Content uploaded by Özlem Yeter
Author content
All content in this area was uploaded by Özlem Yeter on May 24, 2021
Content may be subject to copyright.
Are Refugee Bilingual Children Disadvantaged in Their
Cognitive and Linguistic Abilities?
Özlem Yeter, Hugh Rabagliati, and Duygu Özge
1. Introduction
The Syrian civil war broke out in 2011, and caused has more than five million
people to seek refuge outside the country. In the more than eight years since the
war started, a large population of children have thus grown up as refugees,
particularly in Turkey, which has absorbed the large majority of displaced Syrian
families. However, very little is known about the cognitive and language
development of these children. For instance, it is unclear how their traumatic
displacement experiences might impact their development and how their language
abilities would develop. Our focus group consists of individuals who were forced
to leave their countries because of the life-threatening events they experienced.
Nevertheless, they must learn to make their way in Turkish society, learning a
new language, a new culture, and developing mature cognitive and social skills.
For many displaced children, it is imperative to acquire a new community
language – i.e., Turkish – but it is unclear what environment is optimal to
encourage this. For instance, does immersive education in a Turkish language
school promote strong Turkish language development or does it hinder children’s
cognitive and social development, since they may fail to engage in class or interact
with their peers, leading to worse outcomes?
Previous studies highlight that the children who are exposed to such
experiences caused by war suffer from post-traumatic stress disorder (PTSD) and
mental health problems (Attanayake, et al., 2009; Ehntholt & Yule, 2006; Eruyar,
Maltby, & Vostanis, 2019). A study by zer, Sirin and Oppedal (2016) also
* This study was funded by the British Academy Newton Advanced Fellowship 2017
RD - NAF2R2\100075. We grateful to Middle East Technical University Language and
Cognitive Development Laboratory members for their contribution in coding of the data.
We also thank Prof. Ute Bohnacker and Rima Haddad for providing us with the materials
and the sources we needed. Also, we thank ASAM for helping us reach the refugee
families. And finally, we are grateful to p arents and especially children for participating in
this study.
Özlem Yeter, Middle East Technical University, ozlemyeter1@gmail.com; Hugh
Rabagliati, The University of Edinburgh; hugh.rabagliati@ed.ac.uk; Duygu Özge, Middle
East Technical University, duyguozge@gmail.com.
© 2021 Özlem Yeter, Hugh Rabagliati, and Duygu Özge. Proceedings of the 45th annual
Boston University Conference on Language Development, ed. Danielle Dionne and
Lee-Ann Vidal Covas, 790-804. Somerville, MA: Cascadilla Press.
revealed that the Syrian refugee children who live in refugee camps in Gaziantep
(a city in southern Turkey) show the symptoms of depression and PTSD, as well
as aggression and psychosomatic symptoms at a serious level although 20% of
the participants reported no experiences of war. Tösten, Toprak and Kayan (2017)
interviewed teachers and focused on the learning environments of Syrian refugee
children in a Turkish public school. The teachers participating in the study
observed that the refugee children suffer from PTSD and they show poorer
performance in verbal classes due to the language barrier. This study is in line
with several other studies focusing on the integration of Syrian refugee children
into Turkish school system (Şeker & Sirkeci, 2015; Ugurlu, Akca, & Acarturk
2016; Sarmini, Topçu, & Scharbrodt, 2020). The strong relationship between the
quality of education and children’s cognitive development is well-documented in
the literature (Kaplan, Stolk, Valibhoy, Tucker, & Baker, 2016).
One recent study has examined the association between refugee status and
non-linguistic cognitive abilities. Chen et al. (2019) tested Syrian refugees in
Jordan and Jordanian non-refugees between the ages 12-18 for their working-
memory and inhibitory control abilities and their relation to adversity. They found
that the Syrian refugees had more traumatic experiences, showed more PTSD
symptoms and had a higher poverty rate. However, no association between
traumatic events or PTSD was found with executive functions (EF). Still,
children’s working memory was found to be associated with family poverty and
and the child’s years of education. However, it is important to note that the refugee
children in that study had resettled in an Arabic speaking country, children’s
bilingual status was not specified (i.e. monolinguals or bilinguals), and their
language abilities were not tested. This is noteworthy because there are well-
known (although controversial) claims that executive functions may correlate
with bilingualism.
Specifically, a number of studies show evidence for a “bilingual advantage”
in cognitive abilities such as working memory and executive functions (Bialystok,
2017; Whitford & Luk, 2019). Activating and suppressing multi-lingual system
may indeed be enhancing executive function abilities (Whitford & Luk, 2019).
However, other studies claim there is either no bilingual cognitive advantage
(Hilchey & Klein, 2011) or it arises due to other factors such as IQ (Brydges et
al., 2012) or SES (Morton & Harper, 2007; cf., Blom et.al., 2014).
Although the issues of adaptation, mental health and school environment are
investigated in the literature, we could not identify a publication addressing the
cognitive abilities of bilingual refugee children. We thus aimed to provide an
initial pilot assessment of how refugee bilinguals perform in their language and
cognitive abilities. We know that early life traumas may deteriorate cognitive
functions in monolinguals (De Bellis, Hooper, Spratt, & Woolley, 2009; Pechtel
& Pizzagalli, 2011; Wilson, Hansen, & Li, 2011). Refugee children suffer from
depression or post-traumatic stress disorder (Özer et al., 2016). In the context of
the bilingual advantage, however, it may be that being a bilingual could act as a
protective shield, enhancing cognitive skills despite early life traumas.
791
In this study we investigate the current language and cognitive abilities of the
bilingual Syrian refugee children in comparison to non-refugee bilingual children
living in Turkey. We examine four executive function (EF) components: working
memory, inhibitory control, fluid intelligence and shifting ability. In addition to
these cognitive tests, we assess the children’s language abilities through narrative
and vocabulary tests both in Arabic and Turkish. The findings will provide an
initial evidence base for clinical scientists, educators, and policy makers in
planning intervention studies, curriculums, or educational policies. The results
will also enable us to test theoretical claims about how bilingualism affects
cognitive control abilities.
2. Method
2.1. Participants
49 children participated in this study and they were grouped based on their
refugee status (i.e. refugee bilinguals, non-refugee bilinguals). 25 of the
participants were early bilingual (Turkish-Arabic) Syrian refugee children living
in Turkey (18 girls, 7 boys) at the age of 9 (M=9;5 Range=8;11-10;03). The
remaining 24 children were non-refugee children from a minority group in Hatay
region (a city in southern Turkey) who are simultaneous bilingual speakers of
Arabic and Turkish (12 girls, 12 boys). We considered the Hatay children a natural
comparison group, that could be matched in age (M=9;4 Range=8;06-10;04) and
language background (Arabic and Turkish). Sampling of the refugee children was
made through Association for Solidarity with Asylum Seekers and Migrants
(ASAM) in Ankara and Gaziantep offices. Local authorities and teachers in the
villages with Arabic speaking minorities were contacted for the recruitment of
non-refugee bilingual children.
Mother’s years of schooling ranged between 0-16 (M=5.96 SD=3.91) in
Syrian refugee group. Two of the refugee children had lost both parents at a very
early age and had been raised by their grandparents; their data was excluded in
calculation of mother schooling. Mother’s years of education was between 5-8
years (M=6.13 SD=1.48) in the non-refugee group.
Refugee children’s age of arrival to Turkey varied between 2 and 7 (M= 5.44
SD= 1.71). Although systematic exposure to Turkish starts with schooling, onset
of exposure to the language starts after arriving in Turkey, through interactions in
the neighbourhood, through family members, and through television. No
additional language (e.g. Turkmen, Kurdish) was spoken by refugee families. For
the non-refugee group, both Arabic and Turkish were spoken at home, thus
exposure to both languages starts at home from birth. Families were asked to
informally rate their children’s proficiency in both languages. All refugee children
were reported to be more fluent in Arabic, while Turkish was the stronger
language for all non-refugee children according to their families. Children’s
performance on vocabulary and narrative tests are consistent with families’
language ratings.
792
Regardless of their refugee status, all participants attended Turkish state
schools with Turkish as the medium of instruction and they received no schooling
in Arabic. 80% of the refugee children had been going to Turkish-medium state
schools for at least 3 years at the time of testing, and years of schooling ranged
between 1-5 years (M=2.88 SD=.83). All of the children in the non-refugee Hatay
group had received at least 3 years of schooling at the time of testing (M=3.83
SD=.56) with the range of 3-5 years.
The refugee families reported that they and their children had experienced
significant hardships and traumatic events. Two out of 25 children lost both
parents at the age of 2 and were raised by their grandparents, 5 of the children
have lost their fathers, 1 of the children is abandoned by her mother and 2 of the
children have divorced parents. All of the parents/grandparents reported that their
houses were either destroyed or damaged prior to their arrival to Turkey. The
families also mentioned experiencing financial hardships.
2.2. Materials and Procedure
A set of cognitive and linguistic assessment tasks was administered by two
experimenters for this study. The bilingual experimenter, also the first author of
this paper, is a bilingual speaker of Arabic (Levantine dialect spoken in Hatay)
and Turkish. The bilingual experimenter administered Arabic tests and cognitive
tests in refugee group and all tests in non-refugee group. The children were free
to choose the language of administration for cognitive tasks, because we did not
want their language proficiency to affect their cognitive performance. The second
experimenter, who is a native speaker of Turkish and a master’s student at the
Department of Cognitive Science at METU, was trained and provided with the
manuals by the first author. The Turkish experimenter administered Raven’s
Coloured Progressive Matrices, Turkish narrative and vocabulary tests for Syrian
refugee children. Below we present the materials and the procedure for each test
we used.
2.2.1. Backward Digit-Span Task
Being a component of Wechsler Intelligence Scales for Children-Revised
(Wechsler, 1974), Memory for Digit Span assessment consists of two parts:
forward digit-span dealing with short-term memory and backward digit-span
dealing with working memory. There are 14 sets of increasing digits in both parts,
thus 28 sets in total. The children were first asked to repeat the set of digits uttered
by the experimenter in the same order for forward digit-span task and they were
expected to do the same in reverse order for the backward digit-span task. Both
tasks continue until the child commits two consecutive errors in two series of
digits in the same length. Forward digit span was used only to familiarise the
children with the second part and since we are interested in children’s working
memory abilities, we analysed only the backward digit-span task scores. The
children received 1 point for each series they completed successfully and the
793
maximum score in this test was 14. All the children in the non-refugee group
preferred to complete the task in Turkish while 12 out of 25 children in the refugee
group chose to hear the digits in Arabic.
2.2.2. Raven’s Coloured Progressive Matrices
Raven’s Coloured Progressive Matrices (CPM) is a non-verbal measure of
fluid intelligence and abstract reasoning abilities in children between the ages of
5 and 11 (Raven, Raven, & Court, 1998). For Turkey, reliability and validity of
CPM has been tested for children between the ages of 4-6, and it was found to be
strongly correlated with Bender-Gestalt Test, TONI-3 Test and WISC-R Test
Scores (Kargin, 2017). The test consists of 3 sets (A, AB and B) with 12 items
each. Problem items are ordered in an ascending difficulty. In each problem, the
children are expected to complete a coloured drawing or a matrix by choosing the
appropriate piece among six alternatives. The manual (Raven et al., 1998) was
followed for the administration of the test and the answers were transferred to the
scoring sheet by the experimenters. 1 point is given for every correct response and
highest possible score for this test is 36. The sample sizes of the groups are not
equal for the Raven’s test since data collection process was interrupted by
coronavirus outbreak. 17 out of 25 Syrian children and 10 out of 24 children from
Hatay region took the test.
2.2.3. Stroop-like Happy-Sad Task
We used OpenSeasame (Mathôt, Schreij, & Theeuwes, 2012) to create a
computerised version of the happy-sad task design by following the experiment
design in the paper by Lagattuta, Sayfan and Mansour (2011) to test inhibitory
control in children. There were 20 faces half of which looks happy whereas the
other half looks sad. The faces for this task were taken from NimStim (Tottenham
et al., 2009) from http://www.macbrain.org/faces/index.htm. Assignment of the
emotion faces were counterbalanced by age and gender (i.e. both in refugee and
non-refugee groups, half of the children saw male faces while the other half saw
female faces).The children were instructed and given eight practice trials. Each
child saw 10 happy and 10 sad faces (twenty in total) in mixed order and they
were asked to tell the opposite of the what they see (i.e. say “happy” when they
see a “sad” face and vice versa) as fast as they can. The order of the faces was
automatically randomised every time the test was run.
The same experimenter collected all the children’s responses in both groups
by pushing a keyboard button assigned for “happy” and “sad” as the children
orally responded to the faces. Whole testing process was audio-recorded and the
keyboard responses were checked against oral responses before analysis.
Percentage of the correct responses out of 20 responses and total reaction times
were calculated for this task. The non-refugee group responded in Turkish while
13 out of 25 refugee children responded in Arabic.
794
2.2.4. Berg’s Card Sorting Task
To assess the shifting ability in children, we used 64-trial version of Berg
Card Sorting Test (BCST; Mueller, 2013), a computerised version of Wisconsin
Card Sorting Test (Berg, 1948). 4 stimulus cards differing with respect to colour,
shape and quantity are presented on the screen throughout the test. The
participants are expected to sort the response card with one of the stimulus cards
in every trial through feedback (i.e. right or wrong). The feedback appears on the
screen after each response and the experimenter repeats it orally. The card sorting
rule changed after ten consecutive correct responses, but the children were not
given this information. All children were given the same instructions and were
expected to give five consecutive correct responses to be able to begin the task.
Wrong responses caused by the participant’s use of previous sorting rule are
called perseverative errors and lower number of perseverative errors indicates
better shifting ability (Miyake et al., 2000). We analysed the percentages of
perseverative errors which is found by division of total perseverative errors by
number of trials then multipling by 100.
2.2.5. Turkish Expressive and Receptive Language Test (TİFALDİ)
Turkish Expressive and Receptive Language Test (TİFALDİ) was used to
assess receptive and expressive vocabulary skills of the children (Berument &
Güven, 2013). TİFALDİ consists of black-and-white illustrations with 101 items
in receptive sub-scale and 80 items in expressive sub-scale. On receptive test,
there are four numbered drawings on each image plate and the child points to the
picture that best describes the word uttered by the experimenter. On the expressive
vocabulary test, the child sees only one drawing and is asked to name the related
drawing. The children’s responses were coded to an excel sheet during the testing.
The experimenter starts with the age-relevant item for both tests and the children
are expected to give 8 consecutive answers to achieve a basal score. The number
of correct responses after 8 consecutive correct responses are added on the basal
score. This gives us the raw score and the standard score is determined based on
the manual given for different age-ranges.
2.2.5.1. Modification of TİFALDİ to Test Arabic Vocabulary
To make sure we use the same vocabulary task across the two languages, we
adapted the TIFALDI test into Arabic. Five Syrian university students translated
the items in both sub-tests of TİFALDİ individually. The experimenter compared
the translations. Later, the Arabic version of the test was applied to 6 different
Syrian university students who did not see the drawings before. The final version
of the test was determined after several revisions, feedbacks and discussions with
all the Syrian university students. Item number 73, 100 and 101 (devirmek/to
knock down, viyadük/viaduct, faraş/dustpan) were excluded from receptive sub-
test for not having Arabic equivalents and item number 36 (papatya/daisy) was
795
replaced by “jasmine” for being a more widely-known flower than “daisy”. The
final version of the test was also applied to 4 bilingual speakers of Arabic and
Turkish from Hatay region and they reported that the language and pictures were
appropriate.
2.2.6. Multilingual Assessment Instrument of Narratives (MAIN)
We assessed children’s Turkish and Arabic narrative abilities using
Multilingual Assessment Instrument of Narratives (MAIN, Gagarina et al., 2012;
2019). There are four stories in MAIN, each consisting of 3 episodes, and with
each episode depicted in two pictures. Stories are in parallel with regards to
complexity: Dog Story with Cat Story and Baby Birds with Baby Goats. The latter
parallel stories are slightly more complex than the former two (Kornev &
Balčiūnienė, 2014). Only one language was assessed at a time and the test was
administered in two sessions. The order of the stories and the languages were
counterbalanced with regards to language and story type. The children started
with either of the less complex parallel stories and continued with one of the more
complex ones. In the end, the children told one simpler and one more complex
story for both languages, so four stories in total. The mean of the days between
the testing of two languages in non-refugee children was 9.4 (3 to 30 days), and
9.2 for the refugee children (3 to 50 days).
After explaining the task briefly, the experimenter placed the envelopes
containing the copies of the same story on the table and asked the child to choose
one. This led the child to think that the experimenter did not know the story,
controlling for the shared knowledge effect (Gagarina et al., 2012). When the
child was ready, the experimenter folded the sequence and let only pictures 1-2
(first episode) be visible. After the child finished narrating the first two pictures,
the experimenter unfolded the pictures 3-4 (second episode) next to previous
pictures. When the child finished narrating all episodes, all of the pictures were
visible to him/her. All testing sessions were audio-recorded.
A group of university students contributed to both transcription and scoring
process of the Turkish narratives. The transcriptions were checked by transcribers
and the bilingual experimenter. The Arabic narratives were transcribed by two
bilingual speakers of Arabic and Turkish. The Arabic transcriptions were later
checked and revised both by the bilingual experimenter and two native speakers
of Arabic. The transcriptions were made in CHAT-format and later analysed with
program CLAN (MacWhinney, 2000). The narratives were scored in two main
sections: production at macrostructural level and comprehension.
Macrostructure of a narrative is related to the plot and the organisation of the
story dealing with the sequence of events. MAIN manual was followed for the
scoring of the narratives. The children were awarded one point for each of the
story elements and internal state terms (IST) presented in Table 1. The maximum
score a child can get in this section is 17.
796
Table 1. Story structure components in MAIN
Episode 1 Episode 2 Episode 3
Setting
(time and
place)
IST as initiating
event IST as initiating
event IST as initiating
event
Goal Goal Goal
Attempt Attempt Attempt
Outcome Outcome Outcome
IST as reaction IST as reaction IST as reaction
Each narrative in MAIN is followed by 10 comprehension questions. To be
able to tell a good story, children need to understand the story schema. This
section assesses children’s the ability to interpret the story pictures, story
characters’ goals and their internal states and the ability to explain and reason
them orally. The ability to answer the questions correctly is related to children’s
memory, language abilities as well as Theory of Mind (Lindgren, 2018). Children
were given one point for each correct answer, thus 10 points in maximum.
3. Results
Below we present the results for the cognitive and the linguistic measures
separately. For each measure, we conducted one-way MANCOVA with the score
from the test applied as our dependent variable and Group (refugee, non-refugee)
as our independent variable. All analyses were conducted using the SPSS statistics
program Version 24.0 for windows (IBM Corp., 2016). The label “Hatay” is used
to refer to non-refugee children and the refugee children are labelled as “Syria” in
the graphs.
3.1. Cognitive Tests
The analysis of the backward digit-span working memory task showed that
non-refugee children performed significantly better than refugee children [F(1,45)
= 13.40, p = .001, ηp2 = .23] (Figure 1), so refugee children had poorer working
memory abilities compared to non-refugee children.
For the Raven’s CPM test of nonverbal IQ, there was also a significant effect
of Group [F(1,25) = 9.79, p < 0.05, ηp2 = .28] such that non-refugee children
performed significantly better in fluid intelligence and reasoning abilities than
the refugee children (Figure 2).
On the other hand, although the non-refugee children had more correct
responses in the Stroop-like happy-sad task, the effect of group was not significant
[F(1,45)= .1.32, p = .225, ηp2 = .029], which indicated that the two groups did
not differ with respect to their inhibition abilities (Figure 3).
Finally, in Berg’s card sorting task (BCST), refugee children made
significantly more perseverative errors [F(1,45) = 7.89, p < .05, ηp2 =.14],
indicating more difficulty with cognitive flexibility and shifting abilities
compared to non-refugee children (Figure 4).
797
0
2
4
6
8
10
12
14
Hatay Syria
Scores
0
4
8
12
16
20
24
28
32
36
Hatay Syria
Scores
Figure 1. Backward digit-span task scores Figure 2. Raven’s CPM scores
90
92
94
96
98
100
Hatay Syria
Correct responses
0
5
10
15
20
25
Hatay Syria
Errors
Figure 3. Percentage of correct responses Figure 4. Percentage of errors in
In happy-sad task BCST
3.2. Linguistic Tests
The analysis of vocabulary test showed that refugee children performed
significantly better than non-refugee children in the receptive vocabulary sub-test
in Arabic [F(1,45) = 75.90, p < .001, ηp2=.63] (Figure 5) while an opposite pattern
was observed in the Turkish version of the test such that refugee children
performed poorer than non-refugee children in the receptive vocabulary sub-test
in Turkish [F(1,43) = 11.28, p < .001, ηp2=.72] (Figure 6). In Arabic expressive
vocabulary subtest, we were not able to compare the two groups, because none of
the children in non-refugee group was able to achieve a basal score in Arabic;
however, non-refugee children had significantly higher scores in Turkish
expressive vocabulary subtest compared to refugee children [F(1,38) = 186.03, p
< .00, ηp2 = .83].
798
0
20
40
60
80
100
120
Hatay Syria
Score
0
20
40
60
80
100
120
Hatay Syria
Score
Figure 5. Arabic receptive vocabulary Figure 6. Turkish receptive vocabulary
scores scores
According to the analysis of the production of the narratives in Turkish, the
non-refugee children had significantly higher scores both in the first [F(1,47) =
12.77, p = .001, ηp2 = .21] and the second stories they told, [F(1,47) = 19.719, p
< .001, ηp2 =.29] (see Figure 7).
0
2
4
6
8
10
12
14
16
Hatay Syria Hatay Syria
Story 1 Story 2
Scores
Figure 7. Narrative macrostructure production scores (out of 17) of first and second
Turkish stories by groups
As for the narrative production in Arabic, the refugee children scored
significantly higher both in the first Arabic stories [F(1,47) = 30.54, p < .001, ηp2
= .39] and the second stories [F(1,47) = 23.29, p < .001, ηp2 = .33] (see Figure 8).
When we further analysed the two groups on the basis of their dominant language,
we did not find any difference neither in the first stories [t(23)= .07, p=.9] nor in
the second ones [t(23)=.46, p=.6]. When the analysis was made on the basis of
their non-dominant language, the results were not significant (first stories [t(23)=
-1.21, p=.2], second stories [t(23)= -.22, p=.8]).
799
0
2
4
6
8
10
12
14
16
Hatay Syria Hatay Syria
Story 1 Story 2
Scores
Figure 8. Narrative macrostructure production scores (out of 17) of first and second
Arabic stories by groups
Finally, the analysis of narrative comprehension scores showed an effect of
group only in the second stories told in Turkish [F(1,47)=9.09, p < .05, ηp2=.16]
such that non-refugee children had significantly higher performance compared to
the refugee group; however, the groups did not differ in the second stories they
told in Arabic, and no significant difference was found between groups in first
stories in neither of the languages. Therefore, the analyses are not presented here.
When the groups’ comprehension scores in both stories were analysed on the basis
of their dominant language, the results showed no significance (first stories
[t(23)=.00, p=1], second stories [t(23)=1.58, p=.1]). However, results showed
significance when the comprehension scores in both stories were analysed on the
basis of the groups’ non-dominant languages (first stories [t(23)=2.05, p=.04],
second stories [t(23)=2.27, p=.03] such that non-refugee group scored
significantly higher in their less dominant language (i.e. Arabic) for both stories
while the refugee group scored significantly lower in their less dominant language
(i.e. Turkish) for both stories in comprehension section.
4. Discussion and Conclusion
Previous studies have shown that early life traumas, specifically war
experiences may cause PTSD, stress and mental health problems in children;
however, there
traumas may deteriorate cognitive functions (De Bellis, Hooper, Spratt, &
Woolley, 2009; Pechtel & Pizzagalli, 2011; Wilson, Hansen, & Li, 2011). Our
study provides an initial evaluation of how being a refugee might influence
cognitive and linguistic abilities of a developing bilingual child. We used a
inhibitory control and shifting ability and a set of linguistic tasks testing children’s
vocabulary and narrative abilities.
have been a limited number of studies showing that early life
set of cognitive tasks testing children’s working memory, fluid intelligence,
The Syrian refugee children in this study provided lower scores in tasks of
working memory, fluid intelligence, and shifting ability, compared to non-refugee
800
children. This is in-line with other studies showing that early life traumas may
deteriorate cognitive functions (De Bellis, Hooper, Spratt, & Woolley, 2009;
Pechtel & Pizzagalli, 2011; Wilson, Hansen, & Li, 2011). Both groups performed
similarly in an inhibitory control task and this finding is parallel with the findings
of the study conducted by Chen et al. (2019). Overall, however, the refugee
children showed lower scores in the non-linguistic tasks. We return to this below,
in the context of the bilingual advantage.
The picture for our linguistic tasks was nuanced. While the analysis of
narrative production scores showed no significant difference on the basis of
language dominancy (i.e. Turkish being the dominant language in non-refugee
group, Arabic being the dominant language in refugee group), the comparison on
the basis of their non-dominant language revealed that Turkish receptive
vocabulary abilities of Syrian children were better than the Arabic receptive
vocabulary abilities of non-refugee children. This may be related to Turkish being
the majority/community language, whereas exposure to Arabic in non-refugee
group is limited to the home context. This is in line with the finding that the non-
refugee children were not able to achieve a basal score in Arabic expressive
vocabulary test, and also with families’ rating of the children’s language
proficiency. It appears that linguistic abilities of both groups were roughly
matched: refugee children’s Arabic skills appeared to be similar to non-refugee
children’s Turkish skills, and vice versa. This is important in the context of the
refugee children showing lower performance across the cognitive tasks assessed;
their matched language abilities indicate that these cognitive tasks may not be
picking up on inherent differences between the groups, but rather differences that
are a function of the children’s very different backgrounds, including significant
traumatic experiences in the refugee children.
This initial study was partly motivated by the possibility that becoming
bilingual might offer some cognitive protective factor to refugee children, in the
context of claims of a bilingual advantage. We found no evidence for this. Our
refugee children, despite being bilingual, scored significantly below non-refugee
children in our non-linguistic tasks.
Of course, significant additional data will be required to draw strong
conclusions about any of the questions examined here, such as how displacement
experiences affect cognitive and language development. This initial small-scale
study provides an important first picture, with evidence for strengths in the
refugee’s heritage languages, and evidence consistent with prior claims that early
life traumas (i.e. holding refugee status in our case) affect the performance on
cognitive tasks (De Bellis, Hooper, Spratt, & Woolley, 2009; Pechtel &
Pizzagalli, 2011; Wilson, Hansen, & Li, 2011). But a much larger sample size will
be required to test these claims at scale, while also accounting for the highly
heterogenous backgrounds of these participants. Nonetheless, the results highlight
how holding a refugee status might affect executive functions, and they being a
bilingual has not acted as a clear protective shield for refugee children’s cognitive
skills.
801
References
Attanayake, Vindya, Mckay, Rachel, Joffres, Michel, Singh, Sonal, Burkle, Frederick M.,
& Mills, Edward (2009). Prevalence of mental disorders among children exposed to
war: A systematic review of 7,920 children. Medicine, Conflict and Survival, 25(1),
4-19. doi:10.1080/13623690802568913
Berg, Esta A. (1948). A Simple Objective Technique for Measuring Flexibility in
Thinking, The Journal of General Psychology, 39:1, 15-22, doi:
10.1080/00221309.1948.9918159
Bialystok, Ellen (2017). The bilingual adaptation: How minds accommodate
experience. Psychological Bulletin, 143(3), 233-
262. https://doi.org/10.1037/bul0000099
Blom, Elma, Küntay, Aylin, Messer, Marille, Verhagen, Josje & Leseman, Paul (2014).
The benefits of being bilingual: working memory in bilingual Turkish-Dutch
children. Journal of experimental child psychology, 128, 105-19 .
Brydges, Christopher R., Reid, Corinne L., Fox, Allison M., & Anderson, Mike (2012). A
unitary executive function predicts intelligence in children. Intelligence, 40(5), 458–
469. https://doi.org/10.1016/j.intell.2012.05.006
Chen, Alexandra, Panter‐Brick, Catherine, Hadfield, Kristin, Dajani, Rana, Hamoudi,
Amar, & Sheridan, Margaret (2019). Minds Under Siege: Cognitive Signatures of
Poverty and Trauma in Refugee and Non‐Refugee Adolescents. Child Development,
90(6), 1856-1865. doi:10.1111/cdev.13320
De Bellis, Michael D., Hooper, Stephan R., Spratt, Eve G., & Woolley, Donald P. (2009).
Neuropsychological findings in childhood neglect and their relationships to pediatric
PTSD. Journal of the International Neuropsychological Society, 15(6), 868-878.
doi:10.1017/s1355617709990464
Ehntholt, Kimberly A., & Yule, William (2006). Practitioner Review: Assessment and
treatment of refugee children and adolescents who have experienced war-related
trauma. Journal of Child Psychology and Psychiatry, 47(12), 1197-1210.
doi:10.1111/j.1469-7610.2006.01638.x
Eruyar, Şeyda, Maltby, John., & Vostanis, Panos (2019). How do Syrian refugee children
in Turkey perceive relational factors in the context of their mental health? Clinical
Child Psychology and Psychiatry, 25(1), 260-272. doi:10.1177/1359104519882758
Gagarina, Natalia, Klop, Daleen, Kunnari, Sari, Tantele, Koula, Välimaa, Taina,
Balčiūnienė, Ingrida, Bohnacker, Ute, & Walters, Joel (2012). MAIN: Multilingual
Asssessment Instrument for Narratives. ZAS Papers in Linguistics, 56, 1–156.
Gagarina, Natalia, Klop, Daleen, Kunnari, Sari, Tantele, Koula., Välimaa, Taina.,
Bohnacker, Ute & Walters, Joel (2019). MAIN: Multilingual Assessment Instrument
for Narratives. Revised version. ZAS Papers in Linguistics 63.
Hilchey, Matthew D., & Klein, Raymond M. (2011). Are there bilingual advantages on
nonlinguistic interference tasks? Implications for the plasticity of executive control
processes. Psychonomic bulletin & review, 18(4), 625–658.
https://doi.org/10.3758/s13423-011-0116-7
IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 24.0. Armonk, NY:
IBM Corp.
Kaplan, Ida, Stolk, Yvonne, Valibhoy, Madeleine, Tucker, Alan, & Baker, Judy (2016).
Cognitive assessment of refugee children: Effects of trauma and new language
acquisition. Transcultural Psychiatry, 53(1),81-109.
Kargin, Tevhide (2017). Reliability and Validity of Colored Progressive Matrices for 4-6
Age Children. Turkish Journal of Giftedness and Education, 7(1), 19-38.
802
Kazak-Berument, Sibel, & Güven, Ayşe G. (2013). Turkish Expressive and Receptive
Language Test: I. Standardization, reliability and validity study of the Receptive
Vocabulary Sub-Scale. Turkish Journal of Psychiatry, 24, 192-201.
Kornev, Alexandr Nikolayevitch, & Balčiūnienė, Ingrida (2014). LSCD 2014: Workshop
on Late Stages in Speech and Communication Development (pp. 23-26). London,
United Kingdom: University College London.
Lagattuta, Kristin H., Sayfan, Liat, & Monsour, Michael (2011). A new measure for
assessing executive function across a wide age range: Children and adults find happy-
sad more difficult than day-night. Developmental Science, 14(3), 481-489.
doi:10.1111/j.1467-7687.2010.00994.x
Lindgren, Josefin (2018). What is narrative competence? In Developing narrative
competence: Swedish, Swedish-German and Swedish-Turkish children aged 4–6.
Acta Universitatis Upsaliensis. https://doi.org/10.1177/0142723719844089
MacWhinney, Brian (2000). The CHILDES Project: Tools for analyzing talk (3rd Edition).
Hillsdale, NJ: Lawrence Erlbaum.
Mathôt, Sebastiaan, Schreij, Daniel, & Theeuwes, Jan (2012). OpenSesame: An open-
source, graphical experiment builder for the social sciences. Behavior Research
Methods, 44(2), 314-324. doi:10.3758/s13428-011-0168-7
Miyake, Akira, Friedman, Naomi P., Emerson, Michael J., Witzki, Alexander H.,
Howerter, Amy & Wager, Tor D. (2000). The unity and diversity of executive
functions and their contributions to complex “frontal lobe” tasks: A latent variable
analysis. Cognitive Psychology, 41(1), 49–100.
Morton, J. Bruce, & Harper, Sarah N. (2007). What did Simon say? Revisiting the bilingual
advantage. Developmental science, 10(6), 719–726. https://doi.org/10.1111/j.1467-
7687.2007.00623.x
Mueller, Shane T. (2013). PEBL: The Psychology Experiment Building Language
(Version 0.13). Available at: http://pebl.sourceforge.net
Özer, Serap, Sirin, Selçuk R., & Oppedal, Britt (2016). Bahçeşehir study of Syrian refugee
children in Turkey. Unpublished manuscript. Retrieved
from https://www.fhi.no/globalassets/dokumenterfiler/studier/ungkul/bahcesehir-
study-report.pdf
Pechtel, Pia, Pizzagalli, Diego A. (2011). Effects of early life stress on cognitive and
affective function: an integrated review of human literature. Psychopharmacology,
214(1), 55-70. doi:10.1007/s00213-010-2009-2
Raven, Jean, Raven, John C., & Court, John Hugh (1998). Section 2: Coloured Progressive
Matrices (1998 Edition). Introducing the parallel version of the test. In Manual for the
Raven’s Progressive Matrices and Vocabulary Scales. Great Britain: Oxford
Psychologist Press.
Sarmini, Iman, Topçu, Emel, & Scharbrodt, Oliver (2020). Integrating Syrian refugee
children in Turkey: The role of Turkish language skills (A case study in Gaziantep).
International Journal of Educational Research Open, 1, 100007.
doi:10.1016/j.ijedro.2020.100007
Şeker, B. Dilara, & Sirkeci, İbrahim (2015). Challenges for refugee children at school in
Eastern Turkey. Economics & Sociology, 8(4), 122-133.
Tösten, Rasim, Toprak, Mustafa, & Kayan, M. Selman (2017). An Investigation of
Forcibly Migrated Syrian Refugee Students at Turkish Public Schools. Universal
Journal of Educational Research, 5(7), 1149-1160. doi:10.13189/ujer.2017.050709
Tottenham, Nim., Tanaka, James, Leon, Andrew C., McCarry, Thomas, Nurse, Marcella,
Hare, Todd A., Marcus, David J., Westerlund, Alissa, Casey, B.J., Nelson, Charles
803
A. (2009). The NimStim set of facial expressions: judgments from untrained research
participants. Psychiatry Research, 168(3):242-249.
doi:10.1016/j.psychres.2008.05.006
Ugurlu, Nilay, Akca, Leyla, & Acarturk, Ceren (2016). An art therapy intervention for
symptoms of post-traumatic stress, depression and anxiety among Syrian refugee
children. Vulnerable Children and Youth Studies, 11 (2), 89–102. doi:
10.1080/17450128.2016.1181288 .
UNHCR (2020). Syria Regional Refugee Response. Retrieved December 30, 2020, from
http://data2.unhcr.org/en/situations/syria
Wechsler, David (1974). Wechsler intelligence scale for children-revised. New York:
Psychological Corporation
Whitford, Veronica, & Luk, Gigi (2019). Comparing executive functions in monolinguals
and bilinguals: Considerations on participant characteristics and statistical
assumptions in current research. In I. Sekerina, L. Spradlin, & V. Valian
(Eds.), Bilingualism, executive function, and beyond: Questions and insights (pp. 67-
80). Amsterdam: John Benjamins.
Wilson, Kathryn R., Hansen, David J., & Li, Ming (2011). The traumatic stress response
in child maltreatment and resultant neuropsychological effects. Aggression and
Violent Behavior, 16(2), 87-97. doi:10.1016/j.avb.2010.12.007
804
Proceedings of the 45th annual
Boston University Conference
on Language Development
edited by Danielle Dionne
and Lee-Ann Vidal Covas
Cascadilla Press Somerville, MA 2021
Copyright information
Proceedings of the 45th annual Boston University Conference on Language Development
© 2021 Cascadilla Press. All rights reserved
Copyright notices are located at the bottom of the first page of each paper.
Reprints for course packs can be authorized by Cascadilla Press.
ISSN 1080-692X
ISBN 978-1-57473-067-8 (2 volume set, paperback)
Ordering information
To order a copy of the proceedings or to place a standing order, contact:
Cascadilla Press, P.O. Box 440355, Somerville, MA 02144, USA
phone: 1-617-776-2370, sales@cascadilla.com, www.cascadilla.com