOVVS 4 4PKKSL ZJOVVS ;9 ;LHJOLY ,_ ,_WLYPTLU[ NYV W
* *VU[YVS NYV W ,+ ,K JH[PVUHS ZL[[PUNZ ZJOVVSZ HUK KH JHYL ;3 ;OLYHW YVVT HUK 3HIVYH[VY
4 4P_LK ZL[[PUNZ / /VTL 5: = Not speci ed; 0( 0U[LY]LU[PVU (ZZPZ[HU[ 03 0U[LY]LU[PVU 3LHKLY
VM OPS Y

83 8 HSP[H[P]L YLZLHYJO 85 8 HU[P[H[P]L YLZLHYJO

1 7HYR , )HLR ) 3 3PT HUK / 2V IJSE

REFERENCES
)HL 4 1 ( K VU OL LMMLJ VM H YVIV IH LK WSH PU LY LU PVU WYVNYHT VU OL WSH ILOH PVY VM V UN JOPSKYLU
P O H P T WLJ Y T KP VYKLY +VJ[VYHS KPZZLY[H[PVU +HLN <UP]LYZP[ +HLN 2VYLH
)HL 4 1 3LL / : (U PU]LZ[PNH[PVU PU[V [OL PU[LU[PVU L_WYLZZPVU ILOH]PVY VM V UN JOPSKYLU P[O [OL H [PZT ZWLJ[Y T KPZVYKLY
;OL X HSP[H[P]L HUHS ZPZ VM T ZPJHS HJ[P]P[PLZ P[O YVIV[Z 1V YUHS VM :WLJPHS ,K JH PVU 9LOHIPSP H PVU :JPLUJL
)HL 4 1 3LL / : 7HYR * 1 2PT 1 / ( Z[ K VU [OL WSH ILOH]PVY VM V UN JOPSKYLU P[O KL]LSVWTLU[HS KLSH Z
PU PUJS ZP]L MYLL WSH P[O H MVJ Z VU ZTHSS NYV W MYLL WSH HM[LY [OL PTWSLTLU[H[PVU VM H YVIV[ IHZLK WSH PU[LY]LU[PVU WYVNYHT
1V YUHS VM :WLJPHS ,K JH PVU 9LOHIPSP H PVU :JPLUJL
)LRLSL , *YP[[LUKVU 1 ( : HUZVU ( :HYRHY 5 HYYLU A , 7PSV[ JSPUPJHS HWWSPJH[PVU VM HU HKHW[P]L YVIV[PJ Z Z[LT
MVY V UN JOPSKYLU P[O H [PZT ( P T
*OHYYVU 5 3L PZ 3 *YHPN 4 ( YVIV[PJ [OLYHW JHZL Z[ K +L]LSVWPUN QVPU[ H[[LU[PVU ZRPSSZ P[O H Z[ KLU[ VU [OL H [PZT ZWLJ[Y T
1V YUHS VM ,K JH PVUHS LJOUVSVN : LT
*OLS]HT 2 AHTPU 5 : S]PH:[LLSL . : ( WYLSPTPUHY PU]LZ[PNH[PVU VM 4 +0;9(*2 9
( TLKPJPUL KPZWLUZPUN TVIPSL YVIV[ MVY ZLUPVY JP[PaLUZ 7YVJLKPH VTW LY :JPLUJL
*OVP , 3P[LYH[ YL YL]PL VM YVIV[Z ZLK MVY [OL YLOHIPSP[H[PVU VM JOPSKYLU P[O H [PZ[PJ ZWLJ[Y T KPZVYKLY
1V YUHS VM YLOHIPSP H PVU LSMHYL LUNPULLYPUN H P P L LJOUVSVN
*VU[P + ;Y IPH . ) VUV : +P 5 V]V : +P 5 V]V ( (MMLJ[ YLJVNUP[PVU PU H [PZT H ZPUNSL JHZL Z[ K VU PU[LNYH[PUN H O THUVPK
YVIV[ PU H Z[HUKHYK [OLYHW 8 LY 6WLU HUK 0U LYKP JPWSPUHY 1V YUHS VM LJOUVSVN S YL HUK ,K JH PVU
+H [LUOHOU 2 )PSSHYK ( .HTLZ JOPSKYLU P[O H [PZT JHU WSH P[O 9VIV[H H O THUVPK YVIV[PJ KVSS 0U : 2LH[LZ 71 *SHYRZVU
74 3HUNKVU 7 9VIPUZVU ,KZ UP LY HS HJJL HUK H P P L LJOUVSVN WW 3VUKVU <UP[LK 2PUNKVT :WYPUNLY =LYSHN
+H]PK + 6 *VZ[LZJ * ( 4H[ : :aLU[HNV[HP ( +VIYLHU ( ,MMLJ[Z VM H YVIV[ LUOHUJLK PU[LY]LU[PVU MVY JOPSKYLU
P[O (:+ VU [LHJOPUN [ YU [HRPUN ZRPSSZ 1V YUHS VM ,K JH PVUHS VTW PUN 9L LHYJO
+PLOS 1 1 :JOTP[[ 3 4 =PSSHUV 4 *YV LSS * 9 ;OL JSPUPJHS ZL VM YVIV[Z MVY PUKP]PK HSZ P[O H [PZT ZWLJ[Y T KPZVYKLYZ
( JYP[PJHS YL]PL 9L LHYJO PU ( P T :WLJ Y T P VYKLY
+ X L[[L ( 4PJOH K - 4LYJPLY / ,_WSVYPUN [OL ZL VM H TVIPSL YVIV[ HZ HU PTP[H[PVU HNLU[ P[O JOPSKYLU
P[O SV M UJ[PVUPUN H [PZT ( VUVTV 9VIV
Fachantidis, N., Syriopoulou-Delli, C. K., Vezyrtzis, I., & Zygopoulou, M. (2020). Bene cial effects of robot-mediated class activities
VU H JOPSK P[O (:+ HUK OPZ [ WPJHS JSHZZTH[LZ 0U LYUH PVUHS 1V YUHS VM L LSVWTLU HS P HIPSP PL
-L[[PN ( )HY[VU , , 7HYLU[ PTWSLTLU[H[PVU VM M UJ[PVU IHZLK PU[LY]LU[PVU [V YLK JL JOPSKYLU Z JOHSSLUNPUN ILOH]PVY
( SP[LYH[ YL YL]PL VWPJ PU ,HYS OPSKOVVK :WLJPHS ,K JH PVU
Gersten, R., & Edyburn, D. (2007). De ning quality indicator for special education technology research.
1V YUHS VM :WLJPHS ,K JH PVU LJOUVSVN
/HU ) OL LMMLJ VM H VY PU LY LU PVU PUN LK JH PVUHS O THUVPK YVIV VU U HJ PJ RPSS VM JOPSKYLU
P O H P T WLJ Y T KP VYKLY 4HZ[LY Z KPZZLY[H[PVU , OH VTHUZ <UP]LYZP[ :LV S 2VYLH
/ ZRLUZ ) 7HSTLU ( =HU KLY LYMM 4 3V YLUZ ; )HYHRV]H , 0TWYV]PUN JVSSHIVYH[P]L WSH IL[ LLU JOPSKYLU
P[O H [PZT ZWLJ[Y T KPZVYKLYZ HUK [OLPY ZPISPUNZ ;OL LMMLJ[P]LULZZ VM H YVIV[ TLKPH[LK PU[LY]LU[PVU IHZLK VU 3LNV [OLYHW
1V YUHS VM H P T HUK KL LSVWTLU HS KP VYKLY
/ ZRLUZ ) =LYZJO Y 9 .PSSLZLU 1 +PKKLU 9 )HYHRV]H , 7YVTV[PUN X LZ[PVU HZRPUN PU ZJOVVS HNLK JOPSKYLU
P[O H [PZT ZWLJ[Y T KPZVYKLYZ ,MMLJ[P]LULZZ VM H YVIV[ PU[LY]LU[PVU JVTWHYLK [V H O THU [YHPULY PU[LY]LU[PVU
L LSVWTLU HS UL YVYLOHIPSP H PVU
1LVU 2 / OL LMMLJ VM (( PU LY LU PVU PUN O THUVPK YVIV VU HIPSP VM JVTT UPJH P L L WYL PVU VM WYL JOVVSLY
JOPSKYLU P O TPUPTHSS LYIHS 4HZ[LY Z KPZZLY[H[PVU , OH VTHUZ <UP]LYZP[ :LV S 2VYLH
1LVUN 4 2 3L PJHS YLWYL LU H PVU VM LTV PVU MVY OPNO M UJ PVU H P T PH VY WYVNYHT PU LY LU PVU PUN THY TLKPH
O THUVPK YVIV 7 4HZ[LY Z KPZZLY[H[PVU , OH VTHUZ <UP]LYZP[ :LV S 2VYLH
2HUN 1 / 2PT * . 3LL : / :VUN ) : +L]LSVWTLU[ VM HSRPUN HZZPZ[HUJL YVIV[ MVY [OL ISPUK
1V YUHS VM OL 2VYLHU :LU VY :VJPL
2H[a 9 ;LSL JHYL YVIV[ MVY HZZPZ[PUN PUKLWLUKLU[ ZLUPVY JP[PaLUZ OV SP]L H[ OVTL
: KPL PU LHS O LJOUVSVN 0UMVYTH PJ
2PT : 1 ;OL LMMLJ[Z VM HU PUJYLHZL PU ZLSM KL[LYTPUH[PVU L_WLYPLUJL VU [OL ILOH]PVY VM V UN JOPSKYLU P[O H [PZT ZWLJ[Y T KPZVYKLY
I [LSLWYLZLUJL YVIV[ 1V YUHS VM 9LOHIPSP H PVU LSMHYL ,UNPULLYPUN ( P P L LJOUVSVN
2PT ) 1 /HU 1 / (UHS ZPZ VU WZ JOVSVNPJHS HUK LK JH[PVUHS LMMLJ[Z PU JOPSKYLU HUK OVTL YVIV[ PU[LYHJ[PVU
1V YUHS VM OL 2VYLHU ( VJPH PVU VM 0UMVYTH PVU ,K JH PVU
2PT 1 / 2PT : 1 3LL / : :OPU / 7HYR * 1 2HUN 1 ) 2PT * . <ZPUN H WLLY YVIV[ [V PTWYV]L [OL ZVJPHS ZRPSSZ
VM JOPSKYLU H[ YPZR MVY LTV[PVUHS ILOH]PVYHS KPZVYKLYZ 1V YUHS VM ,TV PVUHS )LOH PVYHS P VYKLY
2PT 2 / 3LL / : *OHUN : 1 )HL 4 1 ;LHJOLYZ WLYZWLJ[P]LZ VU [OL ZL VM H YVIV[ MVY LK JH[PVUHS W YWVZL
MVY JOPSKYLU SHILSLK P[O H [PZT 1V YUHS VM ,TV PVUHS )LOH PVYHS P VYKLY
2PT 2 / 3LL / : 2PT 1 / *OV 1 4 2PT : / 3LHYUPUN L_WYLZZP]L HJ[P]P[PLZ [V JOPSKYLU P[O H [PZT [OYV NO TH[LYPHS
WYLZLU[LK ]PH YVIV[ 1V YUHS VM ,TV PVUHS )LOH PVYHS P VYKLY

PU[LYUH[PVUHSZWLK JVT
IJSE 9VIV P Y PV V O J VTT PJ PV VJP S IPSP P VM OPS Y

2PT 1 / 3LL / : 7HYR * 1 )HL 4 1 2HUN 1 ) ,MMLJ[Z VM PU[LY]LU[PVU WYVNYHT ZPUN VU ZVJPHS ILOH]PVY HUK WSH
HJ[P]P[PLZ MVY V UN JOPSKYLU P[O KL]LSVWTLU[HS KLSH OL 1V YUHS VM :WLJPHS OPSKYLU ,K JH PVU
2PT * . 3LL / : :OPU / 2HUN 1 ) 2PT 4 : *OVP 4 ; :VUN ) : ,Z[HISPZOTLU[ VM YLTV[L Z WWVY[ Z Z[LT
MVY JOPSKYLU P[O KPZHIPSP[PLZ [OYV NO [OL ZL VM YVIV[Z 1V YUHS VM :WLJPHS ,K JH PVU 9LOHIPSP H PVU :JPLUJL
2PT : ) :OPU / ( YL]PL VM Z[ KPLZ VU [OL HWWSPJH[PVU VM YVIV[Z MVY PTWYV]LTLU[ VM ZVJPHS ZRPSSZ MVY JOPSKYLU P[O H [PZ[PJ
ZWLJ[Y T KPZVYKLY HUK LTV[PVUHS HUK ILOH]PVYHS KPZVYKLY 1V YUHS VM :WLJPHS ,K JH PVU HUK 9LOHIPSP H PVU :JPLUJL
2PT * . :VUN ) : :[ KPLZ VU HWWSPJH[PVU VM YVIV[ YLTV[L Z WWVY[ Z Z[LT -VJ ZLK VU ZWLJPHS LK JH[PVU
1V YUHS VM :WLJPHS ,K JH PVU 9LOHIPSP H PVU :JPLUJL
2VZ[Y IPLJ = 2Y JR 1 *VSSHIVYH[P]L YLZLHYJO WYVQLJ[ KL]LSVWPUN HUK [LZ[PUN H YVIV[ HZZPZ[LK PU[LY]LU[PVU MVY JOPSKYLU P[O H [PZT
-YVU PLY PU 9VIV PJ HUK (0
2 VU 1 4 2PT ; 3HUN HNL LU]PYVUTLU[ HUK PU[LYHJ[PVU VM H JOPSK P[O TVKLYH[L H [PZT *VTWHYPZVU VM YVIV[ TLKPH[LK
PUZ[Y J[PVU ZWLJPHS LK JH[PVU HUK PUJS ZP]L LK JH[PVU ZL[[PUNZ 1V YUHS VM OL 2VYLHU ( VJPH PVU MVY 7LY VU PO( P T
3LL 3 OL LMMLJ VM HU HJ P P PUN HU PU LSSPNLU YVIV VU OL JVTT UPJH PVU VM JOPSK P O H P T 4HZ[LY Z KPZZLY[H[PVU
:LV S 5H[PVUHS <UP]LYZP[ VM ,K JH[PVU :LV S 2VYLH
3LL : *OV . 4 ( Z[ K VU WSH JLU[LYLK HJ[P]P[PLZ ZPUN HU PU[LSSPNLU[ YVIV[ PU HU PUJS ZP]L LHYS JOPSKOVVK ZWLJPHS LK JH[PVU
1V YUHS VM ,HYS OPSKOVVK ,K JH PVU
3LL : / 7HYR , / LHJOPUN L JLW PVUHS JOPSKYLU PU PUJS P L L PUN
:WLJPHS LK JH PVU N PKL MVY PU LNYH LK JSH LHJOLY YK LKP[PVU :LV S /HRQPZH
7HYR * 1 :OPU / 3LL / : 2PT 1 / 2PT : 1 2HUN 1 ) 2PT * . ;OL Z[ K VU LMMLJ[ VM PU[LY]LU[PVU MVY PTWYV]PUN
ZVJPHS ZRPSS ZPUN [LSL WYLZLUJL YVIV[ MVY JOPSKYLU P[O LTV[PVUHS HUK ILOH]PVYHS KPZVYKLY H[ YPZR
1V YUHS VM ,TV PVUHS )LOH PVYHS P VYKLY
7VW * ( :PT [ 9 , 7PU[LH : :HSKPLU 1 9 Z ( : =HUKLYMHLPSSPL 1 +H]PK + 6 3LMLILY + =HUKLYIVYNO[ )
:VJPHS YVIV[Z ]Z JVTW [LY KPZWSH KVLZ [OL H ZVJPHS Z[VYPLZ HYL KLSP]LYLK THRL H KPMMLYLUJL MVY [OLPY LMMLJ[P]LULZZ VU (:+ JOPSKYLU&
1V YUHS VM ,K JH PVUHS VTW PUN 9L LHYJO
9VIPUZ ) +H [LUOHOU 2 ,UJV YHNPUN ZVJPHS PU[LYHJ[PVU ZRPSSZ PU JOPSKYLU P[O H [PZT WSH PUN P[O YVIV[Z ( JHZL Z[ K L]HS H[PVU
VM [YPHKPJ PU[LYHJ[PVUZ PU]VS]PUN JOPSKYLU P[O H [PZT V[OLY WLVWSL WLLYZ HUK HK S[Z HUK H YVIV[PJ [V ,UMHUJL
:HHKH[aP 4 5 7LUUPUN[VU 9 * LSJO 2 * .YHOHT 1 / :THSS NYV W [LJOUVSVN HZZPZ[LK PUZ[Y J[PVU ]PY[ HS [LHJOLY
HUK YVIV[ WLLY MVY PUKP]PK HSZ P[O H [PZT ZWLJ[Y T KPZVYKLY 1V YUHS VM H P T HUK KL LSVWTLU HS KP VYKLY
:OPU *OYPZ[PUL HYLNP PUN ILOH PVY YHPUPUN P O HU HUPTHS YVIV MVY JOPSKYLU P O H P T WLJ Y T KP VYKLY 4HZ[LY Z KPZZLY[H[PVU
: UNR UR HU <UP]LYZP[ :LV S 2VYLH
:OPU / 7HYR * 1 3LL / : 2PT 1 / 2PT : 1 2HUN 1 ) 2PT * . ( Z[ K VU [OL YLZWVUZL VM LTWH[OPJ
HUK ULNV[PH[VY PUP[PH[PVUZ VM [OL JOPSKYLU H[ YPZR MVY LTV[PVUHS HUK ILOH]PVYHS KPZVYKLYZ [OYV NO [OL [LSLWYLZLUJL YVIV[ HZZPZ[LK PU[LY]LU[PVU
1V YUHS VM ,TV PVUHS )LOH PVYHS P VYKLY
:PS]LYH ;H PS + )YHKMVYK + 9VILY[Z H[LZ * ;HSR [V TL ;OL YVSL VM O THU YVIV[ PU[LYHJ[PVU PU PTWYV]PUN ]LYIHS
JVTT UPJH[PVU ZRPSSZ PU Z[ KLU[Z P[O H [PZT VY PU[LSSLJ[ HS KPZHIPSP[ 7YVJLLKPUN VM OL O 0,,, 0U LYUH PVUHS : TWV P T
VU 9VIV HUK THU 0U LYHJ P L VTT UPJH PVU 96 4(5
:PS]LYH ;H PS + 9VILY[Z H[LZ * :VJPHSS HZZPZ[P]L YVIV[Z [V LUOHUJL SLHYUPUN MVY ZLJVUKHY Z[ KLU[Z P[O PU[LSSLJ[ HS
KPZHIPSP[PLZ HUK H [PZT 7YVJLLKPUN VM OL O 0,,, 0U LYUH PVUHS : TWV P T VU 9VIV HUK THU 0U LYHJ P L VTT UPJH PVU
96 4(5
:PT [ 9 , =HUKLYMHLPSSPL 1 7LJH ( =HU KL 7LYYL . =HUKLYIVYNO[ ) *OPSKYLU P[O H [PZT ZWLJ[Y T KPZVYKLYZ THRL
H MY P[ ZHSHK P[O WYVIV [OL ZVJPHS YVIV[ (U PU[LYHJ[PVU Z[ K 1V YUHS VM H P T HUK KL LSVWTLU HS KP VYKLY
:V * *OLUN * / 3HT VUN ; 3H / HUN 5N 2 * ; UN / * VUN 9VIV[ IHZLK WSH KYHTH
PU[LY]LU[PVU TH PTWYV]L [OL UHYYH[P]L HIPSP[PLZ VM *OPULZL ZWLHRPUN WYLZJOVVSLYZ P[O H [PZT ZWLJ[Y T KPZVYKLY ( WPSV[ Z[ K
1V YUHS VM ( P T HUK L LSVWTLU HS P VYKLY
:VUN : 4 2PT , / 2 HR 1 9 (UHS ZPZ VM ZVJPHS YVIV[ ZHNL WH[[LYUZ HUK ZH[PZMHJ[PVU VM Z[ KLU[Z P[O WO ZPJHS KPZHIPSP[PLZ
2VYLHU 1V YUHS VM 7O PJHS 4 S PWSL LHS O P HIPSP PL
U : : 2PT / : *OVP 1 : 7HYR : 2 ( YVIV[PJ Z Z[LT P[O ILOH]PVYHS PU[LY]LU[PVU MHJPSP[H[PUN L L JVU[HJ[ HUK MHJPHS LTV[PVU
YLJVNUP[PVU VM JOPSKYLU P[O H [PZT ZWLJ[Y T KPZVYKLYZ 1V YUHS VM 2VYLH 9VIV PJ :VJPL
AOLUN A 5PL . : HUZVU ( LP[SH M ( HYYLU A :HYRHY 5 ( YHUKVTPaLK JVU[YVSSLK [YPHS VM HU PU[LSSPNLU[ YVIV[PJ YLZWVUZL
[V QVPU[ H[[LU[PVU PU[LY]LU[PVU Z Z[LT 1V YUHS VM ( P T HUK L LSVWTLU HS P VYKLY

+VP O[[WZ KVP VYN PQZL