System 66 (2017) 87e99

Contents lists available at ScienceDirect

System
journal homepage: www.elsevier.com/locate/system

CLIL or time? Lexical profiles of CLIL and non-CLIL EFL

learners

n*, M. Pilar Agustín Llach

nez Catala
Rosa M. Jime

n, Departamento de Filologías Modernas, C/ San Jos

Universidad de La Rioja, Facultad de Letras y Educacio e de Calasanz s/n, 26004
~ o, La Rioja, Spain
Logron

a r t i c l e i n f o a b s t r a c t

Article history: Recently, there has been considerable research concerning the effect of CLIL on English
Received 19 October 2015 language learners’ competence. However, it remains unclear if the positive effects found
Received in revised form 18 February 2017 are due to CLIL or to time. To clarify this issue, this paper focuses on the vocabulary output
Accepted 15 March 2017
of CLIL and non-CLIL EFL learners after an equal number of hours of English exposure. The
Available online 25 March 2017
objectives were twofold: (1) to ascertain whether the CLIL group retrieves a higher number
of English words than the non-CLIL group; (2) to determine whether the two groups
Keywords:
produce the same or different words. The sample comprised 70 Spanish EFL learners in
CLIL vocabulary
Lexical availability
their 8th and 10th year of secondary education. The data collection instrument was a
Word frequency lexical availability task consisting of ten prompts. The data were edited, coded, and sub-
Word level jected to quantitative and qualitative analyses. The results showed that the CLIL group
Lexical Frequency Profile retrieved a higher number of words than the non-CLIL group. However, both groups
Cambridge Vocabulary Profile exhibited similarities concerning most and least productive prompts, first word responses,
word frequency, and word level. The findings suggest a need to conduct equal comparisons
of CLILs and non-CLIL groups as well as to examine the task effect, and the vocabulary
input received by learners.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

There is a great deal of variation when it comes to research concerning the effect of Content and Language Integrated
Learning (CLIL) on the development of competence in English as a Foreign Language (EFL). This variation primarily relates to
focus, topic and methodologies. For example, certain studies have examined classroom discourse in CLIL contexts (cf. Nikula,
Dalton-Puffer, & Llinares, 2013; Whittaker, Llinares, & McCabe, 2011; Evnitskaya & Morton, 2011), while others have
compared CLIL and non-CLIL groups in regard to different skills and language components, such as listening and reading
(Pladevall-Ballester & Valbona, 2016), receptive vocabulary size (Ferna
ndez-Fontecha, 2015), and grammar (Aguilar & Mun ~ oz,
2013). The tendency in CLIL studies has been to explore new avenues rather than to build on previous research; indeed, this
makes it difficult to draw definitive conclusions, as there does not currently exist a solid body of empirical evidence related to
each of the aspects investigated thus far. This occurs, for instance, when it comes to vocabulary, as researchers have dealt with
receptive knowledge rather than with productive knowledge. With regard to the latter, scholars have focused on vocabulary
size instead of on the qualitative aspects of the words. However, the most serious limitation of CLIL research, as pointed out by

* Corresponding author.

nez Catala
E-mail address: rosa.jimenez@unirioja.es (R.M. Jime
n).

http://dx.doi.org/10.1016/j.system.2017.03.016
0346-251X/© 2017 Elsevier Ltd. All rights reserved.
88
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

Bruton (2011, 2013), is that most studies have based their conclusions on groups with an unequal number of hours of language
exposure. This makes it almost impossible to disentangle the effect of CLIL from the effect of time. The present study is the
first to attempt to overcome these limitations by looking at the quantity and quality of the vocabulary output of CLILs and
non-CLILs groups, and particularly by comparing these groups when they are exposed to an equal number of hours of English.
Our objectives are two-fold. The first is to ascertain whether CLIL EFL learners produce more English words in a lexical
availability task than EFL learners enrolled in an ordinary formal instruction programme (non-CLIL). Our second objective is to
identify the word types produced by a CLIL group compared to two non-CLIL groups, and to contrast them in terms of three
dimensions of productive vocabulary knowledge: lexical availability, word frequency, and word level.
Vocabulary production has been shown to correlate positively with language level as well as with success in
communication (see Milton, 2013). Likewise, research has also found that infrequent and advanced words are indicators
of lexical richness and language proficiency (Alderson, 2007; Meara, Lightbown, & Halter, 1997; Morris & Cobb, 2004;
Laufer, 1998). Since CLIL learners are exposed to higher and richer language input in their learning experience (Cenoz,
2013; Mun ~ oz, 2015; Va
rkuti, 2010; Xantou, 2011), we might predict that they will retrieve more words and exhibit
more lexical richness than non-CLIL EFL learners; this refers to word frequency and word level, i.e. more infrequent or
advanced ones. As we will see in the background section, some studies have compared the vocabulary output of CLILs and
non-CLILs groups by means of vocabulary tests, compositions, and lexical availability tasks; however, these comparisons
have been based on unequal terms. In addition to this, the qualitative aspects of the vocabulary output produced by the
groups, such as word frequency, patterns of word responses in lexical availability, and word level, have not been
addressed in CLIL studies.
In the remainder of the paper, we first provide a synthesis of productive vocabulary studies within CLIL research, with
emphasis on the task used during data collection, and the implications of the findings. We then explain the assumptions of
word frequency, lexical availability, and word level, as well as the dimensions of word knowledge addressed in the present
study. After posing our research questions, we describe our methodology, report the findings, and discuss said findings in
conjunction with previous research. We end the paper with a summary of conclusions and suggestions for further research.

2. Background

2.1. Research on CLIL learners’ productive vocabulary

Research on the productive vocabulary of CLIL EFL learners is insufficient and rather varied in terms of focus, data
collection task used, and educational context. As shown in Table 1, different aspects of word knowledge have been examined
by previous research, with the letter writing task being the most frequently used data collection instrument; in addition to
this, studies on primary education have been prevailing over those concerning secondary or tertiary education.
Overall, and as illustrated in Table 1, previous studies have corroborated the CLIL advantage. For instance, with regards the
use of vocabulary in letter writing, CLIL EFL learners have, in comparison with non-CLIL learners in Finland and Spain, shown
higher performance in lexical production, higher lexical variation, lower word repetition, and fewer errors (e.g., Pietila €&
Merikivi, 2014; Ruiz de Zarobe, 2010; Celaya & Ruiz de Zarobe, 2010; Agustin Llach & Jime
nez Catal
an, 2007; Agustín
Llach, 2014, 2009). Likewise, in a study concerning the use of borrowings and lexical inventions by 7th and 10th graders
in Catalonia and the Basque Country, Celaya and Ruiz de Zarobe (2010) found that lexical inventions depended on the age of
the learners, but borrowing behaviour was more dependent on the type of learning programme, with CLIL learners borrowing
less than learners who used traditional approaches.
However, inconclusive findings, or counter evidence, have also been found. With regard to letters, in Catalan et al.'s
(2006) study, the non-CLIL group used a higher number of words than the CLIL group, although the latter exhibited
higher lexical variation. In terms of vocabulary tests, Sylve
n (2004) in Sweden, and Seregely (2008) in Austria (in a replica of
the former) reported higher performance for CLIL groups compared to traditional approaches. An important point to
consider when examining Sylven's study is that CLIL was identified as an important factor in encouraging learners to use
English outside the class during, for example, reading activities. As such, it was unclear as to whether the higher number of
words produced by the CLIL groups was due to the effect of CLIL or to extramural activities. As regards Lex30 (Meara &
Fitzpatrick, 2000), the study conducted by Moreno (2009) illustrated that CLIL 6th grade EFL learners did not produce
associations of a higher level than non-CLIL learners; in fact, the word associations observed between the stimulus and
learners' word responses were similar across the two groups. This finding mirrors that reported by Karonen (2003) (as
quoted in Pietila€ & Merikivi, 2014) whose study focused on 6th and 9th graders in Finland; they found that the CLIL and
non-CLIL students produced syntagmatic rather than paradigmatic associations, which are usually more frequent among
English native speakers. With regards the lexical availability of CLIL and non-CLIL EFL learners, the only study of which we
are aware was carried out by Jime
nez Catala
n and Ojeda Alba (2009a), with the authors observing that non-CLIL groups
have an advantage when it comes to lexical output.
While the balance seems to tip in favour of CLIL EFL learners having a larger and more varied productive vocabulary, more
studies are needed since we cannot overlook the existence of inconclusive or opposing findings. Furthermore, most CLIL
research on productive vocabulary has focused on primary school EFL learners. As yet, we do not know whether the same
tendency will be exhibited among secondary school education learners. We are also unsure as to whether CLIL groups will
enjoy an advantage when it comes to the quality of lexical output.

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 89

Table 1
Studies of the effect of CLIL on vocabulary production and use.

Study Focus Instruments Context: Participants background

Jime
nez Catala

n & Lexical phrases Letter 6th Primary (N 60)
Fern

andez Fontecha, 2015 Language level Cloze 11e12 years old
Social Science 1st, 2nd grade
Arts & Crafts 4th, 5th grade
Hours: CLIL ¼ 960; non-CLIL ¼ 629
e
€ & Merikivi, 2014
Pietila Extramural activities VLT receptive & 6th (N 150) vs 9th (N 180)
productive 13e16 years old
Questionnaire Finland
Agustín Llach, 2014 Lexical profile Letter 4th Primary, (N 72)
Transfer 9e10 years old
Social Sciences
Hours: CLIL ¼ 714; non-CLIL ¼ 419
e
Jime
nez Catala
n & Lexical output Lexical availability task 6th (N 25) vs 8th grade CLIL (N 25)
Fitzpatrick, 2014 Lexical frequency 11e12 years old vs (13e14 years old)
Social Sciences
Hours: CLIL ¼ 839 þ 350; non-CLIL ¼ 629
e
Celaya & Ruiz de Zarobe, 2010 Oral and written production The Frog 10th Non-CLIL, age 16, N 18. Hours ¼ 792
borrowings and lexical Letter 10th CLIL, age 16, N 16. Hours ¼ 1120
inventions 10th CLIL, age 16, N 17. Hours: 1155
7th Non-CLIL, age 12.9, N 20. Hours ¼ 416
7th CLIL, age 12.9, N 22. Hours ¼ 1128
Basque country/Catalonia

nez Catala
Jime
n & Lexical availability Lexical 6th Primary (N 86) CLIL (N 42) non-CLIL (N 44)
Ojeda Alba, 2009a Language level availability Task: 11e12 years old
Cloze Hours: non-CLIL ¼ 629, CLIL 629 þ 80
Social sciences
e
Marika English Language proficiency Language Upper secondary School

n
Kjelle Development proficiency: International Baccalaureate
Simes, 2008 Motivation Low frequency (age 16)
Longitudinal words English Medium of Instruction
Motivation (N 43) vs English as a subject
Narratives Sweden
Diagnostic test
Moreno, 2009 Productive Vocabulary Profile Lex30 6th Primary (N 130) Females
Word Associations Science 1st, 2nd grade
Arts & Crafts 4th, 5th grade
Hours: CLIL ¼ 960; Non-CLIL ¼ 629
Spain
Agustín Llach, 2009 Language proficiency Cloze test 6th Primary (N 60) Females
Lexical transfer Letter Science 1st, 2nd; Arts & Crafts 4th, 5th grade
CLIL ¼ 960; Non-CLIL ¼ 629
e
Seregely, 2008 Receptive/Productive Self-report Vocabulary test 11th Secondary; CLIL (N 21), non-CLIL (N 12)
Vocabulary Word in context test Natural sciences, Music, Arts and
Word perception Multiple choice test Physical Education
Test effect Cloze test Austria
Effect of extramural
Activities
Gender
Agustín Llach & Jime
nez Lexical reiteration
Catal

an, 2007 Lexical variation
Jime
nez Catala

n, Ruiz de Receptive vocabulary Cloze 6th Primary (N 130) Females
Zarobe, & Cenoz, 2006 Reading Reading Content: Science 1st, 2nd grade; Arts & Crafts
comprehension VLT (receptive) 4th, 5th grade
Language Level Letter Hours: 960; non-CLIL ¼ 629
Lexical variation e

n, 2004
Sylve Receptive/productive vocabulary Multiple-choice test CLIL students (N 99) versus non-CLILs (N 264)
CLIL/Extramural exposure VKS Science
Longitudinal study: 2 years Test on phrases and Secondary education
collocations Sweden
Cloze test
90
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

2.2. Vocabulary knowledge dimensions: word frequency, lexical availability, word level

Vocabulary learning and development is multifaceted in nature (Schmitt, 2008). In an attempt to capture its essence,
scholars have identified several dimensions of vocabulary knowledge such as breadth, depth, receptive, and productive (see
Milton & Fitzpatrick, 2014); of the latter dimensions, word frequency and lexical availability are considered the most relevant
dimensions of word knowledge (Faerch, Haastrup, & Phillipson, 1984). Word frequency allows us to track the words known to
learners compared with their occurrence in texts produced by native speakers of the target language. Lexical availability, in
contrast, can tell us about the number and the characteristics of the words which learners are capable of retrieving in a given
time in response to word prompts related to daily life situations such as, for instance, transport, or clothes.
Word frequency studies have been mainly concerned with the measurement of vocabulary knowledge on the basis of
frequency levels ranging from high frequent to low frequent words (cf. Laufer & Nation, 1999; Meara & Fitzpatrick, 2000;
Schmitt, Schmitt, & Clapham, 2001). Some studies have also shown that high frequent words are acquired earlier than low
frequent words, and that the higher the number of infrequent words known to language learners, the higher their degree of
lexical richness (e.g. Meara et al., 1997; Nation & Waring, 1997) and the higher their proficiency in the target language (e.g.
Alderson, 2007; Morris & Cobb, 2004; Laufer, 1998). These studies have claimed that vocabulary production beyond the 1k
band predicts advanced vocabulary; in particular, Academy Word List (AWL) and Off-list words have long been considered the
most infrequent words, corresponding to the most advanced level.
Lexical availability research aims to identify speakers' available lexicon. The underlying assumptions are three-fold. First,
the higher the number of informants who retrieve the same word in response to a prompt such as transport, the greater the
availability of this word in the lexicon of the group. Secondly, first word responses are the most available. As such, when the
same first word responses are shared by the majority of speakers, availability is closely related to typicality, understood here
as ‘typical examples of a particular category’ (Evans, 2007: 215). Thirdly, the words retrieved in response to word prompts
uncover different types of associations that point to the organisation of the lexicon (cf. Hern
andez, Izura, & Ellis, 2006).
The lexical availability methodology was used to explore the available lexicon of English monolinguals in Scotland (cf.
Dimitrije
vic, 1969) and English and Spanish bilinguals in the USA (cf. Moreno Ferna
ndez, 2007; Bayley, 1971). More recently,
the lexical availability methodology has been applied to the study of the lexical output of EFL learners in classroom contexts

nez Catal

(cf. Jime an & Ojeda Alba, 2009a,b). This research has been mainly concerned with the description of learners' word
responses in terms of most and least productive prompts and with the identification of the effect of gender, age, and language
level on learners' lexical availability output (cf. see Jime
nez Catal
an & Fitzpatrick, 2014). Lexical availability tasks have also
been employed to study different aspects of CLIL learners’ productive vocabulary, such as number of words retrieved overall
and for each prompt used in the task. With this said, however, as far as we know, these studies have not addressed the quality
aspects of word knowledge.
In addition to frequency and lexical availability, word level is also included in the qualitative analysis of the English lexical
output of CLILs and non-CLILs groups. To this purpose, we follow the classification of word level provided by Cambridge
Vocabulary Profile (CVP), an online resource based on Cambridge Learner Corpus and Bank of English. The use of CVP will
allow us to classify the English words retrieved by CLILs and non-CLILs into CEFR levels: A1, A2, B1, B2, C1 and C2. In addition,
it will also complement the frequency and lexical availability analyses.
CVP is based on two assumptions. The first is that each word can be ranged according to levels. The second assumption is
that there is a progression in word level; that is to say, words at A1 are lower in level than words at A2. In turn, words at A2 are
higher in level compared to A1, but lower than words at B1 level, and so on and so forth. The highest word level, C2, cor-
responds to a proficient user in the CEFR language level classification. The few studies to have included CVP in their analysis
have focused on university students (see Milanovic & Saville, 2015). To our knowledge, no research on CVP levels has been
published with samples of EFL learners from primary or secondary education, let alone with EFL learners from CLIL
programmes.
By including word frequency, lexical availability, and word level as constructs in our study, we hope to provide a more
substantiated analysis of the quality of the words produced by CLILs and non-CLILs, i.e. whether they produce more infre-
quent and advanced words, and whether their words reflect similar or different word association patterns. This quantitative
and qualitative analysis will provide us with more sound evidence which we can use to determine whether there is a real CLIL
advantage concerning vocabulary; vocabulary is, as mentioned earlier, a language component related to language level and to
success in L2 communication. Likewise, by focusing on the lexical output of CLILs and non-CLILs, we hope to contribute to L2
productive vocabulary studies in second languages e an area in need of further research (Schmitt, 2010).
As mentioned earlier, one problem with CLIL research is that it is mostly based on conclusions related to groups with an
unequal amount of language instruction. In parallel, other studies have revealed that an increase in the number of hours of
instruction has an effect on L2 language proficiency (e.g. Hakuta, 2000; Horst & Collins, 2006). Indeed, further research
concerning the relationship between number of hours of instruction and vocabulary acquisition is called for.
Our study aims to develop a clearer understanding of the effectiveness of CLIL on productive vocabulary by means of two
comparison approaches. In comparison 1, we conduct a classical CLIL study examining the lexical output of CLIL and non-CLIL
EFL learners of the same age and from the same school year, but with an unequal number of hours of English instruction. In
comparison 2, the lexical output of the above CLIL group is compared to a non-CLIL group. The members of this group fall into
slightly different age ranges and are in different school years, but have been exposed to an equal number of hours of English
instruction. The research questions are as follows:

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 91

Table 2
Sample summary.

N Grade Hours of exposure to English Approach

Group 1 24 8th 1189 (839 CLIL lessons þ 350 EFL) CLIL
Group 2 26 8th 839 (EFL lessons) Non-CLIL
Group 3 19 10th 1189 (EFL lessons) Non-CLIL

1) Do EFL learners in a bilingual CLIL programme produce a higher number of words than EFL learners in a regular non-CLIL
programme?
2) Do CLIL and non-CLIL EFL learners produce the same or different words in their first word responses to task prompts?
3) Do CLIL learners produce a higher number of infrequent words than non-CLIL learners?
4) Do CLIL learners produce more advanced words than non-CLIL EFL learners?

3. Method

3.1. Informants and study design

A total of 70 Spanish EFL learners participated in this study and were divided into 3 groups depending on their exposure
time and exposure type. They came from two middle-class high secondary education schools located in the same autonomous
community in the North of Spain. We looked at EFL learners' lexical availability output in an English bilingual programme
(CLIL), in which students were taught Natural Sciences and History in the language of English; they also attended regular
English as a school subject lessons. We then compared this output to EFL learners’ lexical availability in the regular foreign
language programme, in which students attended English as a subject classes but no CLIL classes.
With regards the first comparison, we examined the lexical output of 8th grade EFL learners who had been exposed to CLIL
and those who had not: 24 learners belonged to the 8th grade CLIL group, while 26 students were in the 8th grade non-CLIL
group. Both groups attended the same school, were in the same school year, and were in the same age range (13-14 years-old).
However, they differed in terms of the number of hours of English to which they had been exposed during class instruction.
Both the 8th grade CLIL group and the 8th grade non-CLIL group had accumulated 839 h of English as a subject after six years
of primary education and two years of compulsory secondary education (7th and 8th grade). In addition, the 8th grade CLIL
group had benefitted from 350 extra hours through instruction in two curriculum subjects in English (Natural Sciences and
History) throughout the 7th and 8th grades; they had thus accumulated a total of 1189 h of English exposure.
In the second comparison, we tallied the lexical output of the 8th grade CLIL EFL learners along with a group of 10th grade
regular (non-CLIL) EFL learners, comprising 19 students, with a similar number of hours of English exposure: 1189 h.
The comparison of the 8th grade CLIL group and the 8th grade non-CLIL group was made on the basis of the lexical
availability output of the two groups with the same age but a different number of hours of exposure. In order to overcome this
situation, we decided to include another group of learners in the regular EFL non-CLIL strand with the same number of hours
of exposure as the CLIL group, but with different age ranges: the 10th grade non-CLIL group in our sample. Since the members
of these groups were only two years apart and all adolescents, we assumed that the age-related cognitive differences are
small. In this sense, it was expected that differences in lexical production would result from the different teaching approach
used for the two groups as their exposure to English was the same in terms of hours; indeed, the approach used is believed to
determine learners’ proficiency level (see literature review). The samples are summarised in Table 2.

3.2. Data collection instruments and procedures

A background questionnaire and a lexical availability task were used for the two comparisons. The former was included to
gather information on learners' linguistic profiles and previous exposure to English; other factors were also included in the
questionnaire e namely motivation and language difficulties e although they were not analysed in the study itself. The lexical
availability task included ten prompts: ‘Parts of the Body’, ‘Dirty’, ‘Hold’, ‘The Environment’, ‘Food & Drink’, ‘The Earth’,
‘School’, ‘Science’, ‘Town’, and ‘Countryside’. Informants were asked to write, in 2 min, all of the words that came to their mind
in relation to the prompts given. These included, on the one hand, traditional prompts used widely in lexical availability
studies (cf. Samper Herna
ndez, 2002) and, on the other hand, prompts related to the semantic field of “science”, which is one
of the subjects studied in CLIL instruction.
The lexical availability task is a discrete time-controlled task which shows the words retrieved in response to a word
stimulus. Compared to embedded tasks, such as letters or narratives, the lexical availability task has several advantages, the
first being that it focuses exclusively on productive vocabulary rather than on other aspects, such as word recognition or word
grammar. In addition, it gives us an objective account of the different words that learners have stored in their mental lexicons
and which are activated in response to prompts related to specific vocabulary domains.
Each student was provided with a paper booklet containing the prompts, with each prompt displayed on a different page.
The instructions, as briefly explained above, were written in Spanish at the top of the page and were also read aloud to the
92
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

class. The task was administered in class by the first author in the presence of the class teacher. Once instructions were given,
the researcher named each prompt, controlling time by means of a timer (2 min per prompt).

3.3. Data editing and analyses

Learners' word responses were coded according to the following criteria: (i) errors were corrected, (ii) unintelligible
words, Spanish L1 words and words standing for commercial brands (e.g. Schweppes in response to ‘Food and drink’) were
discarded, (iii) repeated words in a prompt were counted only once, (iv) compound words and lexical phrases were regarded
as one lexical unit and counted as one word (e.g. orange juice).
Once the corpus was edited, we created three Excel files (8th grade CLIL, 8th grade non-CLIL, and 10th grade non-CLIL),
into which the words retrieved by each learner were typed before being coded. This process allowed us to run descriptive
statistics for the lexical retrieval of each group, both for the task as a whole, and for each of the prompts. The means
attained by the CLIL and non-CLIL groups were then subjected to a Wilcoxon test,1 which allowed us to analyse the sig-
nificance of the data.
The next step involved the calculation of internal and external frequencies attained by each group in regards to lexical
availability output. Internal frequency was based on learners' lexical output from the lexical availability task. Excel files
containing the words retrieved by the learners were saved as plain texts so as an internal frequency analysis could be run
using Wordsmith Tools (Scott, 2012); indeed, this made it possible to classify learners’ words into alphabetic and frequency
lists. The use of Wordsmith Tools was facilitated by the identification of the shared and non-shared vocabulary of the CLIL
group and the non -CLIL groups.
With regards external frequency, we made use of the Vocabulary Lexical Profile (VLP) e a web program based on the
original Lexical Frequency Profile (LFP), and created by Laufer and Nation (1995). This allowed us to analyse English learners’
lexical richness in writing.2 VLP classifies the word families contained in texts into four frequency bands that stand for 1000
(1k), 2000 (2K), academic words lists (AWL) and Off list words. The program provides users with a classification of words into
word frequency bands, number of running words (tokens), number of different words (types) and proportion of types to
tokens (Type-token ratio). Both the LFP and the web version VLP have been used in studies with EFL learners whose mother
tongues and learning contexts differ (e.g. Horst & Collins, 2006; Laufer & Nation, 1995; Meara et al., 1997).
Finally, and as previously mentioned, we manually classified the first word responses to the two most productive and the
least productive prompts into CEFR word levels (A1, A2, B1, B2, C1 and C2). This was done by using the web resource
Cambridge Vocabulary Profile (CVP).3

4. Results

4.1. Lexical availability output

Research question 1 aimed to ascertain whether 8th grade CLIL EFL learners would retrieve a higher number of words than
8th and 10th grade non-CLIL EFL learners when given the same lexical availability task. As shown by Table 3, the 8th grade
CLIL group retrieved a significantly higher number of words than the 8th non-CLIL group. However, this finding was hardly
surprising, as the CLIL group had benefitted from 350 h of extra exposure.
In contrast, the comparison of the 8th grade CLIL and the 10th grade non-CLIL group was quite revealing, as the key
variable was kept the same, e.g. number of hours of instruction. The means clearly indicated that, on average, the CLIL group
retrieved a higher number of words than the non-CLIL group. The Wilcoxon test, which was applied to the means, revealed
the existence of significant differences in the two comparisons: 8th grade CLIL versus 8th grade non-CLIL (w ¼ 498.
p ¼ 0.00), and 8th grade (CLIL) versus 10th grade (w ¼ 398. p ¼ 0.00) in favour of the 8th grade CLIL group in both cases.
This tendency in favour of the 8th grade CLIL group was also observed when we scrutinised the prompts one by one, since
this group systematically retrieved a higher number of words than the 8th grade and 10th grade non-CLIL groups in response
to each of the ten prompts included in the lexical availability task. Of these, the mean number of responses was significantly
higher for the prompts ‘Body’, ‘Dirty’, ‘Hold’, ‘Food & Drink’, ‘Environment’ and ‘Town’ for comparison 2 (8th grade CLIL group
versus 10th grade non-CLIL group), and for all prompts except for ‘School’ and ‘Countryside’ for comparison 1 (8th grade CLIL
group versus 8th grade non-CLIL group); these were very productive and very unproductive prompts, respectively.
Concerning the most and the least productive prompts, the three groups showed similar behaviour. As indicated by the
means in Table 4, regardless of the school year or type of instruction, prompts can be classified into the most and least
productive in terms of eliciting words from learners' lexicons. With regards the most productive, at the first, second, third and

1
The data followed a non-normal distribution, so we chose the Wilcoxon test to run comparisons.
2
This program has been developed and is maintained by Tom Cobb as software and made available as a free web resource (www.lextutor.ca).
3
Available at http://languageresearch.cambridge.org/wordlists.

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 93

Table 3
Mean number of words retrieved by CLIL and non-CLIL students in the three groups.

Mean SD Max Min

8th grade CLIL (n 24) 139.75 42.61 229 63
8th grade non-CLIL (n 26) 94.77 35.52 176 50
10th grade non-CLIL(n 19) 100.47 41.60 183 32

fourth positions, we found that ‘School’, ‘Food & Drink’, ‘Town’ and ‘Body’ were ranked similarly in the three groups. Indeed,
the only difference was the higher number of words retrieved, on average, by the 8th grade CLIL group. As for the rankings of
the least productive prompts, we noted similarities in the CLIL and non-CLIL groups: ‘Hold’ and ‘Dirty’ held very similar
positions in the rankings of each group. However, there were again significant differences in favour of the CLIL group for most
prompts, as revealed by the Wilcoxon test.

4.2. Shared and idiosyncratic word responses

We now address research question 2, which we used to query whether CLIL and non-CLIL EFL learners would retrieve the
same or different words in their first responses to the prompts included in the lexical availability task. However, due to space
restrictions, in this paper we will only examine the first word responses to two of the most productive prompts (‘School’ and
‘Body’) and the least productive ones (‘Dirty’ and ‘Hold’).
In Table 5, we display the first word responses elicited by each prompt in the CLIL and non-CLIL groups, together with the
number of learners who retrieved each word response. As can be observed, when it came to the two most productive
prompts, first word responses exhibited very similar patterns in the CLIL and non-CLIL groups in terms of coincidence of
responses, as well as word classes to which the responses belong. On the one hand, the most common first responses to the
prompts ‘School’ and ‘Body’ were Teacher and Head respectively. On the other hand, nouns rather than adjectives, verbs or
adverbs emerged in both prompts. The only exceptions were one adjective, which was a first response to ‘School’, and one
verb, which was a first response to ‘Body’, as can be explicitly seen in Table 5.
The expression with regard to the least productive prompts, first word responses to ‘Dirty’ and ‘Hold’ displayed a rather
different pattern. As Table 5 above shows, CLIL learners retrieved a higher number of responses than the non-CLIL partici-
pants. Furthermore, learners' first word responses exhibited major variation. For instance, ‘Hold’ elicited 45 different first
word responses; in other words, as many as 45 participants in the study (64.3%) each retrieved a different word. The least
productive prompts elicited adjectives, verbs and adverbs in addition to nouns; this result contrasts with the predominance of
nouns found in first responses to the most productive prompts.

4.3. Word frequency

Research question 3 enquired as to whether CLIL EFL learners would produce a higher number of infrequent words than
non-CLIL EFL learners. Here, we intended to go a step further in our analysis and examine first word responses from a more
qualitative perspective.
Table 6 displays the results of the word frequency analysis (VLP) applied to CLIL and non-CLIL learners' most and least
frequent first word responses. As can be observed, the CLIL group did not always retrieve more infrequent words than the
non-CLIL group. In fact, word frequency seems to have been affected by the specific prompt. For example, the comparison of
8th grade CLIL learners and 8th grade non-CLIL learners (same school year and age but different hours of instruction),
revealed how the CLIL group retrieved a higher number of words within the 1K band than the non-CLIL group in response to
‘School’, ‘Dirty’, and ‘Hold’.
However, the non-CLIL group produced more 1K words in response to ‘Body’. When we compare both groups at the
upper frequency bands, we see how the 8th grade CLIL group retrieved more Off-list words for ‘Body’ and ‘Dirty’;

Table 4
Mean responses retrieved by each group for each of the prompts.

Rank 8th grade CLIL X 8th grade non-CLIL X 10th grade non-CLIL X
1 School 19.88 School 17.81 School 18.21
2 Food 18.21 Town 13.81 Town 13.95
3 Town 16.71 Food 13.65 Food 13.74
4 Body 15.21 Body 10.58 Body 11.58
5 Earth 14.50 Earth 10.12 Earth 11.00
6 Environment 13.88 Countryside 8.69 Science 9.42
7 Science 11.83 Science 8.12 Countryside 7.84
8 Countryside 11.29 Hold 4.92 Environment 7.79
9 Hold 9.67 Dirty 4.00 Dirty 3.79
10 Dirty 8.58 Environment 3.08 Hold 3.16
94
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

Table 5
Word responses elicited by each prompt distributed by group (In parenthesis, the number of learners who retrieved each word as first response).

PROMPT 8th grade CLIL
SCHOOL Books (1) Boy (1) Building
(1) Classroom (1)
Mathematics (1) Notes (1) Pen/s (4)
Place (1) Science (1) Student (1)
Subjects (2) Table (1) Teacher/s (8)
Window (1)
BODY Arm/s (3) Ears (1) Elbow (1) Eyes (5)
Finger (1) Hand/s (1) Head (11)
DIRTY Adjective (1) Clean (5) Clear (1) Desert (1)
Empty (1) Football (1) Land (1) Many (1)
People (1) Mood (1) Mud (1) Noisy (1)
Old (1) Poor (1) Right (1) Short (1)
Spider (1) Tomato (1) T-shirt (1)
Ugly (1) Window (1)
HOLD Black (2) Candle (1) Celebrate (1)
Cold (2) Dog (1) Door (1) Fast (1)
Field (1) Glasses (1) Go (1) Held (2)
Hot (1) Old (1) Side (1) Spring (2)
Tree (1) Verb (3) Word (1)
8th grade non-CLIL
Bag (3) Book (1) Chair/s
(3) Classroom (2)
Desk (1) Gym (1) Head (1) Maths (1)
Music (1) Pen (1) Pencil (1)
Rubber (1) School-bag (1) Science/s
(2) Secretary (1) Subject (1)
Table (1) Teacher/s (3)
Arm/s (5) Eyes (4) Face (3) Hands (3)
Head (3) Hear (1) Heart (2)
Leg/s (2) Man (1) Nose (1) Shoulders (1)
Bad (1) Beautiful (1) Bedroom (2)
Black (1) Clean (1) Clothes (1) Dance (1)
Dirty dance (1) Dirty dancing (1)
Frightening (1) Happy (1) Hat (1)
Horrible (9) Pig (1) Rubbish (1)
Silly (1)
Bookcase (1) Boy (1) City (1) Hold on
(1) Holdest (1) Horse (1) Lamp/s
(2) Maths (1) One (1) Open (1)
Pen (1) Wardrobe (1) Years (2)
Young (1)
10th grade non-CLIL
Bad (1) Blackboard (1) Book (1)
Chair (1) Class (2) Friends (1) Maths
(1) Pencil (1) School bag (1)
Table (1) Teacher/s (7)
Arms (1) Ears (1) Eyes (2) Feet (1)
Hair (2) Head (7) Hear (1) Leg (2)
Money (1) Mouth (1)
Been (1) Car (1) City (2) Desk (1)
Dust (1) Friends (1) Funny (1)
Paper bin (1) Party (1)
Shit (1) Trash (1)
A paper (1) Black (1) Car (1)
Hand/s (2) Hold the time (1)
Strawberry (1) Holder (1)

however, and somewhat surprisingly, it was the non-CLIL group that retrieved more Off-list words for ‘School’ and ‘Hold’.
Off-list words are those words not included in the frequency lists employed for the VLP analysis e words whose fre-
quency is, generally speaking, very low. The non-CLIL group also retrieved a higher number of 2k words than the CLIL
group for ‘School’, ‘Body’, ‘Dirty’ and ‘Hold’. As for the comparison of the 8th grade CLIL and 10th grade non-CLIL groups
(same hours of instruction), the former retrieved a higher number of words in the following bands and prompts: 1K
(‘School’, ‘Body’, and ‘Hold’), 2k (‘Dirty’ and ‘Hold’), and Off-list (‘School’). In comparison, the 10th grade non-CLIL group
showed an advantage in the following bands and prompts: 1k (‘Dirty’), 2K (‘School’ and ‘Body’), Off-list (‘Body’, ‘Dirty’,
and ‘Hold’).

4.4. Word level

With research question 4, the aim was to ascertain whether CLIL learners would produce more advanced words than
non-CLIL EFL learners. To achieve this objective, the word responses elicited by two of the most productive prompts
(‘School’ and ‘Body’) and the two least productive prompts (‘Dirty’ and ‘Hold’) were classified into CEFR levels by means
of the Cambridge Vocabulary Profile. The results obtained are summarised in Table 7, where the number of word re-
sponses elicited appears to be distributed per prompt, group and word level. Table 8 completes this picture by providing
the words retrieved as first responses by the three groups under examination. These tables reveal that 8th grade CLIL and
8th and 10th grade non-CLIL learners hardly differed in terms of word level; that is to say, regardless of group or prompt,

Table 6
Lexical Frequency Profile of first responses to most and least productive prompts.

1000 Words 1K 2000 Words 2K Off-List

% % %
School
8th grade CLIL 78.57 7.14 14.29
8th grade non-CLIL 47.37 36.84 15.79
10th grade non-CLIL 70.00 30.00 0.00
Body
8th grade CLIL 71.43 14.29 14.28
8th grade non-CLIL 81.82 18.18 0.00
10th grade non-CLIL 54.55 18.18 27.27
Dirty
8th grade CLIL 47.62 33.33 19.05
8th grade non-CLIL 31.58 57.89 10.53
10th grade non-CLIL 53.84 23.08 23.08
Hold
8th grade CLIL 84.21 5.26 10.53
8th grade non-CLIL 60.00 13.33 26.67
10th grade non-CLIL 57.14 0.00 42.86

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 95

Table 7
Raw frequency distribution of first word responses per prompt and word level.

School Body Dirty Hold Total

8th grade CLIL A1 10 6 11 11 38
A2 2 1 5 5 13
B1 1 e 3 2 6
B2 1 e 2 e 3
8th grade non-CLIL A1 13 9 10 8 40
A2 4 1 3 3 11
B1 1 1 3 3 8
10th grade non-CLIL A1 10 9 7 3 29
A2 2 1 e 1 4
B1 e e 4 3 7

the predominant level was A1, followed, at a distance, by A2 and B1. Neither C1 nor C2 words were found in the learners'
responses. B2 words were only observed in the CLIL group with a rather reduced number of learners retrieving them
(only 3).
Similarity in word level is the common tendency in our data. However, some slight differences can be identified at A2 word
level when comparing the 8th grade CLIL group with the 10th grade non-CLIL group; indeed, the CLIL group retrieved a
greater number of words than the non-CLIL group. Another relevant point is the fact that B2 words were retrieved by the CLIL
group, although in a rather negligible number.

5. Discussion

The two comparisons conducted in the present study corroborate the advantage enjoyed by CLIL learners when it comes to
€ & Merikivi, 2014; Ruiz de Zarobe, 2010; Celaya & Ruiz de Zarobe,
lexical output, as claimed by previous research (e.g. Pietila
2010). Simply put, when exposed to a different and the same number of hours of English, the CLIL group retrieved a
significantly greater number of word types than the non-CLIL groups for each of the prompts used as a stimulus. One possible
explanation for this finding is that exposure to content in the language of English could have provided the CLIL group with
meaningful contexts that enhanced their incidental acquisition of English words. Another compatible explanation may well
be related to the intensive practice embarked upon by the CLIL group in comparison with the regular extensive practice
received by the non-CLIL groups; indeed, this could have helped the former to acquire a higher number of English words in an
incidental way. However, we cannot be sure of which of these interpretations might be most plausible, as research has
provided evidence of the positive benefits of enhancing words in the input provided to language learners (e.g., Mashhadi &
Jamalifar, 2015; Qian, 2008; Barcroft, 2003), and of incidental vocabulary learning through authentic texts (e.g. Pellicer-
Sa
nchez & Schmitt, 2010; Brown, Waring, & Donkaewbua, 2008). Likewise, certain studies have shown the positive effect
of intensive practice compared to extended or distributed practice (e.g., Collins & White, 2011; Serrano & Mun ~ oz, 2007;
Hinger, 2006; Spada & Lightbown, 1989). Enhanced language input, authentic materials, and intensive language practice
are among the main features of CLIL instruction (Mun ~ oz, 2007, pp. 17e26), but there remains a great deal of uncertainty
regarding which of these variables is more closely related to CLIL gains.
The 8th CLIL group and the 8th and 10th non-CLIL groups differed significantly in terms of the number of words retrieved
overall and for most of the prompts out of the ten used. Nevertheless, they did show a striking similarity when it came to the
most and least productive prompts; that is to say, the CLIL group and the non-CLIL groups retrieved more words in response to
‘School’ or ‘Parts of the body’ and fewer words in response to ‘Dirty’ and ‘Hold’. From an educational stance, this finding points
to a rather limited knowledge of the words that are usually associated with ‘Dirty’ and ‘Hold’ such as dirty play, dirty trick, and
dirty joke for the former, and hold on, hold a (meeting), or hold the (door) open for the latter.4 However, from a research
perspective, the similar tendencies observed in previous lexical availability research in Spanish as L1 and L2 (e.g., Herna
ndez
et al., 2006; Urrutia, 2003; Samper Hern
andez, 2002) as well as in English as L1 and L2 (e.g., Catal

an & Ojeda Alba, 2009 a,b;
Ferreira & Echeverría, 2010; Germany & Cartes, 2000) point to the existence of common patterns in the organisation of
learners' lexicon, where nouns and some semantic categories seem to predominate. Moreover, the comparative study con-
ducted by Ferreira and Echeverría (2010) on the lexical availability of English native and non-native speakers corroborated the
existence of the most and least productive prompts in the two groups: female English native secondary school students
retrieved a higher number of words than female EFL learners at university, but both groups reflected the same pattern of
retrieving more words for ‘Parts of the body’ and ‘Food & drink’ and less words for ‘Terrorism and crime’, and ‘Health and
medicine’. The authors also provided evidence supporting the predominance of certain words over other word classes in both
groups.

4
Examples selected from http://www.oxforddictionaries.com/es.
96
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

Table 8
First word responses classified into CEFR.

PROMPT 8th grade CLIL 8th grade non-CLIL 10th grade non-CLIL
SCHOOL A1 Books Boy Classroom Pen/s Place A1 Bag Book Chair/s Classroom A1 Bad Book Chair Class Friends Pencil
(A1) Student Subjects Table Teacher/s Desk Head Music Pen Pencil School bag Table Teacher/s
Window School-bag Subject Table Teacher/s A2 Blackboard Maths
A2 Building Mathematics A2 Gym Maths Rubber Secretary
B1 Science B1 Science/s
B2 Notes
BODY A1 Arm/s Ears Elbow Eyes Hand/s Head A1 Arm/s Eyes Face Hands Head A1 Arms Ears Feet Hair Head
(A1) A2 Finger Hear Leg/s Man Nose Hear Leg Money Mouth
A2 Heart A2 Eyes
B1 Shoulders
DIRTY A1 Clean Football Many People Old A1 Bad Beautiful Bedroom Black A1 Been Car City Desk Friends Funny Party
(A2) Poor Right Short Tomato T-shirt Clean Clothes Dance Happy Hat Pig B1 Dust Paper bin Shit Trash
Window A2 Dirty dance Dirty dancing Horrible
A2 Adjective Clear Desert Empty Noisy B1 Frightening Rubbish Silly
B1 Mood Spider Ugly
B2 Land Mud
HOLD A1 Black Cold Dog Door Fast Glasses A1 Bookcase Boy City Horse One A1 A paper, Black Car
(A2) Go Hot Old Tree Word Pen Years Young A2 Hand/s
A2 Field Held Side Spring Verb A2 Lamp/s Maths Open B1 Hold the time Strawberry, Holder
B1 Candle Celebrate B1 Hold on, Holdest Wardrobe

In the present study, ‘School’ and ‘Body’ seemed to trigger instances of words deeply entrenched in learners' minds. Of
particular note here are the words Teacher and Head; these are first word responses retrieved by the CLIL and the two non-
CLIL groups for the two most productive prompts. Indeed, these words could be considered, following Evans (2007), as typical
members of their corresponding semantic category, respectively. In contrast, the 45 different first word responses to the
prompt ‘Hold’, together with the diversity of word classes linked to ‘Dirty’ and ‘Hold’ revealed a low degree of typicality,
compared to ‘School’ or ‘Body’. This low degree of typicality is characterised by the predominance of nouns as well as high
agreement between learners' word responses. The fact that the least productive prompts were an adjective (Dirty) and a verb
(Hold) respectively might explain their low productivity and the variation in the responses they elicited. Further research in
this area is certainly warranted.
During our qualitative analysis, we examined learners' first word responses on the basis of frequency. In terms of the CLIL
group, we expected a greater number of infrequent words than the other groups, as this would mean a higher level, according
to the lexical frequency assumption advocated, for instance, by Alderson (2007), Morris and Cobb (2004) and Laufer (1998).
However, this was not corroborated by our data; what emerged instead was the predominance of 1k words and some in-
cursions in upper bands (infrequent words) with both the CLIL and the non-CLIL groups. Our results are in tandem with those
reported by Agustín Llach and Canga Alonso (2014), who also studied CLILs’ receptive vocabulary. In their comparison of CLIL
and traditional EFL learners over 5 years from primary school to secondary school, they found that the vocabulary of both
groups fell into the 1000 most frequent words. Likewise, our results are similar to those attained by Horst and Collins (2006),
who examined 6th grade French Canadian ESL learners, and the results reported by Jime
nez Catala

n and Fitzpatrick (2014),
whose sample comprised 6th and 8th grade Spanish EFL learners of an identical language level. Far from finding an increase of
words in the infrequent bands as the number of hours of instruction increase, these two studies found something fairly
different. Conducted in different contexts and with groups of learners whose mother tongues differed, the two studies found
an increase of 1k words.
In our view, the similarity of findings may be related to the fact that young English learners at primary and secondary
education levels continue to acquire basic vocabulary corresponding to the most frequent bands all through compulsory
education. Likewise, we underestimate the fact that a great percentage of hours of the target language exposure resulted from
English as a subject in the school curricula in the CLIL and the non-CLIL groups. In English as a subject, the vocabulary input
contained in course books and teaching materials is usually based on a mixture of criteria such as basic/core vocabulary,
frequency, range, or just suitability to learners’ age and interests (Nation, 2010).
The CVP classification of word level applied to the words retrieved by the CLIL and non-CLIL group corroborates the ev-
idence provided by the VLP analysis; again, here we observed similarities rather than differences. The CLIL group hardly
differentiated from the other groups in regards word level. Although they retrieved a significantly greater number of words
than the other groups, the words retrieved were predominantly A1 level, which was not very different from the other groups.
However, despite the similar quality of the words, some signs of further progress in the CLIL group were also revealed. For
instance, B2 and A2 words were found in the CLIL group, but not in the non-CLIL groups. The retrieval of compound nouns and
lexical units in the 8th grade non-CLIL group also provided further evidence in this regard, as it indicated that the use of
routinised lexical units is more frequent in low language levels (cf. Wray, 2008, 2002). The absence of lexical units (lexical
phrases) from CLILs' first word responses may point to a higher word level; indeed, this was also suggested by their responses
for the Adjective ‘Dirty’ and the Verb ‘Hold’, which involved some degree of metacognitive knowledge.
While these examples are quite inspiring, we must nevertheless be cautious in our interpretation of them, as they were
produced by only a few learners. Moreover, research by Jime
nez Catala

n and Ferna
ndez Fontecha (2015) with 6th grade

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 97

primary CLIL and non-CLIL EFL learners showed an insignificant relation between lexical phrases in writing and language
level, as measured by a Cloze test. Although we cannot ignore the reality of the above examples, it is convenient not to lose
sight of the strong similarities observed in the words retrieved by the CLIL and non-CLIL groups in the present study.
In light of this, it may well be the case that more CLIL instruction is needed for quality differences to appear. Although the
8th grade CLIL and the 10th grade non-CLIL groups in the present study had accumulated the same total of English exposure,
the fact is that only 350 h out of the 1189 h was CLIL instruction. However, it may also be the case that CLIL and non-CLIL
learners are not exposed to a different type of vocabulary regarding word frequency and word level. This issue must be
addressed in future CLIL research.

6. Conclusions

The present study included two comparisons, the aim of which was to examine quantitative and qualitative aspects of the
lexical availability of CLIL and non-CLIL EFL learners in secondary school education. The first comparison, mirroring previous
research, looked at a CLIL group and a non-CLIL group at the same course level, but with a different number of hours of English
exposure. The second comparison aimed to solve this problem by comparing learners of a similar age and educational level
who had received the same amount of English instruction.
We combined the analysis of lexical availability output with a word frequency analysis (VLP) and Cambridge Word Level
Analysis (CVP) so as to gain insights into the quantity and quality of words retrieved by the CLIL group and the non-CLIL
groups in a lexical availability task. We aimed to establish whether 8th grade EFL learners on a CLIL programme had an
advantage over 8th grade non-CLIL groups receiving regular EFL instruction when it came to higher lexical output, lexical
variation, infrequent words, and advanced words.
Our findings indicated that, when exposed to a different number of hours of English and the same number of hours of
English, the 8th CLIL group significantly retrieved more words than the non-CLIL groups for all the prompts included in the
lexical availability task. However, the CLIL group did not differ from the non-CLIL groups in terms of their most and least
productive prompts, nor in the specific words retrieved. Of particular note here is the similarity observed in the words
retrieved by the CLIL and non-CLIL learners, as first responses to the most and less productive prompts. Indeed, this similarity
was also corroborated by the word frequency and word level analysis which was applied to the data, as assessed by using both
the VLP and the CVP.
Our study has provided evidence of the advantage enjoyed by CLIL learners in terms of a higher lexical availability output;
however, there was not so much evidence to show an advantage in terms of lexical frequency or word level. With this said,
care should be taken when extrapolating these findings. This is because, on the one hand, the results are based on a small
sample of learners belonging to a specific CLIL context while, on the other, the data were elicited from a specific lexical
availability task. Although it is clear that this research is fairly robust, there is a need for further studies with larger samples of
CLIL students, as well as different content subjects.
Likewise, it might well be the case that some tasks favour EFL learners rather than CLIL learners or vice versa. With
regard to future comparisons, tasks should cater equally to the needs of CLIL learners and non-CLIL learners. Another
essential step is to examine whether bilingual CLIL programmes and EFL regular instruction prepare learners for
communication by means of the appropriate vocabulary. One approach to investigating this issue is to use contextualised
prompts to activate learners' words related to daily situations such as ‘At the School’ or ‘At the Supermarket’ instead of
using ‘School’ or ‘Supermarket’. Another way in which to obtain further insights into CLILs' effectiveness is by conducting
research into the lexical availability of CLIL learners in relation to specific content subjects. In this regard, it is particularly
important to study the relation between the vocabulary input and the lexical availability output of CLIL learners in specific
school subjects at different course grades.
The last, but far from least important thing to consider is that, although the analysis of first responses to the most and least
productive prompts has provided us with relevant insights into the lexical availability of CLIL and non-CLIL learners, further
analysis of all the responses is required to complement this study. In particular, it is essential to further analyse all the re-
sponses retrieved for prompts closely related to content such as ‘The Earth’, ‘The Environment’, ‘Countryside’ and ‘Town’. In
this respect, the prospective investigation of the word associations between the first and second responses per prompt
retrieved by learners is a promising line of research which would provide further insights into the effectiveness of CLIL when
it comes to learners' vocabulary.

Acknowledgements

This work was supported by the Spanish Ministerio de Ciencia e Innovacio
n under Grant number FFI 2010-19334. We are
grateful to the students and teachers who agreed to collaborate in this study. Our thanks also go for the directors of the two
high-schools and students' parents who granted permission to collect the data.

References

~ oz, C. (2013). The effect of proficiency on CLIL benefits in engineering students in Spain. International Journal of Applied Linguistics, 24(1),
Aguilar, M., & Mun
1e18.
98
n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala

Agustín Llach, M. P. (2014). Exploring the lexical profile of young CLIL learners: Towards an improvement in lexical use. Journal of Immersion and Content-
based Language Education, 2(nº 1), 53e73.
Agustín Llach, M. P., & Canga Alonso, A. (2014). Vocabulary growth in young CLIL and traditional EFL learners: Evidence from research and implications for
education. International Journal of Applied Linguistics. http://dx.doi.org/10.1111/ijal.12090.
Agustín Llach, M. P. (2009). The role of Spanish L1 in the vocabulary use of CLIL and non-CLIL EFL learners. In Y. Ruiz de Zarobe, & R. M. Jime
nez Catala

n (Eds.),
Content and language integrated learning. Evidence from research in Europe (pp. 112e129). Bristol: Multilingual Matters.
Agustín Llach, M. P., & Jime
nez Catala
n, R. M. (2007). Lexical reiteration in EFL learners’ essays: Does it relate to the type of instruction? International Journal
of English Studies, 7(2), 85e103.
Alderson, C. (2007). Judging the frequency of english words. Applied Linguistics, 28(3), 383e409.
Barcroft, J. (2003). Distinctiveness and bidirectional effects in input enhancement for vocabulary learning. Applied Language Learning, 13, 133e159.
Bayley, V. J. (1971). A study of lexical availability among monolingual-bilingual speakers of Spanish and English. Unpublished Thesis. Houston: Rice University.
Brown, R., Waring, R., & Donkaewbua, S. (2008). Incidental vocabulary acquisition from reading, reading-while listening, and listening to stories. Reading in
a Foreign Language, 20(2), 136e163.
Bruton, A. (2011). Is CLIL so beneficial, or just selective? Re-evaluating some of the research. System, 39, 523e532.
Bruton, A. (2013). Some of the reasons why…and why not. System, 41(3), 587e597.
Celaya, M. L., & Ruiz de Zarobe, Y. (2010). First languages and age in CLIL and non-CLIL contexts. International CLIL Research Journal, 1, 3e5.
Cenoz, J. (2013). Discussion: Towards an educational perspective in CLIL language policy and pedagogical practice. International Journal of Bilingual Education
and Bilingualism, 16(3), 389e394.
Collins, L., & White, L. (2011). An intensive look at intensity and language learning. TESOL Quarterly, 45(1), 106e133.
Dimitrije
vic, N. (1969). Lexical availability. Heidelberg: Julius Gross Verlag.
Evans, V. (2007). A glossary of cognitive linguistics. Edinburgh: Edinburgh University Press.
Evnitskaya, N., & Morton, T. (2011). Knowledge construction, meaning-making and interaction in CLIL science classroom communities of practice. Language
and Education, 25(2), 109e127.
Faerch, C., Haastrup, K., & Phillipson, R. (1984). Learner language and language learning. Multilingual Matters.
Fern
andez-Fontecha, A. (2015). Motivation and vocabulary breath in CLIL and EFL contexts. different age, same time of exposure. Complutense Journal of
English Studies, 23, 79e96.
Ferreira, R., & Echeverría, M. S. (2010). Redes sem
anticas en el le
xico disponible de ingle
s L1 e ingle
s LE. Onoma
zein, 21, 133e153.
Germany, P., & Cartes, N. (2000). Le
xico disponible en ingle
s como segunda lengua en instruccio
n personalizada. Estudios Pedago
gicos, 26, 39e144.
Hakuta, K. (2000). How long does it take english learners to attain proficiency. University of California Linguistic Minority. Research Institute; University of
California. https://escholarship.org/uc/item/13w7m06g.
Hern
andez, N., Izura, C., & Ellis, A. W. (2006). Cognitive aspects of lexical availability. European Journal of Cognitive Psychology, 18, 734e755.
Hinger, B. (2006). The distribution of instructional time and its effect on group cohesion in the foreign language classroom: A comparison of intensive and
standard format courses. System, 34(1), 97e118.
Horst, M., & Collins, L. (2006). From weak to strong: How does their vocabulary grow? The Canadian Modern Language Review, 63(1), 83e106.
Jime
nez Catal

ndez Fontecha, A. (2015). Lexical phrases in compositions by CLIL and non- CLIL learners. In D. Marsh, M. L. Pe
an, R. M., & Ferna
rez Can
~ ado, & J.
Ra
ez Padilla (Eds.), Spanish CLIL in action: Voices from the classroom (pp. 82e98). Newcastle upon Tyne: Cambridge Scholars Publishing.
Jime
nez Catala
n, R. M., & Fitzpatrick, T. (2014). Frequency profiles of EFL learners’ lexical availability. In R. M. Jime
nez Catal
an (Ed.), Lexical availability in
English and Spanish as a second language (pp. 83e103). Dordrecht Heidelberg New York London: Springer. http://dx.doi.org/10.1007/978-94-007-7158-
1_11.
Jime
nez Catala
n, R. M., & Ojeda Alba, J. (2009a). Disponibilidad le
xica en ingle
s como lengua extranjera en dos tipos de instruccio
n. Lenguaje y Textos, 30,
167e176.
Jime
nez Catala
n, R. M., & Ojeda Alba, J. (2009b). Girls’ and Boys’ lexical availability in EFL. ITL International Journal of Applied Linguistics, 158, 57e76. http://dx.
doi.org/10.2143/ITL.158.0.2046920.
Jime
nez Catalan, R. M., Ruiz de Zarobe, Y., & Cenoz, J. (2006). Vocabulary profiles of English Foreign Language learners in English as a subject and as a vehicular
language. Viewz: Viena English Working Papers (Vol 15, pp. 23e27).
Karonen, S. (2003). Semantic organization of L2 learners' mental lexicon: Comparing pupils in mainstream education and CLIL classes. Unpublished MA Thesis.
University of Turku.
Laufer, B. (1998). The development of passive and active vocabulary in a second language: Same or different? Applied Linguistics, 19(2), 255e271.
Laufer, B., & Nation, P. (1995). Vocabulary size and use: Lexical richness in L2 production. Applied Linguistics, 307e322.
Laufer, B., & Nation, P. (1999). A vocabulary-size test of controlled productive ability. Language Testing, 16, 33e51.
Mashhadi, F., & Jamalifar, G. (2015). Second language vocabulary learning through visual and textual representation. Procedia-Social and Behavioral sciences,
192, 298e307.
Meara, P., & Fitzpatrick, T. (2000). Lex30: An improved method of assessing productive vocabulary in an L2. System, 28, 19e30.
Meara, P., Lightbown, P., & Halter, R. H. (1997). Classrooms as lexical environments. Language Teaching Research, 1, 28e46.
Milanovic, M., & Saville, N. (2015). English profile studies 5. english profile in practice. Cambridge: Cambridge University Press.
Milton, J. (2013). Measuring the contribution of vocabulary knowledge to proficiency in the four skills. EUROSLA Monograph Series, 2, 57e78.
Milton, J., & Fitzpatrick, T. (Eds.). (2014). Dimensions of vocabulary knowledge. New York: Palgrave Macmillan.
Moreno Fern
andez, F. (2007). Anglicismos en el le
xico disponible de los adolescentes hispanos de Chicago. In K. Potowski, & R. Cameron (Eds.), Spanish in
contact: Policy, social and linguistic inquiries (vol 22, pp. 41e58). Studies in Language and Society.
Moreno, S. (2009). Young learners’ L2 word association responses in two different learning contexts. In Y. Ruiz de Zarobe, & R. M. Jime
nez Catala
n (Eds.),
Content and language integrated learning. Evidence from research in Europe (pp. 93e111). Bristol: Multilingual Matters.
Morris, L., & Cobb, T. (2004). Vocabulary profiles as predictors of the academic performance of teaching english as a second language trainees. System, 32,
75e87.
Mun ~ oz, C. (2007). CLIL: Some thoughts on its psycholinguistic principles. RESLA, Volumen monogra
fico, 17e26.
~
Munoz, C. (2015). Time and timing in CLIL: A comparative approach to language gains. In M. Juan-Garau, & Salazar-Noguera (Eds.), Content-based language
learning in multilingual educational environments, educational linguistics (vol 23). http://dx.doi.org/10.1007/978-3-319-11496-5_6.
Nation, P. (2010). Learning vocabulary in another language. Cambridge: Cambridge University Press.
Nation, P., & Waring, R. (1997). Vocabulary size, text coverage and word lists. In N. Schmitt, & M. McCarthy (Eds.), Vocabulary: Description, acquisition and
pedagogy (pp. 6e19). New York: Cambridge University Press.
Nikula, T., Dalton-Puffer, C., & Llinares, A. (2013). CLIL classroom discourse. research from Europe. Journal of Immersion and Content-Based Language Edu-
cation, 1(1), 70e100.
Pellicer-S
anchez, A., & Schmitt, N. (2015). Incidental vocabulary acquisition from an authentic novel: Do things fall apart? Reading in a Foreign Language,
22(1), 31e55.
€
Pietila, P., & Merikivi, R. (2014). The impact of free-time reading on foreign language vocabulary development. Journal of Language Teaching and Research,
5(1), 28e36.
Pladevall-Ballester, E., & Valbona, A. (2016). CLIL in minimal input contexts: A longitudinal study of primary school learners' receptive skills. System, 58,
37e48.
Qian, D. (2008). From single words to passages: Contextual effects on predictive power of vocabulary measures for assessing reading performance. Language
Assessment Quarterly, 5(1), 1e19.

n, M.P. Agustín Llach / System 66 (2017) 87e99
R.M. Jim
enez Catala 99

Ruiz de Zarobe, Y. (2010). In C. Dalton-Puffer, T. Nikula, & U. Smit (Eds.), Language use and language learning in CLIL classrooms (pp. 191e213). Amsterdam/
Philadelphia: John Benjamins Publishing Company.
Samper Herna
ndez, N. (2002). Disponibilidad l
exica en alumnos de espan ~ ol como lengua extranjera. M
alaga: ASELE.
Schmitt, N. (2008). Review article: Instructed second language vocabulary learning. Language Teaching Research, 12(3), 329e363.
Schmitt, N. (2010). Researching vocabulary. a vocabulary research manual. Basingstoke, New York: Palgrave Macmillan.
Schmitt, N., Schmitt, D., & Clapham, C. (2001). Developing and exploring the behaviour of two new versions of the vocabulary levels test. Language Testing,
18(1), 55e88.
Scott, M. (2012). WordSmith tools version 6. Stroud: Lexical Analysis Software.
Seregely, E. M. (2008). A comparison of lexical learning in CLIL and traditional EFL classrooms. Wien: University of Wien, Magistra der Philosphie.
Serrano, R., & Mun ~ oz, C. (2007). Same hours, different time distribution: Any difference in EFL? System, 35(3), 305e321.
Simes, M. K. (2008). Room for improvement? A comparative study of Swedish learners' free written production in english in the foreign language classroom and in
immersion education. Karlstad: Faculty of Arts and Education, University Studies (PhD. Dissertation).
Spada, N., & Lightbown, P. (1989). Intensive ESL programmes in quebec in primary schools. TESL Canada Journal/Revue TESL du Canada, 7(1), 11e33.

n, L. K. (2004). Teaching in english or english teaching? On the effects of content and language integrated learning on Swedish learners' incidental vocabulary
Sylve
acquisition. Unpublished doctoral dissertation. Gothenburg: University of Gothenburg.
Urrutia, M. A. (2003). Redes sema
nticas en línea: Una tarea de acceso le
xico a partir de un estudio experimental. RLA, 41, 119e141.
V

arkuti, A. (2010). Linguistic benefits of the CLIL approach: Measuring linguistic competences. International CLIL Research Journal, 1(3), 67e79.
Whittaker, R., Llinares, A., & McCabe, A. (2011). Written discourse development in CLIL at secondary school. Language Teaching Research, 15(3), 343e362.
Special Issue: Content-based language Teaching.
Wray, A. (2002). Formulaic language and the lexicon. Cambridge: Cambridge University Press.
Wray, A. (2008). Formulaic language. Language Teaching Research, 46(3), 316e334.
Xantou, M. (2011). The impact of CLIL on L2 vocabulary development and content knowledge. English Teaching: Practice and Critique, 10(4), 116e126.