Abstract
The present study seeks to explore the effects of participating in a content and language integrated learning (CLIL) programme in terms of second language (L2) content-specific vocabulary learning in a primary education school. CLIL instruction has been found to benefit the development of English L2 vocabulary, more receptive than productive. Nonetheless, little is known on content-specific vocabulary learning in CLIL and how it is supported through classroom input with young learners. We explore the impact of a 4-week CLIL programme implemented in a Catalan primary education school with a group of 16 Catalan/Spanish first language (L1) children, aged 10–11 years. Vocabulary learning was measured by means of ad hoc tests, designed on the basis of the input from the CLIL materials and targeting vocabulary depth and breadth. The results indicate that CLIL instruction positively impacted children’s depth and breadth of content-specific vocabulary knowledge. More specifically, the children experienced a significant improvement both in their receptive and productive knowledge of word meaning and form with regard to content-specific lexis. Moreover, with regard to children’s productive vocabulary knowledge as elicited through a written content task, statistical improvement was found in terms of lexical diversity and sophistication, but not lexical density. We argue for the need to use input profiled assessment instruments to appraise L2 vocabulary gains in CLIL.
Keywords
I Introduction
Vocabulary learning is considered to be key in the acquisition of a second or foreign language (L2), and it is also closely related to enhanced language proficiency: the greater the lexical repertoire, the more likely one is to communicate effectively and become a more proficient user of the language in focus (Meara, 1996; Nation, 1990; Read, 1998). Vocabulary knowledge has been shown to develop with increasing exposure to the target language and proficiency level (Goldberg et al., 2008; Schmitt, 1998). Such development is considered to follow a systematic order related to frequency of input, that is, at initial stages learners get familiar with the words they are exposed to more frequently and, as they become more proficient, less frequent words are incorporated into their word stock (Milton, 2009; Vermeer, 2001).
In instructed contexts, a great amount of effort has been put into identifying more efficient ways of teaching and learning L2 vocabulary. Content and language integrated learning (CLIL) classroom contexts are considered to provide numerous opportunities in this regard since in this type of instruction the L2 is used for genuine communication while covering subject-specific topics related to a range of school disciplines. The dual-focus on content and language (Coyle et al., 2010) further stretches CLIL students’ L2 skills since they must understand new subject-related concepts and apply them by performing communicative tasks with varying cognitive and linguistic demands (Tragant et al., 2015). As a result, CLIL instruction is considered to reunite the psycholinguistic conditions for more authentic and meaningful L2 learning than in the mainstream English as a foreign language (EFL) class (Muñoz, 2007).
The L2 learning benefits of CLIL have been extensively documented, particularly in the domain of vocabulary acquisition, in different contexts and at different educational stages (for an overview, see Dalton-Puffer, 2011; Pérez Cañado, 2012). Yet, the bulk of the empirical evidence available is associated with the impact of CLIL as an educational approach, as opposed to EFL instruction, with little insight into actual pedagogical practices and materials. Nonetheless, such insight is crucial for understanding the learning that takes place in CLIL classrooms. In our study, we aim to explore the L2 learning outcomes of a Generació Plurilingüe (GEP; ‘plurilingual generation’) CLIL programme implemented in a state primary school in Catalonia (Spain) during the 2021–22 academic year in relation to the input that the children were exposed to during class time through teacher talk and class materials.
GEP programmes were first launched by the Catalan Department of Education in 2017–18 with the objective of improving students’ L2 competence in order to prepare them to become fully-fledged citizens able to interact critically with the plurilingual and multicultural world around them. From its beginnings since the school year 2020–21, 611 educational centers in Catalonia have participated in this innovative programme. GEP provides a 3-year training scheme addressed to teachers and school management teams who want to design, implement and assess CLIL programmes using information and communication technology (ICT) at all pre-university educational stages, funded by the Catalan Department of Education. The ultimate aim is to increase the contact hours with the curricular foreign languages and to develop and consolidate the schools’ plurilingual linguistic project.
Moreover, while numerous studies have looked into the benefits of CLIL for general L2 vocabulary development, little evidence is available on the learning outcomes in terms of content-specific vocabulary, more tied to the learning objectives of actual CLIL implementations. Therefore, this study is concerned with examining the effects of participating in a CLIL intervention within the GEP programme on primary school children’s knowledge of content-specific vocabulary in EFL with the ultimate objective of assessing the effectiveness of implementing this particular innovative programme in Catalan schools. Gauging the impact of localized interventions is necessary if we want to bridge the gap between research and educational policies.
II Review of the literature
1 Vocabulary learning in CLIL contexts
Most of the early research on L2 vocabulary learning in CLIL was carried out in secondary education schools, and overall findings evidenced CLIL students’ gains in general vocabulary knowledge, in particular receptive vocabulary, as compared to that of their non-CLIL EFL peers (Jiménez Catalán & Ruiz de Zarobe, 2009; Jiménez Catalán et al., 2006). More recent studies also support the benefits that participating in CLIL programmes might bring about for teenagers’ lexical knowledge in the target language, which appears to grow as experience with the CLIL approach, age and proficiency increases (Agustín-Llach & Canga Alonso, 2014; Castellano-Risco et al., 2020). Castellano-Risco et al. (2020) measured the impact of exposure on the development of receptive vocabulary knowledge of 138 secondary education CLIL learners (with different starting ages) and non-CLIL EFL learners. While statistically significant differences could be identified in the recognition of the 2,000 (2K) most frequent general, non-academic vocabulary in favor of CLIL learners, no significant differences were found within the different CLIL groups who had embarked on the programme at different ages all along their primary and secondary schooling. All in all, these findings help shed some further light onto the fact that it is not just a matter of increasing contact hours through CLIL alone, but of offering high quality CLIL experiences.
Having a more authentic and richer exposure to the foreign language through CLIL also seems to stimulate productive vocabulary learning, and even more so among more mature and proficient learners, as is the case with secondary education students. Positive outcomes have been reported by some comparative studies (Jexenflicker & Dalton Puffer, 2010; Ruiz de Zarobe, 2010) in favor of students participating in the CLIL strand, particularly in terms of their lexical variety and use of low frequency words. Canga Alonso and Arribas Garcia (2015) investigated the productive vocabulary knowledge of 10th grade students (15–16 years of age) participating in a CLIL programme and those students following a non-CLIL programme. The results revealed a larger productive vocabulary size in the case of CLIL learners. But, still, the productive vocabulary size of both groups was below the first 1,000 (1K) most frequent words in English according to the Productive Vocabulary Levels Test (PVLT) (Laufer & Nation, 1995, 1999). These findings also highlight the fact that, despite the advantage shown by CLIL learners, both CLIL and non-CLIL students might still face some difficulties to produce basic spoken and written texts in English. Additionally, it may take time for the positive impact of CLIL to show up, as evidenced by Reynaert (2019), since significant changes in general productive vocabulary knowledge in her study were only observable after two years of experiencing CLIL education. Such results are in line with research studies examining classroom vocabulary use in detail since CLIL students also appear to be more able to notice their own lexical gaps and, as a result, initiate some self-repair more frequently than their non-CLIL EFL counterparts (Dalton-Puffer, 2007).
If we now turn to the context of primary education, there are also a range of studies which support the advantage of learners participating in CLIL programmes. Xanthou (2011) examined the impact of CLIL on L2 vocabulary knowledge with two groups of 6th graders (10–11 years old). The first group was taught three 80-minute Science lessons through the medium of L2 English, while the second group was taught the same content through the medium of first language (L1) Greek. The outcomes demonstrated a significant effect of CLIL on receptive L2 vocabulary knowledge, evidencing how comprehension and retention were enhanced by attaching L2 words to the students’ immediate reality. Furthermore, Canga Alonso (2015) compared the receptive general vocabulary knowledge of 79 CLIL primary learners with 331 non-CLIL learners at the end of primary and secondary school. Results revealed that learners’ receptive vocabulary sizes lay within the most frequent 1K words. In addition, non-CLIL secondary school students obtained better results than primary students but the differences between the non-CLIL EFL secondary education group and the CLIL primary group were not statistically significant. Thus, these findings support the fact that the CLIL approach offers a benefit for general vocabulary acquisition since CLIL learners had been exposed to the L2 for a shorter period of time.
With regard to productive L2 vocabulary knowledge of primary education students receiving CLIL instruction, studies are scarce if compared to those addressing receptive vocabulary knowledge. CLIL learners seem to show a higher performance in lexical production, higher lexical variation and lower word repetition, mainly examined through a letter-writing task or a cloze (Agustín Llach, 2014; Agustín Llach & Jiménez Catalán, 2007; Celaya & Ruiz de Zarobe, 2010). Nevertheless, there also exists counter evidence supporting the advantage of non-CLIL groups when it comes to lexical availability (Jiménez Catalán & Ojeda Alba, 2009; Jiménez Catalán et al., 2006). All in all, the research findings obtained so far appear to be rather inconclusive as for the quantity and quality of lexical output of CLIL and non-CLIL groups. More studies are needed to ascertain whether CLIL groups benefit from such content-based approach, particularly when it comes to the quality of their lexical production.
2 Academic and content-specific vocabulary learning in CLIL contexts
Very few studies have looked into the impact of CLIL on academic and content-specific vocabulary, even though the input received in CLIL settings targets the development of learners’ CALP (Cognitive Academic Language Proficiency), as opposed to the focus on BICS (Basic Interpersonal Communicative Skills) (Cummins, 1979), traditionally found in the EFL class. Research findings so far indicate that the gap between CLIL and non-CLIL learners in the receptive academic band is even greater than the one in general vocabulary knowledge. In a study already mentioned, Castellano-Risco et al. (2020) found statistically significant differences between the CLIL and the non-CLIL learners with regard to their academic vocabulary size as measured by the Vocabulary Levels Test (Schmitt et al., 2001). The type of exposure received in CLIL classes was accountable for this advantage.
Another strand of research has explored the impact of CLIL on learners’ content-specific vocabulary learning, namely lexis that is needed to access the disciplinary content and the learning activities in the CLIL class. Mastery of this lexis is closely related to specific teaching and learning objectives within contextualized CLIL interventions/programmes. To our knowledge, Heras and Lasagabaster (2015) and Tragant et al. (2015) are the only studies approaching L2 vocabulary outcomes in CLIL from this perspective. Tragant et al.’s (2015) study was conducted with a group of primary education students, aged 8 years, that were exposed to EFL instruction in the fall term and to CLIL instruction (Science) in the winter term. Tailor-made productive vocabulary tests based on the previous analysis of CLIL and EFL textbook materials were created. The results obtained showed that students were exposed to a greater number of words and to more abstract and specific vocabulary in the CLIL materials, but that they made significant progress in vocabulary learning in both contexts. In the case of secondary education students, Heras and Lasagabaster (2015) also proved the overall positive benefits of CLIL instruction on vocabulary learning related to the discipline of Physical Education via a written vocabulary test specifically created for the study.
Finally, standardized tests, widely used in almost all the studies reporting on vocabulary learning in CLIL, cannot tap into content-specific vocabulary learning, which requires vocabulary tests more aligned with the actual content and language learning aims pursued in the CLIL class. Moreover, in all the reviewed studies attention has been paid to learners’ L2 vocabulary size or breadth, but not to the depth of their vocabulary knowledge, that is how well one knows a word, including spelling, meaning, collocation, frequency, morphological and syntactic properties, among other aspects (Laufer, 1996; Liu & Nation, 1985). An exploration of both breadth and depth of content-specific vocabulary knowledge can help provide a more realistic picture of what participating in a CLIL programme entails in terms of L2 learning opportunities and outcomes in the school context.
3 The study
The main objective of this study is to explore the effects of participating in a GEP CLIL intervention on primary school children’s knowledge of content-specific vocabulary in EFL. More specifically, we aimed to answer the following research questions:
Research question 1: Is there a significant difference between children’s depth of L2 content-specific word knowledge before and after the intervention?
Research question 2: Is there a significant difference between children’s breadth of L2 productive vocabulary in a content task before and after the intervention?
Research question 3: Are there any correlations among the different dimensions of children’s (content-specific) L2 vocabulary knowledge before and after the intervention?
III Method
1 Participants
Our study was carried out at a Catalan public school located in a medium-sized town near Barcelona. The participants were 16 bilingual Spanish/Catalan primary school children (6 girls and 10 boys) aged 10–11 years and enrolled in the same 5th grade class. The class size had been reduced from the regular 25 pupils to 16 in the context of the Covid-19 pandemic, so the participants represent an intact class. The children were instructed in minimal exposure conditions, receiving 3 hours of instrumental English classes per week, with no prior CLIL experience. At the onset of the study, they had accumulated approximately 500 hours of in-class exposure to the target language. Six out of the 16 children (38%) were attending extracurricular EFL afternoon sessions, although the majority of them had only started at the same year or the year before, and the number of hours per week ranged from 1 to 2, thus keeping exposure low. The sample was homogeneous as regards their socio-economic and cultural background, according to the biodata questionnaire that was administered to all the participants.
2 Design of the study
A pre-test/post-test design was used to measure the impact of a CLIL pedagogical intervention on the learners’ depth of content-specific word knowledge and breadth of productive vocabulary knowledge in a content task. Written consent was obtained from the children’s families and the teachers involved. The pedagogical intervention took place during the 2020–21 academic year, from April to May, and consisted of 6 sessions of 55 minutes each over a 3-week span (2 sessions/week) on the topic of ‘healthy nutrition’. The children had not had any other contact with CLIL and, during the intervention, they received no other EFL instruction. The intervention was part of the school’s GEP innovation programme which had started the previous academic year (2019–20). In our study, the CLIL pedagogical intervention was co-taught by two teachers (female, aged 42 and 28 years, respectively). They had planned and designed the sessions together under the supervision of experienced CLIL teacher trainers. The teachers had previous experience co-teaching as they had implemented a shorter CLIL module in the first year of their GEP training. One of the teachers (Teacher 1) was a primary school English specialist, with a certified B2 level of English and 18 years of experience teaching EFL in pre-school and primary education. The other teacher (Teacher 2) was a preschool teacher, with a certified B2 level of English and 1 year of experience teaching EFL. In the Catalan education system, a homeroom preschool or primary teacher can teach English provided they can certify a B2 level of proficiency.
The intervention consisted of a 3-week unit on the topic of ‘Healthy Nutrition’ (6 sessions of 55 minutes each). The final aim of the project was to create a balanced and healthy menu for the school. To do this, the children had to investigate what a balanced menu is and which food groups it must include in order to be considered healthy. They also had to analyse their own school menu and make suggestions on how it could be improved. The intervention had a dual focus on content and language learning and followed the CLIL principles of planning and material design including the 4Cs (Content, Cognition, Communication and Culture) framework, the Language Triptych (Coyle, 2006, 2007; Coyle et al., 2010), and other conceptual tools such as the taxonomy of thinking skills (Krathwohl, 2002)
3 Instruments
The vocabulary learning outcomes of the CLIL intervention were measured by means of an ad hoc test designed on the basis of the class materials that were used for instruction (i.e. student worksheets, teacher’s notes, class PowerPoints, reading texts, classroom videos). Following Heras and Lasagabaster (2015), the ad hoc design was considered necessary to gauge the lexical gains specifically attributable to the CLIL intervention under study. The teachers shared the materials with the researchers prior to the intervention. The materials were analysed with Lextutor (Cobb, 2014) to identify those lexical items (i.e. content words) that appeared three or more times and which were related to the topic of nutrition. The frequency of the items in the materials was assumed to make them more salient to the children and presumably more likely to be retained (Vermeer, 2001). The resulting pool of items (i.e. 44 items) was screened for cognates, reducing it to a shortlist of 25 items. The shortlisted items were further categorized in terms of their frequency of use in the general language by means of the BNC Sampler corpus (both written and spoken). This was considered necessary because some of the lexical items frequently used in the materials had low frequency of use in the general language and were, hence, more likely to have been encountered by the children only during the intervention (e.g. ‘sparingly’). The final selection consisted of 20 items, 1 frequently used in the class materials and with both low and high frequency of use in the general language, and representing different lexical categories (nouns, verbs, adjectives and adverbs). The test was inspected by the teachers who considered that no formal and content modifications were necessary.
The test consisted of two parts which tapped into different types of vocabulary knowledge. The first part focused on children’s depth of content-specific vocabulary knowledge, targeting two specific dimensions: receptive and productive knowledge of meaning and form (Webb, 2005). More specifically, the children had to provide the translation into Catalan or Spanish of 20 lexical items that appeared in the class materials (i.e. receptive meaning and form knowledge) as well as use them in a sentence (i.e. productive meaning and form knowledge). The second part of the test focused on children’s breadth of productive vocabulary knowledge in a content task. According to Schmitt (2019), productive vocabulary knowledge involves knowing a lexical item well enough to use it in a communicative task and should, hence, be measured with skill-based instruments. Consequently, the children were asked to write a potato omelet recipe, including the ingredients, the kitchen utensils needed, and the different steps. The recipe was a genre targeted in the CLIL intervention.
4 Data collection and analysis
The test was administered a week before the onset of the CLIL intervention (T1) and again one week after the end of the intervention (i.e. 4 weeks apart) (T2). The test was administered in class and lasted 45 minutes on both occasions. At the beginning of the test, clear instructions together with an example were given both orally and in written form in the students’ mother tongue to clarify what they had to do. To mitigate the test–retest effect, the order of the lexical items in the first part of the test was changed in the post-test.
The two parts of the test were corrected in different ways. In the first part (i.e. depth of content-specific vocabulary knowledge), words correctly translated received 1 point, whereas incorrect or no translations received 0 points. The children were awarded 1 point for sentence-use of the words that was both accurate and adequate and 0.5 points for use that was adequate but not accurate (e.g. problems with morphosyntactic features of the word). Pallotti (2009) described adequacy as ‘the degree to which learners’ performance is more or less successful in achieving the task’s goals efficiently’ (p. 596). The second part of the test (i.e. breadth of productive vocabulary knowledge) was rated for lexical diversity, lexical density and lexical sophistication, once the students’ writing were transcribed. The measure of lexical diversity used was the Corrected Type Token Ratio (CTTR), a measure that is approximately independent of sample size. It represents the number of different words (i.e. types) divided by the square root of twice the number of words (i.e. tokens) in the sample (Carroll, 1964). To calculate the CTTR, there was no lemmatization (i.e. inflected forms of one word were considered different types) and we excluded numbers, cognates (‘minutes’), false starts, L1 words, and illegible words. Lexical sophistication was measured by means of Advanced Guiraud’s Index for 1K and 2K words. Finally, lexical density was calculated as the ratio of lexical words divided by all tokens. Calculations were performed with VocabProfiler Classic v. 4 (Cobb, 2014).
The six sessions were video-recorded to gain additional insight into the input the children received in class and how the materials were used in practice. While a detailed description of the class sessions is beyond the scope of this study, the observations confirm that the teachers’ performance and the in-class activities favored the learning of the targeted vocabulary.
Inferential statistics were performed with the SPSS 19 statistical package. A Kolmogorov–Smirnov test was carried out to check if the data met the normality assumption. As some of the variables in the study had normal distribution, but not all of them, paired t-tests and Pearson Product Moment correlations 2 were used applying bootstrapping to compensate for the partial violation of the normality assumption. Finally, Cohen’s d values were calculated to obtain effect sizes of the differences between pre- and post-intervention scores. Following Plonsky and Oswald (2014), in SLA (Second Language Acquisition) values of 1.00 and above indicate a large effect size, values between 0.5 and 0.99 a medium effect size and values below 0.5, a small effect size.
IV Results
Tables 1 and 2 present the descriptive and inferential statistics for the two dimensions of depth of content-specific vocabulary knowledge explored in this study, namely receptive meaning and form knowledge (Receptive_MFK) and productive meaning and form knowledge (Productive_MFK), at T1 and T2. With regard to children’s receptive MFK, means were greater at T2 than at T1 (4.75 vs. 7.75), and the difference was statistically significant (t(15) = −4.24, p = .01, 95% CI [−4.44, −1.81]) with a medium effect size (d = 0.93). Productive MFK was also greater at T2 than at T1 (3.06 vs. 6.19), with children using the words in sentences more accurately and adequately after the intervention, with a statistically significant difference (t(15) = −3.15, p = .02. 95% CI [−5.22, −1.41]) and a medium effect size (d = 0.86). It should also be pointed out that at T2 the minimum values of both variables were above 0 (2 vs. .50), indicating that all children were able to translate at least two words and produce one adequate sentence.
Descriptive statistics for second language (L2) content-specific word knowledge measures at T1 and T2.
Notes. N = number of participants. SD = standard deviation. SE = standard error. BCa = bias corrected and accelerated. CI = confidence interval. Unless otherwise noted, bootstrap results are based on 1,000 bootstrap samples.
Paired-samples t-tests results for receptive and productive meaning and form knowledge (MFK).
Notes. SE = standard error. CI = confidence interval. Unless otherwise noted, bootstrap results are based on 1,000 bootstrap samples.
To further analyse learners’ receptive MFK, Table 3 presents the percentage of children who correctly translated each of the 20 content-specific words at T1 and T2, grouped according to their frequency band (i.e. 1K, 2K and off-list words), and the average improvement in each word category. As expected, the greatest meaning and form recall occurred in relation to 1K words, both at T1 and T2 (53.2% vs. 60.4%), and the largest improvement was found with regard to off-list words (+25%), even though receptive MFK improved across categories.
Receptive meaning and form knowledge (MFK) with regard to content-specific word frequency band.
Notes. * average percentage (with gain in parentheses).
With regard to learners’ productive MFK, Table 4 presents the percentage of children who used the content-specific words in a sentence adequately and/or accurately at T1 and T2, according to the frequency band of the words, and the average improvement in each word category. As can be seen, rates of productive MFK are rather low at T1 across categories, especially in terms of 2K and off-list words (only 3.9% and 1.2% of the children were able to use these words in a sentence) and much lower than the rates for receptive MFK at the same time point. These rates improve substantially at T2, notably in the categories of 2K and off-list words (+271.8% and +716.7% respectively). Nevertheless, the rates for productive MFK are still below the rates observed for receptive MFK at the same time point.
Productive meaning and form knowledge (MFK) with regard to content-specific word frequency band.
Notes. * average percentage (with gain in parentheses).
Turning now to learners’ productive vocabulary knowledge in a content task, Tables 5 and 6 present the descriptive and inferential statistics for the different measures used in the study. All measures improved from T1 to T2, except for lexical density. The Corrected Type Token Ratio (CTTR) increased from T2 to T1 (1.42 vs. 2.17), the difference being statistically significant (t(15) = −3.51, p = .01, 95% CI [−1.17, −.38]) with a strong effect size (d = 1.23). This indicates that learners’ written productions were lexically more diverse (i.e. contained more word types and tokens) at the end of the intervention than at its onset. With regard to lexical sophistication, children used more 1K and 2K words at T2 than at T1 (.97 vs. 1.56 and .12 vs. .18, respectively). The improvement was statistically significant only for the 1K measure (t(15) = −2.72, p = .02, 95% CI [−1.03, −.21]). As for lexical density, it decreased from T1 to T2 (.57 vs. .53): unlike at T1 when learners’ texts consisted mostly of lexical words (i.e. high density), at T2 the learners produced more elaborate texts, comprising both lexical and function words, with an impact on the lexical density of the text. The difference between lexical density measures at T1 and T2 was not statistically significant.
Descriptive statistics for productive vocabulary knowledge measures at T1 and T2 and their progress rates.
Notes. N = number of participants. SD = standard deviation. SE = standard error. BCa = bias corrected and accelerated. CI = confidence interval. Unless otherwise noted, bootstrap results are based on 1,000 bootstrap samples.
Paired-samples t-tests results for productive vocabulary knowledge measures.
Notes. SE = standard error. CI = confidence interval. Unless otherwise noted, bootstrap results are based on 1,000 bootstrap samples.
With regard to the correlations between the different dimensions of learners’ (content-specific) L2 vocabulary knowledge, Table 7 presents the bootstrapped Pearson product moment correlations at T1. There is a positive and moderate correlation between receptive and productive MFK (statistically significant at p = .006), meaning that children who translated more words were also more able to use them in a sentence, but they were generally better at translating than at producing sentences. There is also a positive and moderate correlation between learners’ productive MFK and the proportion of 1K words used in the texts (statistically significant at p = .02). With regard to the different dimensions of productive vocabulary knowledge, there is a positive and strong correlation between the lexical diversity of the texts as measured by the CTTR and the number of 1K words used (statistically significant at p < 0.01) and a positive but moderate correlation between lexical diversity and lexical density (statistically significant at p = .03).
Pearson product moment correlations between depth and breadth of vocabulary knowledge measures at T1.
Notes. * Correlation significant at the .05 level (two-way). ** Correlation significant at the .01 level (two-way).
Most of the correlations observed at T1 are maintained at T2. Table 8 presents the bootstrapped Pearson product moment correlations at T2. The correlation between receptive and productive MFK strengthens (statistically significant at p < 0.01), indicating that at the end of the intervention children were more likely to both translate the words and use them in a sentence adequately and/or accurately. The correlation between learners’ productive MFK and the proportion of 1K words used in the texts remains moderate and positive (statistically significant at p = .04). The correlation between the lexical diversity measure (CTTR) and the number of 1K words used is maintained (statistically significant at p = .005) but the correlation between lexical diversity and lexical density is no longer statistically significant.
Pearson product moment correlations between depth and breadth of vocabulary knowledge measures at T2.
Notes. * Correlation significant at the .05 level (two-way). ** Correlation significant at the .01 level (two-way).
V Discussion
The aim of the current study was to explore the effects of participating in a GEP CLIL intervention on primary school children’s knowledge of content-specific vocabulary in EFL. Our results indicate that the intervention had a positive impact on children’s depth of L2 content-specific vocabulary knowledge (i.e. their receptive and productive knowledge of word meaning and form) and breadth of L2 productive vocabulary knowledge in a content task despite its relatively short span (3 weeks).
With regard to research question 1, our findings are in line with previous research on receptive and productive vocabulary gains in CLIL (vs. non-CLIL) primary education learners (Canga Alonso, 2013; Jiménez Catalán & Ruiz de Zarobe, 2009; Jiménez Catalán et al., 2006; Xanthou, 2011) in that the children experienced a significant improvement both in their receptive and productive knowledge of word meaning and form with regard to the selected content-specific lexis. The fact that they had had multiple opportunities to encounter the words in the learning materials and the lexis was made salient through teacher talk, as corroborated by class transcripts (example 1, excerpt from the last session; for transcription conventions, see Appendix 1), can be assumed to have contributed to the children’s gains.
(1) Teacher (T): (. . .) okay a:: what do you remember? ↑ for example do you remember the food groups?↑ S1 (Student 1): e:: no S2 (Student 2, etc.): yes Teacher: S2↑ S2: proteins and foods x x x Teacher: goo:d S2: dairy products T: ye::s S2: a:nd fats oils and sugars T: oils and suga:rs = anything e:lse? ↑ S2: and grains T: and gra:ins ↑ very good so do you remember the servings? ↑ (.) per da:y ↑ (.) more or less? S3. S3: for example the proteins 3 5 servings a week Teacher: 3 to 5 servings a wee:k? ↑ the proteins? ↑ S3: yes Teacher: good ↑ = so a da:y S3: a da::y two: or three Teacher: two or three a day (.) okay ↑ = a:: S4, what do you remember? S4: fa:ts and sugars Teacher: ye::s S4: e::m e:m:: (.) o::n:e Teacher: one time? S4: one time a week Teacher: one time a week (.) more or less (.) yes? ↑ spa:ringly = do you remember this? S4: yes Teacher: strange word? spa:ringly? oka:y (.) xx S4: e::: (.) grains Teacher: gra:ins S4: two (.) three (2) two three::: servi:ngs per week
The gains were across-the-board, in high, low-frequency and off-list words, which indicates that children can learn lexical items typically acquired at more advanced proficiency levels provided they are exposed to them in a meaningful context. The CLIL sessions in our study provided plenty of support to comprehensible input and numerous opportunities for output and interaction through content tasks. Nonetheless, similar to what was found in relation to general vocabulary learning in CLIL, receptive knowledge of content-specific lexis is more robust than productive knowledge: children are more capable of recalling the meaning of the words than of using them in sentences. This could be attributed to the fact that productive use of vocabulary is cognitively more demanding than reception (Webb, 2008), difficulty which is most likely incremented in the case of lexis associated with specialized knowledge domains (i.e. nutrition). As a matter of fact, some of the children relied on formulaic structures such as ‘I like . . .’ which allowed them to use accurately, though not always adequately, the nouns in the test.
Turning now to research question 2, the CLIL intervention brought about an improvement in children’s productive vocabulary knowledge as elicited through a written content task (i.e. a recipe), in line with previous research on writing skills in CLIL which also explored the productive vocabulary dimension (Lasagabaster, 2008; Ruiz de Zarobe, 2010). By the end of the intervention, the children produced lexically more diverse and sophisticated texts. In some cases, the children went from no production to fully fledged texts (example 2), which indicates that even short-term CLIL interventions like the one in our study can be beneficial for the productive use of L2 lexis provided they incorporate the necessary scaffolding. The particular genre targeted in the production task had been presented in class, and the children had had an opportunity to write recipes collaboratively and receive feedback from the teachers. Moreover, production beyond word-level is a reliable indicator of learners’ progress in their development of subject matter literacy (Llinares et al., 2012; Pavón, 2018).
(2) Peel the potatoes, cut the potatoes and cut the onion, put oil in the freyking [frying] pan, united the potatoes with the onion, and put in the freyking [frying] pan, break the eggs in a boul and put salt. Stire the eggs, put the onion and the potatoes, and put all on the freyking [frying] pan, and wait the recipe finish when the omelette is gold, and jay of your omelett. (S1, T2)
Finally, with regard to research question 3, our findings indicate that reception and production as well as depth and breadth are interconnected dimensions of learners’ L2 lexicon (see also Vermeer, 2001). Children who recall more content-specific word meanings tend to be more capable of using them in sentences, a correlation which strengthens by the end of the intervention. Nonetheless, one should bear in mind that overall children are better at reception than at production, meaning that we cannot assume that receptive mastery will effortlessly develop into productive mastery (Schmitt & Schmitt, 2014). Moreover, children who can use content-specific words, and hence have more elaborate lexicons, are also more prone to producing high-frequency words in the content task, implying an interplay between Cummins’ (1979) BICS and CALP (Llinares et al., 2012). With regard to the different dimensions of productive L2 vocabulary knowledge, lexical diversity, sophistication and density seem to be connected from the initial stages of CALP (for evidence on this connection with secondary education CLIL learners, see Lorenzo and Rodríguez,2014).
VI Pedagogical implications
The present study has shed additional light on current EFL teaching practices applying a CLIL approach in Catalan primary schools and, more specifically, on the impact of a 4-week GEP CLIL programme on students’ L2 subject-specific vocabulary learning. Despite the short span of time, positive effects on both students’ breadth and depth of content-specific vocabulary have been documented. This implies that the CLIL approach, with the amount and type of scaffolding provided through the GEP programme, starts enhancing vocabulary learning even after a few weeks of implementation. However, further research is needed to see whether there are greater gains as a result of longer periods of instruction and if these gains are sustained in time.
Examining teacher input and classroom materials in this particular GEP CLIL intervention has made it possible to see how the quality of L2 learning in CLIL programmes is closely related not only to the type of instruction, but also to the pedagogical quality of its classroom implementation. The fact that the CLIL teachers in our study received constant guidance and tuition from experts on CLIL lesson planning and material design as part of the GEP teacher training enabled them to integrate content learning and language learning more effectively and to offer a rich, meaningful and engaging learning experience to this particular group of primary education students. Thus, participating in this innovative programme seems to help teachers gain expertise in developing high quality CLIL lesson plans and materials which guarantee an effective and successful implementation afterwards. Nevertheless, more studies are needed to explore how CLIL teaching is put into practice in other GEP schools with CLIL programmes of different lengths and intensities, for instance. A larger participant population as well as a proper experimental design, including a control group, would be needed to explore these issues possibly through a university-school collaboration.
Finally, our study highlights the importance of obtaining vocabulary profiles of classroom input to better gauge the type of L2 learning taking place in CLIL classrooms, which is just partially revealed by standardized vocabulary tests. While these tests provide an estimate of students’ vocabulary size, they say little about the content-specific vocabulary the CLIL learners have been exposed to and have actually practiced in the CLIL lessons. Practitioners would benefit from more training on vocabulary profiling tools to design tailored tests when implementing CLIL projects.
Footnotes
Appendix 1. Transcription conventions
Source. Escobar Urmeneta, 2013.
(.) unmeasured (micro-)pause of less than two tenths of a second
= latching (no gap) between utterances produced by the same speaker or different speakers
↑ rising intonation
: stretching the preceding sound; more colons, more stretching
Acknowledgements
The authors wish to acknowledge the teachers and the children who participated in the study. Alexandra Vraciu is a Serra Húnter Fellow at the Faculty of Education (Universitat de Lleida, Spain).
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
