On self-reported use of communication strategies by CLIL learners in primary education

Abstract

The use of communication strategies (CSs) in oral and written second language (L2) production has been widely investigated (e.g. Muñoz, 2007). As for content and language integrated learning (CLIL) settings, learners seem to resort to the first language (L1) less often than in traditional foreign language instruction (e.g. Celaya & Ruiz de Zarobe, 2010). However, few studies have examined what L2 learners say about their use of CSs by means of questionnaires – e.g. Ehrman & Oxford (1990), with adult English as a foreign language (EFL) learners – and little is known about the reported use of CSs by young learners (Purdie & Oliver, 1999), and much less by young CLIL learners. This study examines learners’ self-reported opinions about the use of CSs (guessing, miming, morphological creativity, dictionary, predicting, paraphrasing, borrowing, calque, foreignizing, avoidance and appeal for assistance). An adapted survey (Kellerman, Bongaerts, & Poulisse, 1987; Oxford, 1989; O’Malley & Chamot, 1990; Yule & Tarone, 1990) was administered to CLIL learners of English in grades 5 and 6 of primary education. Quantitative differences in terms of the type of strategies used were explored. Analyses showed striking similarities between grades 5 and 6 as well as significant differences in the use of the different CSs, paraphrasing and appeal for assistance being the most frequent strategies, whereas morphological creativity and miming obtained the lowest frequency. Findings are discussed in the light of learners’ age and the nature of CLIL instruction.

Keywords

communication strategies (CSs)content and language integrated learning (CLIL)children learners of English as a foreign language (EFL)questionnaires

I Introduction

Communication strategies (henceforth CSs) refer to all those devices language learners use to overcome linguistic difficulties encountered when trying to communicate in a foreign language (FL) with a reduced interlanguage system (Fernández Dobao, 2002). The first systematic analysis of CSs dates back to 1970s, when Váradi (1973) introduced some of the types of CS which subsequent researchers would use.

The use of CSs by second language (L2) learners (e.g. Muñoz, 2007) in oral and written production has been widely investigated. Recently, the use of the first language (L1) as a CS during oral production has been researched in content and language integrated learning (CLIL) contexts, where, compared to their mainstream counterparts, CLIL learners resort to their L1 less frequently (Martínez-Adrián & Gutiérrez-Mangado, 2015). However, we are in the need of studies that tackle the whole inventory of CSs. In addition, not many studies have examined what L2 learners say about their CSs in written questionnaires; e.g. Ehrman & Oxford (1990), with adult English as a foreign language (EFL) learners. To our knowledge, little is known about the use of CSs by young L2 learners, and much less by young CLIL learners, Purdie and Oliver’s (1999) being the only study that has addressed this issue with young schoolchildren. The present study uses an adapted version from Purdie and Oliver’s (1999) survey with young CLIL learners, whose repertoire of CSs makes up an underexplored area in CLIL research.

The present study investigates young CLIL (where English is the vehicular language) learners’ reported use of CSs as well as their preferences in the type of strategies employed during production of the target language. This article is structured as follows: the second section presents the theoretical framework and the main findings regarding CSs in mainstream EFL and CLIL settings in oral and written production and their study through questionnaires. This section finishes with the research questions posed in the study. A third section describes the setting and participants, the research instruments employed and the procedure we followed. Sections 4 and 5 present the results and discussion respectively, and the last section concludes the article pointing out limitations and offering lines for further research.

II Literature background

1 Taxonomies of CSs

The study of CSs in L2 acquisition has been approached from two different perspectives: the psycholinguistic and the interactional perspective. From a psycholinguistic point of view, CSs are treated as an individual mental response to a problem rather than as a joint response (Bialystok, 1990; Færch & Kasper, 1980, 1983, 1984; Poulisse, 1993, 1997; Poulisse, Bongaerts, & Kellerman, 1990). From an interactional perspective, a CS is a shared enterprise in which both the speaker and the hearer are involved rather than being only the responsibility of the speaker (Corder, 1978; Tarone, 1977, 1981; Váradi, 1973).

Dörnyei and Scott (1997) provided a review of nine taxonomies of CSs, but three of them have been the most prevalent in the area: Tarone’s taxonomy (1977), Færch and Kasper’s taxonomy (1983) and the Nijmegen group’s taxonomy (Poulisse, 1990). For space constraints, we will devote our attention to the classification developed by Tarone (interactional perspective) and the one by the Nijmegen group (psycholinguistic perspective), which both Purdie and Oliver’s (1999) survey and the questionnaire administered to participants in the present study are built on.¹ Tables 1 and 2 show both classifications.

Table 1.

Tarone’s typology of communication strategies (CSs).

Communication strategies
1. Avoidance
a. Topic avoidance b. Message abandonment
2. Paraphrase a. Approximation (pipe for waterpipe; from Tarone, 1977, p. 198) b. Word coinage (airball for balloon; from Váradi, 1973, as cited in Poulisse, 1990) c. Circumlocution (when describing a waterpipe, she is uh, Persina, and we use in Turkey a lot of; from Tarone, 1977, p. 198)
3. Conscious transfer a. Literal translation (translation of an equivalent Mandarin expression ‘He invites him to drink’; from Poulisse, 1990, p. 60) b. Language switch (The use of Turkish balon for English balloon; from Poulisse, 1990, p. 60)
4. Appeal for assistance (What is this? What called, from Tarone, 1977, p. 199)
5. Mime (clapping one’s hands to illustrate applause, from Tarone, 1977, p. 199)

Source. Based on Tarone, 1977.

Table 2.

The Nijmegen project’s typology of CSs.

Archistrategies	Communication strategies
Conceptual	1. Analytic (circumlocution, description, and paraphrase) (this you use for a baby so, uh, that it can’t uh, make, uh, his clothes erm 3 uh dirty (from Poulisse, 1990, p. 60)
Conceptual	2. Holistic (the use of a superordinate, co-ordinate, or subordinate term) vegetables for peas (from Poulisse, 1990, p. 61)
Linguistic	3. Transfer (borrowing, foreignizing, and literal translation) (Dutch etalage for English shop-window or cuffer from French coiffeur; from Poulisse, 1990, p. 62)
Linguistic	4. Morphological creativity (representator for representative; from Poulisse, 1990, p. 62)

Source. Based on Poulisse, 1990.

In order to tackle the whole inventory of CSs, the following categories from these well-known taxonomies were incorporated into our questionnaire: Avoidance and appeal for assistance from Tarone (1977), and linguistic strategies such as transfer, which is broken down into borrowing, calque and foreignizing, were also adopted from Poulisse’s (1990) taxonomy.

2 Research on CSs

Empirical studies on the use of CSs by L2 learners have been devoted to the following topics: classification of CSs, effectiveness of CSs, factors affecting choice of CSs and teachability of CSs. Several factors have been found to affect the use of CSs: proficiency level in the target language (i.e. Bialystok, 1983; Jourdain, 2000; Liskin-Gasparro, 1996; Paribakht, 1985; Poulisse et al., 1990; Tarone, 1977); influence of the native language (Tarone & Yule, 1987; Si-Qing, 1990); personality (Haastrup & Phillipson, 1983; Luján Ortega & Clark, 2000); learning and cognitive style (Luján Ortega & Clark, 2000; Littlemore, 2001); gender (Jiménez Catalán, 2003; Wang, 2008); task-related features such as cognitive demands, time constraints and interlocutor’s role (Poulisse et al., 1990; Khanji, 1993); and L1 strategic behaviour (Poulisse et al., 1990). Among these factors, proficiency has received the greatest attention. Research has tested the effect of this factor in terms of frequency and choice of CSs. Less proficient learners have been found to use more CSs than more proficient learners due to their limited command of the target language (Fernández Dobao, 2002; Hyde, 1982; Liskin-Gasparro, 1996; Paribakht, 1985; Poulisse et al., 1990). However, proficiency has a slightly limited effect on the choice of particular types of CSs. While several authors (Bialystok, 1983; Bialystok & Fröhlich, 1980; Jourdain, 2000; Wannaruk, 2003) have concluded that low-proficient learners tend to make a greater use of avoidance, mime and L1-based strategies, and more advanced learners tend to prefer L2-based strategies instead, this is not always the trend observed in research. Poulisse et al. (1990) found that the impact of proficiency was overruled by the effect of other factors such as the nature of the communicative task. This trend was also confirmed some years later by Fernández Dobao (2002). In highly demanding tasks, advanced learners used as many avoidance and transfer strategies as the low-proficient learners did. All in all, Ellis (2008) claims that little has yet been discovered about the developmental nature of CSs in L2 production. Thus, this article will try to contribute to fill this gap by investigating the use of CSs in two groups of young English learners in a CLIL setting.

3 Research on L1-based CSs in CLIL settings

The use of CSs during oral and written production has been widely investigated in EFL contexts (Cenoz, 2003; Gost, & Celaya, 2005; Muñoz, 2007; Navés, Miralpeix, & Celaya, 2005; Poulisse & Bongaerts, 1994; Viladot & Celaya, 2007). More limited research has been conducted in CLIL contexts in this respect, the majority of studies having been conducted with secondary-school learners. In addition, these studies have tackled L1-based strategies instead of examining the whole inventory of CSs, limitations that we will try to overcome in the present article. Given the scarcity of investigations conducted on CSs other than L1-based ones, the review of the studies that we provide below will illuminate the discussion of the results found in the present survey study. Some of these studies are longitudinal or pesudolongitudinal investigations of CLIL learners (Arratibel-Irazusta, 2015; Gutiérrez-Mangado, 2015; Lázaro Ibarrola & García Mayo, 2012; Serra, 2007) and other studies have compared CLIL to mainstream EFL learners (Agustín Llach, 2009; 2014; Celaya, 2008; Celaya & Ruiz de Zarobe, 2010; Gallardo-del-Puerto, 2015; García Mayo & Lázaro Ibarrola, 2015; Martínez-Adrián, 2015; Martínez-Adrián & Gutiérrez-Mangado, 2015).

In the case of the (pseudo)longitudinal studies conducted to date, Arratibel-Irazusta (2015) investigated the use of the L1 (interactional strategies, transfer lapses – i.e. borrowings and foreignizings – code-switching and discourse markers) in an oral narration task together with general proficiency in two different age CLIL groups of secondary-school learners. Older learners were found to outperform younger learners in general proficiency. However, both groups behaved in the same way in terms of L1 use except for the production of foreignizings. The comparison of the different categories examined revealed that discourse markers was the category that presented more L1 use in both groups. Finally, significant correlations were found between general proficiency and L1 influence. Lázaro Ibarrola and García Mayo (2012) have also examined the use of the L1 (discourse markers and repair sequences) in secondary school learners immersed in a CLIL context in the Basque Country. They concluded that L1 use significantly decreased along the 2-year period investigated.

Other longitudinal studies have been conducted with primary school children. Serra (2007) analysed L1 use in repair sequences by German-speaking primary-school children learning Italian or Romansch in a CLIL context. L1 use in repair sequences decreased as learners gradually attained a higher proficiency in the target language (TL).

More recently, Gutiérrez-Mangado (2015) examined the pseudolongitudinal development of two groups of primary-school CLIL and mainstream EFL learners in terms of L1 use in interactional strategies, transfer lapses and codeswitching. In the lapse of the two-year period explored, mainstream EFL learners were found to use the L1 in these three categories. In contrast, a decrease in appeals for assistance and an increase in borrowings were observed in the CLIL learners. Namely, they were found to abandon more cooperative strategies (appeals for assistance) in favor of more uncooperative ones.

With respect to the studies that have compared CLIL to mainstream EFL learners, Celaya and Ruiz de Zarobe (2010) examined the production of borrowings and foreignizings in written production by secondary school students in Grades 7 (age 13) and 10 (age 16). CLIL groups in both grades produced fewer borrowings than their mainstream EFL counterparts in a writing task. However, no clear pattern emerged for lexical creations.

Likewise, Martínez-Adrián and Gutiérrez-Mangado (2015) have compared secondary school CLIL to mainstream EFL learners as regards L1 use during an oral production task. This task was analysed both in terms of use of the L1 in interactional strategies and transfer lapses. A lower use of the L1 and a greater use of the target language during interaction were found in the case of CLIL learners.

Other studies that have compared CLIL and mainstream EFL learners have examined primary school children. Celaya (2008) explored lexical transfer (borrowings and lexical creations) in the written compositions by CLIL vs. mainstream EFL learners at grades 5 (age 11) and 7 (age 13). Results showed a lower percentage of borrowings in CLIL than in mainstream EFL but similar percentages of lexical creations in both settings. However, no inferential statistical analyses were carried out to rule out the effect of probability. In contrast, Agustín Llach (2009) observed fewer instances of L1 lexical influence not only in borrowings, but also in foreignizings and calques in the writings of Grade 6 (age 12) CLIL learners when compared to their mainstream EFL peers. Nevertheless, differences turned out to be significant only in the case of borrowings. These results were explained in terms of CLIL learners’ higher command of the FLs as well as their more frequent use of English as a means of communication.

Apart from the studies that have investigated L1 use in written production, other studies carried out with young learners have devoted their attention to oral production. García Mayo and Lázaro Ibarrola (2015) analysed the production of 40 age- and proficiency-matched dyads in terms of overall L1 use. CLIL learners displayed slightly higher linguistic abilities and were found to resort to the L1 less frequently than mainstream EFL learners, as they were more fluent and were used to speaking English with a meaningful purpose.

Pladevall-Ballester and Vraciu (2017) compared the longitudinal development of L1 use patterns in the oral production of 5th grade (age 11) and 6th grade (age 12) CLIL and mainstream EFL learners matched for exposure hours. More specifically, the study examined the learners’ use of the L1 in an individual narrative task in terms of content and function words, codeswitching, lexical transfer (i.e. borrowings, foreignizings) and interactional strategies. Results revealed no significant differences between the groups in the total number of L1 words. These findings seem to contradict other studies, where CLIL learners made a significant lower use of the L1 (Gallardo-del-Puerto, 2015; Gutiérrez-Mangado, 2015; Martínez-Adrián & Gutiérrez-Mangado, 2015; Martínez-Adrián, 2015). Results also pointed to a decreasing tendency in the use of borrowings and codeswitching at different testing times in both groups: findings that contradict previous studies where CLIL learners produced fewer borrowings than their mainstream EFL counterparts (e.g. Agustín Llach, 2009, 2014). The results in this study might be accounted for by the fact that, unlike the majority of the studies that examine the effect of CLIL instruction on the use of L1 patterns, the groups are matched for exposure hours. However, the increased use of foreignizings observed from the beginning of the study in the CLIL group (and in later stages in the mainstream EFL group) goes in line with the tendency observed in a considerable number of studies reporting a higher use of this strategy on account of this group’s higher proficiency (Agustin Llach, 2014; Celaya, 2008; Celaya & Ruiz de Zarobe, 2010).

In a similar vein, Gallardo-del-Puerto (2015) gathered data from primary school learners performing a peer interaction task. Two groups of CLIL learners in grades 4 and 6 were compared to two other groups of age-matched mainstream EFL learners as regards their production of codeswitching and transfer lapses in a dyadic story-telling task. Grade 4 CLIL learners produced a lower rate of codeswitching and transfer lapses than their mainstream EFL counterparts, these differences not being statistically significant though. In contrast, statistical significance was reached when Grade 6 participants were compared. Codeswitching, borrowings and foreignizings were more frequent in mainstream EFL than in CLIL learners, whereas CLIL learners produced a greater number of calques than mainstream EFL students. The study also concludes that the idea that foreignizing is characteristic of higher proficiency learners is not supported by the data obtained, which contradicts most CLIL research on this matter (Agustin Llach, 2014; Celaya, 2008; Celaya & Ruiz de Zarobe, 2010; Pladevall-Ballester & Vraciu, 2017).

Martínez-Adrián (2015) also analysed the oral production of 44 age- and proficiency-matched dyads (21 CLIL, 23 mainstream EFL) in terms of L1 use in appeals, clarification requests and metacomments. The results indicated that mainstream EFL learners produced more instances of L1 use in interactional strategies and that, as in Gallardo-del-Puerto (2015), greater differences emerged between both types of learners as grade increased. A qualitative inspection of the results also revealed that there were differences between groups in their preference for either the L1 or the target language in the case of appeals and clarification requests. However, metacomments were always produced in the L1 in both groups, which appeared to support previous classroom observation data (Gené Gil, Garau, & Salazar Noguera, 2012).

Even though quite recently research has been conducted on the use of CSs by young learners in CLIL settings, research in this area is still in its infancy and more studies aimed at the investigation of the whole inventory of CSs are needed.

4 The study of CSs through questionnaires

When one examines research on the use of CSs, the picture that emerges is that studies have mainly analysed oral or written production and that little is known about learners’ self-reported opinions concerning the use of CS. Some survey studies conducted with adults have investigated CSs together with other learning strategies (Ehrman & Oxford, 1990). However, as reported in Khan and Victori (2011), questionnaires used to study learning strategies and/or CSs are considered problematic by some researchers because they assume a stable aptitude or trait with respect to language learning in general (Tseng, Dörnyei, & Schmitt, 2006). They have not taken into account that learners may adjust their strategic approach depending on the situation or task, as some researchers have suggested (Macaro, 2006; Oxford, Cho, Leung, & Kim, 2004). In fact, a call for the administration of questionnaires that focus on oral communication in conjunction with specific tasks has been made. Thus, studies triangulating data have been clearly advocated in recent research (Gao, 2007). In order to fill this gap, Khan and Victori (2011) analysed the opinions gathered through three different questionnaires administered after the performance of three different linguistic tasks (picture story, art description and information-gap). These authors concluded that the instrument was not completely reliable in the sense that some of the items were prone to multiple interpretations and that these differed from the interpretation intended by the researcher. Overall, students reported the strategies they typically used rather than the specific strategies used on a task.

In the case of young learners, there is a lack of research investigating their use of CSs by means of questionnaires. The study by Purdie and Oliver (1999) is one exception. These authors examined the self-reported opinions on the use of learning strategies by 9- to 12-year-old children learning L2 English in an English as a second language (ESL) context. In comparison with the rest of strategy types analysed, the use of CSs were found to be the least frequently used by the learners, which contrasts with what has been found in young EFL learners during oral production. The study that we present below will try to add more evidence to the field by investigating the opinions gathered through a questionnaire administered to young English learners in a CLIL context. This will allow us to verify the trend usually observed in the case of young CLIL learners when performing oral production tasks.

5 Research questions

Based on the theoretical background and the findings from the research summarized in the aforementioned sections, the present study addresses the following research questions:

Research question 1: What is the amount of self-reported use of CSs in 5th and 6th grade CLIL learners?

Research question 2: What type of strategies do 5th and 6th grade CLIL learners prefer?

III Methodology

1 Setting and participants

The present study was carried out in a CLIL context in a middle-size town in the Autonomous Community of the Basque Country with 146 Basque–Spanish bilingual EFL learners (85 males, 61 females; age range 10–12) in 6 intact classrooms: 3 in their 5th (n = 75) and 3 in their 6th (n = 71) year of primary education, respectively. The learners had been exposed to Spanish and Basque since birth and/or early childhood in the school. They belonged to an instruction model in the Basque school system in which the school subjects are taught through Spanish and Basque (i.e. partial immersion); in this school English is gradually introduced as a vehicle for instruction (CLIL proper) from 3rd year of primary education, and taught by 3 co-ordinated content teachers. In 5th and 6th years learners are exposed to English in the classroom context for 5 to 7 weekly hours in 1-hour lessons scheduled as English as a school subject (2 to 3 hours in 5th year and 3 to 4 hours in 6th year – also by 3 co-ordinated language teachers), as well as in content lessons such as science, arts and crafts or physical education (2 to 3 hours in 5th year and 3 to 4 hours in 6th year). Lessons scheduled as English are gradually reduced throughout the academic year, while lessons scheduled as CLIL are gradually increased in these two grades, as follows: In 5th year English takes up 3 hours of instruction in the first term, and 2 hours in the 2nd and 3rd terms, while CLIL takes up 2 hours in the first term – 1 hour a week of science and one hour of physical education – and 3 hours in the subsequent terms – 2 hours a week of science and 1 hour of arts and crafts. In the case of 6th year, English takes up 4 hours in the first term and 3 hours in the following terms, while learners receive 3 hours a week of CLIL instruction in the first term in science – 2 hours – and physical education – 1 hour – and 4 hours in the following terms; 3 hours of science and one hour of arts and crafts.

All participants had been exposed to English since preschool in a classroom setting. At the moment of data gathering 5th year learners had received 714 hours of exposure (162 hours of CLIL lessons), and 6th graders 884 hours (291 hours of CLIL lessons); 59% of the learners received exposure extramurally, with an average of 2 hours of extra lessons per week in the past few years. Table 3 summarizes the participants’ mean previous FL contact in terms of onset age of exposure, length of exposure and FL proficiency level at the moment of testing. As far as their English proficiency, both groups were considered beginner learners, although the older group had reached a slightly better command of the language. According to a proficiency level test administered to both groups at the beginning of the study, 5th year students were in the A2-, and 6th year ones in the A2 level, according to the Common European Framework of Reference for Languages (CEFR; Council of Europe, 2001).

Table 3.

Participants.

	Onset age	Length of exposure to English	Hours of exposure to English	Hours of content and language integrated learning (CLIL) exposure	English proficiency level according to the Common European Framework of Reference (CEFR)
5th year	3.93	6-and-a-half academic years	714	162	A2–
6th year	4.28	7-and-a-half academic years	884	291	A2

2 Data collection, instruments and procedure

At the outset of the study, the learners were asked to fill in a biographical questionnaire on language(s) spoken for social use and at school, on the amount of time learning English (at school and/or extramurally), and a proficiency level test, consisting of the reading, listening and writing sections of the Cambridge English Flyers (Cambridge English, n.d.).³

After filling the background questionnaire and the proficiency level test, the participants underwent – with the help of the corresponding teacher during regular lessons – a self-report questionnaire adapted from Purdie and Oliver (1999) consisting in 40 statements (in Spanish) that surveyed on learning strategies in general, out of which 11 randomized items focused on CSs. The data presented here focuses on the latter. The instrument employed was a survey adapted from several taxonomies, given ‘[…] the diversity of strategy types suggested by descriptive taxonomies to be found in the literature’ (Kellerman, Bongaerts, & Poulisse, 1987, p. 100). The survey included conceptual, linguistic and interactional strategies.

From Purdie and Oliver (1999) – who in turn based their questionnaire on Oxford (1989) and O’Malley and Chamot (1990) – the following strategies were selected: guessing, miming, morphological creativity, dictionary, predicting and paraphrasing.² Strategies such as transfer, which falls into borrowing, literal translation and foreignizing – included in the taxonomy by Poulisse (1990) (see Table 2) – were also included in the survey. Finally, strategies such as avoidance and appeal for assistance included in the classification by Tarone (1977) (see Table 1) were adopted. A five-point Likert-type scale was used, in which the minimum score for each item was 1 (I strongly disagree) and the maximum 5 (I strongly agree). Table 4 summarizes the distribution of categories with their corresponding items.

Table 4.

Distribution of communication strategies (CSs).

Purdie and Oliver (1999)	Guessing	Si no entiendo algo en inglés, trato de adivinar lo que quiere decir.
	Miming	Si no sé cómo decir algo en inglés, uso las manos para mostrar lo que quiero decir.
	Morphological creativity	Si no sé cómo decir algo en inglés, me invento palabras nuevas.
	Dictionary	Si no entiendo lo que significa algo cuando leo en inglés, lo miro en el diccionario.
	Predicting	Cuando alguien me habla en inglés, trato de adivinar lo que va a decir justo a continuación.
	Paraphrasing	Si no sé cómo decir algo en inglés, uso otras palabras que significan lo mismo.
Poulisse (1990)	Borrowing	Cuando no sé decir algo en inglés, lo digo en euskera o castellano.
	Calque	Cuando no sé decir algo en inglés, traduzco palabra por palabra del euskera o castellano (por ejemplo, ‘my favorite plate’ en vez de ‘my favorite dish’.
	Foreignizing	Cuando no sé decir algo en inglés, adapto la palabra del euskera o castellano al inglés (por ejemplo, ‘go to the bosqu’ en vez de ‘go to the forest’.
Yule and Tarone (1990)	Avoidance	Cuando no sé decir algo en inglés, evito referirme a ello.
Yule and Tarone (1990)	Appeal for assistance	Cuando no sé decir algo en inglés, pido ayuda a otra persona (profesor, compañero, mamá, papá, …).

3 Data analysis procedure

Data were analysed in three different rounds: for the whole participant sample and for Grade 5 and Grade 6 separately. In each of these data sets, mean scores (1–5) and standard deviations (SD) were calculated both for each CS and for the whole set of strategies.

Kolmogrov–Smirnow tests were run to verify the normality of distribution of the samples, which were not found to be normal. Friedman tests were computed to investigate if there were any differences among the means. Subsequently, the mean scores of each of the individual strategies were compared to the mean score of the whole category of CSs (Wilcoxon tests) so as to verify which strategies significantly differ from the average use of CSs. This analysis would render three different degrees of strategy use –a ‘low use’ for those strategies whose means are significantly lower than that of the whole category, an ‘average use’ for those strategies whose means do not significantly differ from it, and a ‘high use’ for those strategies presenting mean scores significantly higher than that of the whole set of CSs. Regarding significance, an alpha level of .05 was used for all statistical tests. However, p-values below .01 and .001 were also indicated.

IV Results

In this section we will present the results as regards the whole sample (Table 5) and for each specific age group separately (Table 6 for 5th year and Table 7 for 6th year). These tables display strategy use organized in descending order, that is, individual strategies are ordered top-down from highest to lowest means. For the sake of clarity, highest means (those found to be significantly superior to the average overall use of CSs) are shaded in dark grey, and lowest means (those found to be significantly inferior to the average overall use of CSs) in light grey. Intermediate means (those not found to be significantly different from the average overall use of CSs) are not shaded.

Table 5.

All participants.

Strategies	Mean	SD	Z	Significance
Appeal for assistance	4.57	.62	−9.413	< .001
Paraphrasing	4.09	.96	−6.494	< .001
Dictionary	3.80	1.13	−4.175	< .001
Borrowing	3.75	1.33	−3.907	< .001
Guessing	3.66	1.32	−2.862	< .01
Calque	3.39	1.21	−.569	> .05
Predicting	3.16	1.28	−1.964	< .05
Avoidance	3.07	1.30	−3.531	< .001
Foreignizing	3.01	1.40	−3.464	< .001
Miming	2.74	1.36	−5.725	< .001
Morphological creativity	2.24	1.27	−8.730	< .001

Table 6.

Grade 5 participants.

Strategies	Mean	SD	Z	Significance
Appeal for assistance	4.59	.66	−6.381	< .001
Paraphrasing	4.21	.90	−4.580	< .001
Dictionary	3.85	1.06	−3.055	< .01
Borrowing	3.64	1.44	−1.543	> .05
Guessing	3.58	1.41	−1.356	> .05
Calque	3.50	1.26	−1.098	> .05
Foreignizing	3.23	1.43	−.923	> .05
Predicting	3.20	1.31	−1.431	> .05
Avoidance	3.00	1.34	−2.899	< .01
Miming	2.94	1.46	−2.994	< .01
Morphological creativity	2.35	1.35	−5.861	< .001

Table 7.

Grade 6 participants.

Strategies	Mean	SD	Z	Significance
Appeal for assistance	4.56	.58	−6.909	< .001
Paraphrasing	3.97	1.01	−4.606	< .001
Borrowing	3.87	1.20	−4.002	< .001
Dictionary	3.75	1.20	−2.766	< .01
Guessing	3.74	1.23	−2.728	< .01
Calque	3.28	1.17	−.374	> .05
Predicting	3.13	1.24	−1.271	> .05
Avoidance	3.13	1.27	−2.121	< .05
Foreignizing	2.79	1.36	−3.927	< .001
Miming	2.54	1.24	−5.107	< .001
Morphological creativity	2.13	1.20	−6.450	< .001

1 All participants

When we considered the whole sample, learners reported a moderate-to-high use of these strategies, with a mean of 3.39 (SD = .56). As for the results of the different individual strategies for all participants (Table 5), a non-parametric Friedman test of differences among repeated measures rendered a Chi-square value of 332.52 which was significant (p < .001), revealing that there were significant differences among the various strategies. As for the comparisons of the means for each individual strategy with that of the general category, Wilcoxon tests reached statistical significance in all comparisons made except when the strategy calque was involved in the comparison. Participants reported using the strategies guessing, borrowing, dictionary, paraphrasing and appeal for assistance to a larger extent than the rest. Paraphrasing and appeal for assistance were the ones which obtained mean scores above 4. On the contrary, predicting, avoidance, foreignizing, miming and morphological creativity were reported to be less frequently used. Miming and morphological creativity were the two least used CSs with a mean score below 3.

2 Grade 5 participants

Analyses computed with Grade 5 students also revealed a moderate to high use of CSs, with a mean score of 3.45 (SD = .56). With respect to the analysis of the different CSs (Table 6), the Chi-squared value rendered by the non-parametric Friedman test of differences among repeated measures was of 137.52 (p < .001), indicating that there were significant differences among the various individual strategies. As for the comparisons of the means for each individual strategy with that of the general category, Wilcoxon tests reached statistical significance in all comparisons made except when the strategies predicting, foreignizing, calque, guessing and borrowing were involved in the comparison. Three strategies were found to be used the most, namely dictionary, and above 4, paraphrasing and appeal for assistance. Unlikely, avoidance, and below 3, miming and morphological creativity yielded the lowest mean scores.

3 Grade 6 participants

Analyses computed with Grade 6 students indicated that the use of CSs as a whole remained moderate to high, with a mean score of 3.32 (SD = .55). As regards the distribution of the different CSs (Table 7), the Friedman test of differences among repeated measures rendered a Chi-square significant value of 202.90 (p < .001), which revealed that there were significant differences among several individual strategies. As for the comparisons of the means for each individual strategy with that of the general category, Wilcoxon tests reached statistical significance in all comparisons made except when the strategies predicting and calque were involved in the comparison. On the one hand, guessing, dictionary, borrowing, paraphrasing and above 4, appeal for assistance displayed statistically higher means than that of the category. On the other hand, avoidance and below 3 foreignizing, miming and morphological creativity attained the lowest frequency of use.

V Discussion

This study aimed to investigate young CLIL learners’ reported use of CSs as well as their preferences in the type of strategies employed during production of the target language. It was found that these learners exhibit a moderate-to-high use of these strategies, results that differ from those of the child sample of Purdie and Oliver (1999), in which they reported a lower use of CSs. This difference may be due to the fact that apart from the CSs included in their study, following the idea that the taxonomies found in the literature are far from homogeneous (Kellerman et al., 1987), we also included others such as appeal for assistance, a strategy reported to be used the most frequently in our study – findings in line with those reported in Victori and Tragant (2003) – and which has probably affected the difference in the use of CSs in general. In addition, our study was conducted in an EFL context, whereas in Purdie and Oliver, the participants were learners of L2 English in a naturalistic environment. A logical corollary is the higher use of this type of strategies in the EFL context, as previously found in other studies conducted in similar environments (i.e. Fernández Dobao, 2002; Poulisse et al., 1990).

Before examining the type of strategies preferred by these learners, it is worth clarifying that in our study a broad categorization scheme was adopted, even though it is our interest to report the main contrasts. The most preferred strategies were appeals for assistance and paraphrasing and the least, morphological creativity. In this respect, the use of paraphrasing and morphological creativity were also found to be the most and the least frequently used strategy in the study conducted by Poulisse et al. (1990). Research has shown that together with avoidance, L1-based strategies (borrowing, calque and foreignizing) are more frequent in low proficient learners. However, these CLIL learners exhibit a tendency to use more conceptual, i.e. L2-based strategies (not typically the focus of investigation, as compared to L1 influence, in CLIL studies), which are generally attributed to more advanced learners, such as paraphrasing (second most frequently used strategy). On the other hand, L1-based strategies such as foreignizing or other strategies widely employed by low-proficient learners such as avoidance, are used less frequently, with statistically lower means than the mean of the whole category, a finding which aligns with previous research that has examined oral production in CLIL settings (i.e. Gallardo-del-Puerto, 2015; Martínez-Adrián, 2015; Martínez-Adrián & Gutiérrez-Mangado, 2015). This may indicate that these CLIL learners risk more as they may feel more comfortable in using the target language and resort to strategies more characteristic of advanced learners, which seems to suggest that their educational context – a CLIL program – is positively affecting their rate of development in strategy use. The input characteristics the learners are exposed to in a content-based language teaching context seem to be a factor affecting their use of more advanced CSs. Teachers in these input-rich contexts paraphrase frequently, reformulating their own and the learners’ utterances, or scaffolding (Wood, Bruner, & Ross, 1976) the learners’ messages to make meaning come through and avoid communication breakdowns. On the basis of their answers, the learners in the study might be mimicking the teachers’ behavior. In addition, the fact that these learners immersed in a CLIL setting use the FL as a means of communication might lead them to be able to manipulate concepts in the target language to a larger extent. In sum, the effect of proficiency is suggested to be overruled by the effect of CLIL in this respect.

It should be noted though, that even if they are immersed in a CLIL context, learners report making an extensive use of appeal for assistance, which contrasts with what has been previously found in CLIL research with respect to this strategy (see Martínez-Adrián, 2015; Martínez-Adrián & Gutiérrez-Mangado, 2015). In this case, their younger age may be affecting their choice of this particular strategy to a larger extent. This finding would be in line with previous studies that investigated the use of the developmental nature of learning strategy use. In this respect, Victori and Tragant (2003) examined EFL students’ reported use of strategies longitudinally. In their open-ended questionnaire, they found that social strategies were reported to be used quite abundantly at the age of 10, but four years later, at age 14, they nearly disappeared from learners’ answers.

When comparing 5th and 6th grade learners, results show similar tendencies in the overall frequency use of CSs for both ages. However, more significant differences were found in a wider range of strategy types used within 6th graders. As for choice, 6th graders ranked higher than 5th graders in their reported use of borrowing and guessing. In this respect, previous studies examining the oral production of primary school learners have also attested a greater use of borrowings in 6th graders than in younger graders (Gutiérrez-Mangado, 2015). On the contrary, foreignizing ranked lower in grade 6 than in grade 5, a general trend reported in previous CLIL research where this type of strategy has not been found to be characteristic of more proficient learners (Arratibel-Irazusta, 2015; Gallardo-del-Puerto, 2015).

VI Conclusions

In conclusion, this article has revealed a moderate-to-high use of CSs in formal acquisition context learners even when immersed in a CLIL setting in which a greater focus on meaning and communication exists. However, even if CLIL is not associated with lower levels of CS use, as happens in naturalistic acquisition settings (Purdie & Oliver, 1999), it does seem to be linked to CS selection. CLIL is suggested to overrule the effect of proficiency as learners immersed in this type of meaning-oriented classrooms tend to favor the use of L2-based strategies, which are more typical of learners with a higher proficiency, rather than avoidance and L1-based strategies, which typically characterize low-proficient learners.

In terms of pedagogical implications, further research should explore the teachability of CSs (see Dörnyei, 1995) and how strategy training could affect their use. In this vein, research has examined what type of strategies good language learners use (Lennon, 1989; Naiman, Fröhlich, Stern, & Todesco, 1978; Reiss, 1983; Rubin, 1975). On the other hand, further investigation into FL learner strategy use seems timely. In this context, the present study has shown that, although CLIL learners report using some strategies which are more typical of more advanced learners, they state that they still resort to their L1, a finding which is in line with previous production data-based studies on L1 use in EFL (Cenoz, 2003; Gost & Celaya, 2005; Muñoz, 2007; Navés et al., 2005; Poulisse & Bongaerts, 1994; Viladot & Celaya, 2007) and CLIL (Arratibel-Irazusta, 2015; Gutiérrez-Mangado, 2015; Lázaro Ibarrola & García Mayo, 2012; Serra, 2007) settings. Research has alternatively shown that the use of the L1 can be a useful resource in bilingual educational contexts (see Gené Gil et al., 2012; Martínez-Adrián, 2015; Williams, 2002) so the typical penalization of L1 use in classrooms should be reconsidered when learning the language in CLIL programs. While we wait for further research to address this issue, FL teachers could benefit from taking cognisance of learner strategy use. Teachers should know, for instance, that social strategies are more typical of younger children whereas paraphrasing seems to be more characteristic of higher levels of proficiency. Future research should also shed more light on the effect of CLIL on CS use by comparing learners immersed in these educational contexts to mainstream EFL learners so that some of the potential effects of CLIL suggested in this article are proven experimentally. This would also enable a better comparison with findings from production data-based studies on L1 use comparing CLIL to mainstream EFL learners (Agustín Llach, 2009, 2014; Celaya, 2008; Celaya & Ruiz de Zarobe, 2010; Gallardo-del-Puerto, 2015; García Mayo & Lázaro Ibarrola, 2015; Martínez-Adrián, 2015; Martínez-Adrián & Gutiérrez-Mangado, 2015). Similarly, triangulation (as in Khan & Victori, 2011) of the self-reported opinions analysed in this study with oral data would be desirable in order to verify the reliability of learners’ self-reported use of CSs. Additionally, given the dynamic nature of the process of L2 acquisition, longitudinal studies (as in Victori & Tragant, 2003; Serra, 2007), analysing self-reported opinions over time would also be advisable.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors would like to acknowledge the grants awarded by the Spanish Ministry of Economy and Competitiveness (FFI2012-32212) and (FFI2016-74950-P), the University of the Basque Country (UPV/EHU) (UFI 11/06) and the Basque Government (IT904-16).

Notes

References

Agustín Llach

M.P

. (2009). The role of Spanish L1 in the vocabulary use of content and non-content EFL learners. In: Ruiz de Zarobe

Jiménez Catalán

R.M.

(Eds.), Content and language integrated learning: Evidence from research in Europe (pp. 112–129). Bristol: Multilingual Matters.

Agustín Llach

M.P

. (2014). L1 use in children EFL learners in traditional versus CLIL instruction. Paper presented at the International Conference on Child Foreign Language Acquisition, Vitoria-Gasteiz, Spain, 16–17 October.

Arratibel-Irazusta

. (2015). ‘They start to busq the jump’: A pseudolongitudinal study of crosslinguistic influence in L3 English CLIL learners. Unpublished master’s thesis, University of the Basque Country, Vitoria-Gasteiz, Spain.

Bialystok

(1983). Some factors in the selection and implementation of communication strategies. In: Færch

Kasper

(Eds.), Strategies in interlanguage communication (pp. 100–118). London: Longman.

Bialystok

(1990). Communication strategies: A psychological analysis of second language use. Oxford: Basil Blackwell.

Bialystok

Frohlich

(1980). Oral communication strategies for lexical difficulties. Interlanguage Studies Bulletin, 5, 3–30.

Cambridge English (n.d.). Cambridge English: Flyers (YLE Flyers). Cambridge: Cambridge English Language Assessment. Available at: http://www.cambridgeenglish.org/exams/young-learners-english/flyers (accessed July 2017).

Celaya

M.L.

(2008). ‘I study natus in English’: Lexical transfer in CLIL and regular learners. In: Manroy

Sánchez

(Eds.), 25 Años de lingüística aplicada en España: Hitos y retos (pp. 43–49). Murcia: Editum (Ediciones de la Universidad de Murcia).

Celaya

M.L.

Ruiz de Zarobe

(2010). First language and age in CLIL and non-CLIL contexts. International CLIL Research Journal, 1, 60–66.

10.

Cenoz

(2003). Cross-linguistic influence in third language acquisition: Implications for the organization of the multilingual mental lexicon. Bulletin VALS-ASLA (Vereinigung für angewandte Linguistik in der Schweiz), 78, 1–11.

11.

Corder

(1978). Strategies of communication. In: Færch

Kasper

(Eds.), Strategies in interlanguage communication (pp. 15–19). London: Longman.

12.

Council of Europe (2001). Common European framework of reference for languages: Learning, teaching, assessment. Cambridge: Cambridge University Press. Available at: http://www.coe.int/t/dg4/linguistic/Source/Framework_en.pdf (accessed July 2017).

13.

Dörnyei

(1995). On the teachability of communication strategies. TESOL Quarterly, 29, 55–81.

14.

Dörnyei

Scott

M.L.

(1997). Communication strategies in a second language: definitions and taxonomies. Language Learning, 47, 173–210.

15.

Ehrman

Oxford

(1990). Adult language learning styles and strategies in an intensive training setting. Modern Language Journal, 74, 311–326.

16.

Ellis

(2008). The study of second language acquisition. Oxford: Oxford University Press.

17.

Færch

Kasper

(1980). Processes and strategies in foreign language learning and communication. Interlanguage Bulletin Studies, 5, 47–118.

18.

Færch

Kasper

(1983). Strategies in interlanguage communication. London: Longman.

19.

Færch

Kasper

(1984). Two ways of defining communication strategies. Language Learning, 34, 45–63.

20.

Fernández Dobao

A.M

. (2002). The effect of language proficiency on communication strategy use: A case study of Galician learners of English. Miscelánea: A journal of English and American studies, 25, 53–75.

21.

Gallardo-del-Puerto

(2015). L1 influence in CLIL vs. EFL schoolchildren: A study of codeswitching and transfer lapses. Unpublished paper presented at the 33th AESLA International Conference, Universidad Politécnica de Madrid, Spain, 16–18 April.

22.

Gao

(2002). Has language learning strategy research come to an end? A response to Tseng et al. 2006. Applied Linguistics, 28, 615–620.

23.

García Mayo

M.P.

Lázaro Ibarrola

. (2015). Do children negotiate for meaning in task-based interaction? Evidence from CLIL and EFL settings. System, 54, 40–54.

24.

Gené Gil

Garau

M. J.

Salazar Noguera

. (2012). A case study exploring oral language choice between the target language and the L1s in mainstream CLIL and EFL secondary education. Revista de Lingüística y Lenguas Aplicadas 7, 133–146.

25.

Gost

Celaya

M. L.

(2005). Age and the use of the L1 in EFL oral production. In: Carrió Pastor

M.L.

(Ed.), Perspectivas Interdisciplinares de la Lingüística Aplicada (pp. 129–136). Valencia: Universitat Politècnica de València.

26.

Gutiérrez-Mangado

(2015). The development of L1 use by school children in CLIL and non-CLIL contexts. Unpublished paper presented at the 33rd AESLA International Conference, Madrid, Spain, 16–18 April.

27.

Haastrup

Phillipson

(1983). Achievement strategies in learner/native speaker interaction. In: Færch

Kasper

(Eds.), Strategies in interlanguage communication (pp. 140–158). London: Longman.

28.

Hyde

(1982). The identification of communication strategies in the interlanguage of Spanish speakers of English. Anglo-American Studies, 2, 13–30.

29.

Jiménez Catalán

R.M

. (2003). Sex differences in L2 vocabulary learning Strategies. International Journal of Applied Linguistics, 13, 54–77.

30.

Jourdain

(2000). A native-like ability to circumlocute. The Modern Language Journal, 84, 185–195.

31.

Kellerman

Bongaerts

Poulisse

(1987). Strategy and system in L2 referential communication. In: Ellis

(Ed.), Second language acquisition in context (pp. 100–112). London: Prentice-Hall.

32.

Khan

Victori

(2011). Perceived vs. actual strategy use across three oral communication tasks. International Review of Applied Linguistics in Language Teaching, 49, 27–53.

33.

Khanji

(1993). Interlanguage talk: The Relation between task types and CS among EFL Arab learners. In: Alatis

J.E.

(Ed.), Georgetown University Round Table on Language and Linguistics 1993 (pp. 428–436). Washington D.C.: Georgetown University Press.

34.

Lázaro Ibarrola

García Mayo

M.P

. (2012). L1 use and morphosyntactic development in the oral production of EFL learners in a CLIL context. International Review of Applied Linguistics, 50, 135–160.

35.

Lennon

(1989). Introspection and intentionality in advanced second-language acquisition. Language Learning, 39, 375–396.

36.

Liskin-Gasparro

J.E.

(1996). Circumlocution, communication strategies, and the ACTFL proficiency guidelines: An analysis of student discourse. Foreign Language Annals, 29, 317–330.

37.

Littlemore

(2001). An empirical study of the relationship between cognitive style and the use of communication strategy. Applied Linguistics, 22, 241–265.

38.

Luján-Ortega

Clark-Carter

(2000). Individual differences, strategic performance and achievement in second language learners of Spanish. Studia Linguistica, 54, 280–287.

39.

Macaro

(2006). Strategies for language learning and language use: Empirical implications for questionnaire studies. JACET Bulletin, 31, 21–32.

40.

Martínez-Adrián

(2015). L1 use in interactional strategies by primary school learners in two learning contexts: CLIL vs. non-CLIL. Unpublished paper presented at the 33rd AESLA International Conference, Madrid, Spain, 16–18 April.

41.

Martínez-Adrián

Gutiérrez-Mangado

M.J.

(2015). L1 use, lexical richness, accuracy and complexity in CLIL and non-CLIL learners. Atlantis, Journal of the Spanish Association for Anglo-American Studies, 37, 175–197.

42.

Muñoz

(2007). Cross-linguistic influence and language switches in L4 oral production. Vigo International Journal of Applied Linguistics, 4, 73–94.

43.

Naiman

Fröhlich

Stern

Todesco

(1978). The good language learner. Toronto: Ontario Institute for Studies in Education.

44.

Navés

Miralpeix

Celaya

M.L.

(2005). Who transfers more … and what? Cross-linguistic influence in relation to school grade and language dominance in EFL. International Journal of Multilingualism, 2, 113–134.

45.

O’Malley

J.M.

Chamot

A.U.

(1990). Learning strategies in second language acquisition. Cambridge: Cambridge University Press.

46.

Oxford

R.L.

(1989). Use of language learning strategies: a synthesis of studies with implications for strategy training. System, 17, 235–247.

47.

Oxford

Cho

Leung

Kim

H.J.

(2004). Effect of the presence and difficulty of task on strategy use: An exploratory study. International Review of Applied Linguistics, 42, 1–47.

48.

Paribakht

(1985). Strategic competence and language proficiency. Applied Linguistics, 6, 132–146.

49.

Pladevall-Ballester

Vraciu

(2017). Exploring early EFL: L1 use in oral narratives by CLIL and non-CLIL primary school learners. In: García Mayo

M.P.

(Ed.), Learning foreign languages in primary school: Research insights. (pp. 124–148). Clevedon: Multilingual Matters.

50.

Poulisse

(1990). The use of compensatory strategies by Dutch learners of English. Dordrecht: Foris.

51.

Poulisse

(1993). A theoretical account of lexical communication strategies. In: Schreuder

Weltens

(Eds.), The bilingual lexicon, (pp. 157–189). Amsterdam: John Benjamins.

52.

Poulisse

(1997). Some words in defense of the psycholinguistic approach. Modern Language Journal, 81, 324–328.

53.

Poulisse

Bongaerts

(1994). First language use in second language production. Applied Linguistics, 15, 36–57.

54.

Poulisse

Bongaerts

Kellerman

(1990). The use of compensatory strategies by Dutch learners of English. Enschede: Sneldruk.

55.

Purdie

Oliver

(1999). Language learning strategies used by bilingual school-aged children. System, 27, 375–38.

56.

Reiss

M.A.

(1983). Helping the unsuccessful language learner. Forum, 21, 2–24.

57.

Rubin

(1975). What the ‘good language learner’ can teach us. TESOL Quarterly, 9, 41–51.

58.

Serra

(2007). Assessing CLIL at primary school: A longitudinal study. The International Journal of Bilingual Education and bilingualism, 10, 582–602.

59.

Si-Qing

(1990). a study of communication strategies in interlanguage production by Chinese EFL learners. Language Learning, 40, 155–87.

60.

Tarone

(1977). Conscious communication strategies in interlanguage: A progress report. In: Brown

H.D.

Yorio

C.A.

Crymes

R.C.

(Eds.), On TESOL 77: Teaching and Learning English as a second language: Trends in research and practice (pp. 194–203). Washington, DC: TESOL.

61.

Tarone

(1981). Some thoughts on the notion of communication strategy. TESOL Quarterly, 15, 285–295.

62.

Tarone

Yule

(1987). Communication strategies in East–West interactions. In: Smith

L.E.

(Ed.), Discourse across cultures: Strategies in world Englishes (pp. 49–65). New York: Prentice Hall.

63.

Tseng

Dörney

Schmidt

(2006). A new approach to assessing strategic learning: The case of self-regulation in vocabulary acquisition. Applied Linguistics, 27, 78–102.

64.

Váradi

(1973). Strategies of target language learner communication: Message adjustment. International Review of Applied Linguistics, 18, 59–72.

65.

Victori

Tragant

(2003). Learner strategies: A cross-sectional and longitudinal study of primary and high-school EFL learners. In: García Mayo

M.P.

García Lecumberri

M.L.

(Eds.), Age and the acquisition of English as a foreign language (pp. 182–209). Clevedon: Multilingual Matters.

66.

Viladot

Celaya

(2007). ‘How do you say preparar?’: L1 use in EFL oral production and task-related differences. In: Losada Friend

et al (Eds.), Proceedings of the 30th international AEDEAN conference. Huelva: U de Huelva.

67.

Wang

L.M.

(2008). EFL 学习者习得交际策略的性别差异研究 [A study of gender differences in communication strategies by EFL learners]. Foreign Languages and Their Teaching, 8, 37–41.

68.

Wannaruk

(2003). Communication strategies employed by EST students. Studies in Language and Language Teaching, 12, 1–18.

69.

Williams

(2002). Extending bilingualism in the education system: Education and lifelong learning committee ELL-06-02. Available at: http://www.assembly.wales/Committee%20Documents/ELL%2006-02(p.4)%20Dr%20Cen%20Williams%20paper-20032002-28970/3c91c7af00023d820000595000000000-English.pdf (accessed July 2017).

70.

Wood

Bruner

Ross

(1976). The role of tutoring in problem solving. Journal of Child Psychology and Psychiatry, 17, 89–100.

71.

Yule

Tarone

(1990). Eliciting the performance of strategic competence. In: Scarcella

Andersen

Krashen

(Eds.), Developing communicative competence in a second language (pp. 179–184). Boston: Heinle & Heinle.