A Longitudinal Study on the Impact of CLIL on Affective Factors

Abstract

Content and Language Integrated Learning (CLIL) programmes are burgeoning in European school contexts due to the widespread belief that they help to significantly improve foreign language learning while content learning is not negatively affected. However, some of its purported benefits still need to be empirically confirmed. Thus, while this approach is widely believed to positively influence students’ affective stance, research studies are scant, and hardly any are longitudinal in nature. In an attempt to fill this gap, the present longitudinal study aims to analyse the impact of CLIL on different affective components. The participants are 304 secondary education students who were enrolled in CLIL and non-CLIL programmes. Contrary to expectations and the reviewed literature, the downward motivational trend observed in non-CLIL students in previous studies was not found. The results also indicate that CLIL does not help to sustain students’ motivation over time, but motivation to learn the subject matter is maintained in CLIL classes.

INTRODUCTION

European language policy recommends competency in the local and/or national language and two other European languages (Commission of the European Communities 1995). Against this backdrop, Content and Language Integrated Learning (CLIL), an umbrella term for a teaching approach in which content-based subjects are taught and learnt in a foreign language (FL), has been embraced all over Europe by stakeholders as a ‘way to transcend the perceived weakness of traditional FL (foreign language) teaching’ (Dalton-Puffer 2011: 185), including the communicative approaches in language learning which have failed to ‘motivate and produce a highly skilled plurilingual, pluricultural workforce’ (Coyle 2008: 99). Unlike in the case of EFL (English as a foreign language) approaches, the main goal of CLIL is to ‘produce competent and confident target language users, while at the same time teaching subject content’ (Mearns 2012: 176), consequently, the same importance is attributed to the content of the subject, and the language (Marsh 2002), which is normally a FL.

After approximately two decades since the widespread introduction of CLIL in European educational contexts, research has shown that this new approach improves FL learning (see Dalton-Puffer 2011 for an overview of the research on this topic; Lorenzo et al. 2010). In addition, CLIL is also associated with the development of skills such as problem solving, risk-taking, linguistic confidence, communication skills, vocabulary, self-expression, spontaneous talk, cultural awareness, and global citizenship (Coyle 2006). However, CLIL is not a panacea, and some downsides have also been identified in the literature. First, there seems to be some evidence (Otwinowska 2013) which shows that if CLIL is introduced too early and is not carefully implemented from a methodological standpoint, it may negatively affect children’s attitudes and motivation to learn. In line with this result, Muñoz (2015:10) reports that research conducted by Lorenzo et al. (2010) revealed that ‘older CLIL students benefit from CLIL more than younger CLIL students’ as far as their achievement in the target language is concerned. The reasons for the relevance of the students’ age in CLIL may be that the older students have more refined analytic acquisitional strategies for the L2 than the younger students, and they may have better developed L1 literacy skills which can harness literacy development in an L2 (Genesee 2014). Secondly, Seikkula-Leino (2007) has observed that there are more overachievers in the content subjects in the non-CLIL groups because learning in their mother tongue puts students in a better position to obtain maximum results and, conversely, learning in a FL may negatively influence the ability to learn the content (e.g. mathematics) among the CLIL students. Thirdly, Seikkula-Leino (2007) has also noted that CLIL students may have low self-concept (low self esteem) in FLs as they are frequently exposed to language which is above their current competence. Finally, CLIL students may feel overwhelmed by the task of learning content through a FL (Smit 2008), although the difficulty may also lie in the subject itself regardless of the language of instruction (Hellekjaer 2010).

The aim of this article is to study the interaction between CLIL and an increase in student motivation, one of the main purported benefits associated with CLIL. We analyse and compare the levels of motivation in CLIL and non-CLIL classes in five secondary schools in the Basque Autonomous Community (BAC) in Spain from a longitudinal perspective. In order to do so, we have organized this article as follows. First, we provide an overview of the research on motivation in CLIL followed by the characterization of the scales in motivation that are used in our study. Next, we state the principal features of the study and then present its main results. In the last section we provide some final conclusions.

Motivation in language learning in CLIL

It is difficult to generalize researchers’ findings regarding the beliefs certain groups hold about their own CLIL realities (Dalton-Puffer and Smit 2013: 549). This is because the contexts in which CLIL are implemented may vary widely, and CLIL may be realized differently depending on the socio-cultural settings and educational policies of the countries involved (Coyle 2008). However, researchers do seem to agree that one of the main benefits of CLIL is the high levels of motivation perceived among students: ‘one of the most powerful findings of CLIL groups centers on increased motivation in both learners and teachers’ (Coyle 2006: 11). Various reasons have been found to be responsible for this effect. First, unlike in EFL where language is the main focus of attention and content is not usually evaluated, in CLIL both content and the FL play crucial roles in assessment. Secondly, CLIL provides a cognitively challenging situation which is associated with an improved sense of achievement. Thirdly, students may be influenced by stakeholders’ belief that CLIL has positive effects on learning and teaching a FL while learning other subjects in the FL (Dafouz et al. 2007; Doiz et al. 2010; Lasagabaster and Sierra 2010; Hüttner et al. 2013—in Dalton-Puffer and Smit 2013). Fourthly, CLIL also tends to promote teacher motivation through collaboration between content and language teachers and offers cross-curricular opportunities leading to fruitful discussions on pedagogical issues and practices (Coyle 2006). Finally, it is characterized by ‘flexible non-prescriptive models which encourage context driven changes’ (Coyle 2006:12) and it provides teachers and students with a sense of ownership of their teaching practice and the learning process. As Coyle (2006: 12) nicely puts it, ‘motivated teachers “breed” motivated learners’ and motivated learners breed motivated teachers in turn.

The effect of CLIL on students’ motivation in secondary schools in the BAC, the context for this study, has been tackled in Sierra (2011), Lasagabaster (2011), and Doiz et al. (2014). These three studies confirmed the pedagogical benefits of CLIL students as compared with non-CLIL students with regard to their degree of motivation to learn English. In particular, Lasagabaster (2011) revealed that CLIL students were more motivated than non-CLIL students with respect to the degree of interest, instrumental motivation, attitudes towards learning English at school, and effort made. In Doiz et al. (2014), better means were also obtained among the CLIL students than the non-CLIL students in intrinsic motivation, instrumental orientation, and interest in FLs and foreign cultures. Yet, concerning the results of some of these studies, doubts on the efficacy of the CLIL approach in isolation may have arisen. This is because Rumlich (2014) has shown that prospective CLIL learners are more motivated and more competent language learners before entering a CLIL program. While this may be the case with the subjects in the present study as well, we intend to study the evolution of motivation within the CLIL and within the non-CLIL groups respectively (intra-group comparisons) from a longitudinal perspective rather than comparing CLIL and non-CLIL groups (inter-group comparisons).

More research on the relationship between CLIL and motivation is needed as acknowledged by Breidbach and Viebrock (2012: 11), who underscore that ‘Motivation is among the marginally examined concepts in German-speaking CLIL research’ in their review of CLIL research in the German context and in the light of the existence of some more recent critical studies on the positive effects of CLIL (see Breidbach and Viebrock, 2012, for a review of a number of studies on this issue). The need for more research is particularly remarkable in the case of longitudinal studies of motivation in CLIL. It is striking to see that little attention has been paid so far to the evolution of student motivation in CLIL programmes over time despite the fact that motivation may fluctuate over time. In fact, with the exception of the results in Tragant (2006), student motivation in traditional EFL classrooms tends to decrease over the years of language study and education (Chambers 1999; Davies and Brember 2001; Madrid 2002; Williams et al. 2002; Lorenzo et al. 2009; Fernández and Terrazas 2012), as students’ involvement habitually deteriorates with time (Littlejohn 2008). To the best of our knowledge, one of the few studies undertaken in CLIL contexts is the one by Papaja and Rojczyk (2013). These authors studied motivation from CLIL students’ perspective in a Polish high school and noted sustained high ratings of pleasure from learning English in CLIL during a three-year period regardless of the gender of the students. In addition, high levels of satisfaction with learning content in English were observed, although the satisfaction which reached its peak in the second year waned towards the end of the third year due to the pressure of the final high school examinations.

Hence, the question arises as to whether motivation may also wane once CLIL ceases to be a novelty: ‘[ … ] will attitudes, practices and outcomes be the same once it has lost its aura of innovation and become “normal practice”?’ (Dalton-Puffer and Smit 2013: 557). This article seeks to compare the levels of language motivation of students in CLIL and EFL (or non-CLIL) secondary students from a longitudinal point of view and to clarify whether the purported positive effects of CLIL on students’ initial motivation is sustained over time, or whether such impact wanes once the initial novelty of the approach wears off. In order to do so, we next introduce the affective scales which will be considered against our data.

Affective factors

While the importance of cognitive variables in FL learning is widely accepted, the importance of affective factors should not be downplayed as evidenced by the research conducted in the past decade. Traditionally, affective factors have dealt with anxiety and scales within the motivation domain such as intrinsic motivation, integrative and instrumental orientation, interest in FLs/cultures, and motivational strength. These scales have been successfully applied, among others, by Gardner (1985) and by Schmidt and Watanabe (2001) in studies of motivation in different contexts. Following Schmidt et al. (1996, in Schmidt and Watanabe 2001), we group these scales, with the exception of the integrative orientation scale that has been excluded for reasons that we will explain below, under the three overarching dimensions that make up Gardner’s definition of motivated learning behaviour (1985), namely, (i) affect (the attitudes towards learning a language), (ii) the desire to learn the language or goal orientation, and (iii) effort/expectancy. In other words, we follow Papaja and Rojczyk’s (2013: 243) characterization that ‘a highly motivated individual will want to learn the language, enjoy learning the language, and strive to learn the language’. A fourth dimension, namely parental encouragement (Gardner, 1985), is also included in the present study, for reasons that will be provided below. Connections between these scales and Dörnyei’s (2005, 2009) L2 Motivational self system are established when appropriate. In addition, students’ intrinsic motivation and motivational strength towards content learning in CLIL are also analysed, since subject (content) matter relevance has been pointed out above as an argument for the attractiveness of CLIL.

Affect: Intrinsic motivation and anxiety

Affect refers to the attitudes towards learning a language. It includes three measures primarily: intrinsic motivation, anxiety, and liking for challenge (Schmidt et al. 1996 in Schmidt and Watanabe 2001). Intrinsic motivation is a crucial motivational factor (Ushioda 2001) that refers to the enjoyment of the activity itself, in the present case, learning English. High levels of enjoyment of learning English will lead to more motivated students and as a result, language learning is enhanced. In contrast, anxiety is reversely correlated to intrinsic motivation (Schmidt and Watanabe 2001). That is to say, anxiety or ‘the lack of confidence in oneself as a learner, uneasiness, frustration, self-doubt, apprehension and tension which are specifically related to language learning situations’ (Arnold and Brown 1999; in Turula 2013: 256) has traditionally been argued to constitute an obstacle for language learning. In this vein, Naiman et al. (1975) showed that teachers believed that students’ anxiety level affected students’ FL learning success, and Horwitz (2001), Krashen (1981), and Rivers (1964), to mention just a few studies, believed that it may hamper learning and performance. The sources for anxiety in the language classroom are multiple, such as testing, communication apprehension, peer evaluation, stage fright, the fear of being laughed at, the learner’s personality, the instructor’s teaching styles, the learning context itself, the L1 skills, the process of learning a FL and, perhaps most importantly, the instructor’s support to the students (Marcos-Llinás and Juan-Garau 2009). Thus, it seems reasonable to ask teachers to promote positive affect and keep anxiety levels low (Arnold 1999; Dewaele 2013). However, the link between anxiety and language learning success depends on the understanding of the concept of anxiety and the situation in which it is researched (Gardner 1985). Moreover, there is mounting evidence which ‘challenges the long-held tenet in anxiety research that foreign language anxiety interferes with foreign language proficiency and achievement’ (Marcos-Llinás and Juan-Garau 2009: 104). Finally, students who value learning a language, that is to say, students who are motivated to learn a language, are more likely to perceive being challenged positively than students for whom learning a language responds to the need to meet the language requirement (Schmidt and Watanabe 2001). Thus liking of challenge is closely related to motivation.

Goal orientation: desire to learn a language

The category of goal orientation or the desire to learn an L2 taken in this article includes the measures of instrumental orientation and interest in FLs and cultures. We have excluded from it Gardner and Lambert’s (1972) integrative motivation, that is to say, the desire to identify with or become integrated in the L2 culture/community. While this kind of motivation may be appropriate for the Canadian context it was meant to describe when this concept was posited, it may not be applicable to other situations in which learners of English have no urge to identify with the L2 community or have little contact with L2 speakers (Lamb 2004; Dörnyei and Ushioda 2009). This is the case in many European countries, including our own where English is not spoken outside of the classroom. Moreover, the current role of English as a lingua franca, the development and acquisition of a variety of world Englishes (Coetzee-Van Rooy 2006) and consequently, the blurring of a clear-cut association of English with particular anglophone cultures (Lamb 2004) have rendered the integrative construct inappropriate for many contexts in which English language learning as a FL takes place (Dörnyei 2010).

Instrumental orientation

An instrumentally orientated student’s reason for learning a language is utilitarian. It is part of the students’ education, which will serve to advance in their profession or career. Instrumentality however, may have two different facets in Dörnyei’s L2 motivational self system (2010): a promotion or a prevention focus. The former fits within the Ideal L2 self, that is, the ideal L2 speaker one would like to become or the idealized image associated with being professionally successful. The latter is part of the Ought-to self as it embraces the properties one believes one ought to have to meet external expectations (e.g. parental expectations) and to avoid possible negative consequences, such as the fear of failing an exam or to disappointing one’s parents. The items included in our questionnaire deal with the promotion-focus of instrumentality alone.

Interest in FLs/cultures

The possibility of being in contact with people from cultures different from one’s own may be a motivating factor for FL learning. Students may be interested in learning FLs, including English, because knowing languages opens the possibility for interacting with more people and getting to know other cultures. Thus, learning languages facilitates access to a world that lies beyond the borders of the students’ country. This is clearly one of the outcomes of learning English, since English is an international language that ‘is most often used in intercultural encounters’ (Kormos and Csizér 2008: 330). In addition, in the current era of globalization, the media and the Internet are leading to the development of a global identity among young people ‘that gives them a sense of belonging to a worldwide culture and includes an awareness of the events, practices, styles and information that are part of the global culture’ (Arnett 2002: 777, in Dörnyei et al. 2006: 92). It is therefore becoming commonplace to find people all over the world who have developed a bicultural identity which is grounded in their local culture as well as in global culture (Arnett 2002). Global identity is an ‘international posture’ (Yashima 2002), which includes ‘interest in foreign or international affairs, willingness to go overseas to study or work, readiness to interact with intercultural partners … and a non-ethnocentric attitude toward different cultures’ (Kormos and Csizér 2008: 330).

Effort/expectancy

Unlike other skills in life, learning an L2 requires ‘intensive practice and increased effort’, hence it is crucial for learners to become autonomous learners who know how to study on their own and practice the L2 whenever the opportunity arises (Kormos and Csizér 2008: 350). Effort is the motivational strength or intensity the individual exerts on language learning. A highly motivated individual will do her best to learn the language. She will work hard to attain her goal despite the difficulties that may lie in the process. In addition, the term expectancy proposed by Schmidt et al. (1996, in Schmidt and Watanabe 2001) includes the component of effort but also interrelated factors such as self-confidence, positive thinking, and determination, all of which are crucial for motivation.

Parental encouragement

Parents’ influence on their children’s competence in the FL has been considered in numerous studies (see Lindgren and Muñoz 2013 for a review of the impact of factors such as ‘parents’ literacy level, parents’ involvement and attitudes towards the FL, and parents’ proficiency in the FL’ on school performance). In particular, Desforges and Abouchaar (2003: 86) have concluded that parental involvement that is specifically manifested in ‘at-home relationships’ and through the ‘modelling of aspirations’ is crucial for school outcomes because it helps children build a ‘pro-social, pro-learning self-concept and high educational aspirations’.

In a similar vein, Csizér and Kormos (2009) hypothesized that, since Dörnyei’s (2009) Ought-to L2 self dimension of his L2 motivational self system contains extrinsic motivational forces, motivation would be affected by parental encouragement. Their model revealed that for secondary education students the relationship between parental encouragement and the Ought-to L2 self was very strong, as the latter ‘is entirely socially constructed. Consequently students’ views of what attributes they should possess to meet the expectations of their environment are formed by the attitudes of their immediate learning environment’ (Csizér and Kormos 2009: 107). In another study of Hungarian minority students in the Ukraine, Henkel (2009) also found a statistically significant correlation between the Ought-to L2 self component and parental encouragement, and concluded that the students’ Ought-to L2 self reflects the Ought-to L2 self that the parents would like their children to achieve. The consideration of this scale in our study thus seems to be fully justified.

Research questions

This study aims to answer the following three research questions:

1. Does the CLIL approach help to maintain students’ motivation to learn English as a FL over time in comparison with the downward trends of non-CLIL students?

2. Does CLIL equally affect different affective factors?

3. Does CLIL also positively affect student’s motivation to learn the subject content over time?

The study

The participants

Out of the initial 393 participants, 304 eventually completed the questionnaires in the different phases. The students were enrolled in five schools in the BAC and were divided into two groups according to the school year they were enrolled in at the time of the study. The first group consisted of 158 first-year secondary school students (12–13-year-old students when the study began), while the second group was made up of 146 third-year secondary school students (14–15-year-old students at the onset of the study). In addition, each group was split into two sub-groups, the non-CLIL group and the CLIL group. Following the school policies, the students in the CLIL groups were selected by the schools from the first-grade pool of students and only the students with the highest level of English were admitted into the CLIL groups. The CLIL students in grade 3 were not selected, as all of them were invited to participate in the plurilingual experience. The subjects taught in English varied from school to school and included the following: social sciences, citizenship, arts, music, and the weekly tutorial. Since the participants in our study attended bilingual programmes in Basque and Spanish, English is their L3.

The participants were divided approximately equally across gender in both groups, although there were slightly more female students in both groups and grades than male students. There were also more students in the CLIL groups, especially in the third year, as displayed in Table 1.

Table 1:

Distribution and gender in both groups (SE = secondary education)

	1st year SE (per cent)		3rd year SE (per cent)
Gender	CLIL (per cent)	Non-CLIL (per cent)	CLIL (per cent)	Non-CLIL (per cent)
Female	59 (60.8)	33 (54.1)	54 (52.9)	24 (54.5)
Male	38 (39.2)	28 (45.9)	48 (47.1)	20 (45.5)
Total	97 (61.4)	61 (38.6)	102 (69.9)	44 (30.1)
	158 (100)		146 (100)

The instrument

The questionnaire was based on scales already used by Gardner (1985) and Schmidt and Watanabe (2001). Gardner (1985) developed his instrument in the Canadian context (an L2 context), but his scales have been successfully applied in different FL contexts. Since Schmidt and Watanabe (2001) specifically designed their scales for an FL context, these scales were suitable for our context. Other aspects of motivation were not included in the instrument because the aforementioned scales were considered to be the most suitable for our context; see our previous description of the scales in this paper for our reasoning behind the exclusion of integrative motivation in contexts such as the BAC where English is a FL. In contrast, anxiety and parental encouragement were included and hypothesized to play an important role in L2 motivation, as other scholars have proposed (Schmidt and Watanabe 2001; Kormos and Csizér 2008; Csizér and Kormos 2009).

After the first piloting of the instrument in its English version, the decision was made that the non-CLIL group would complete the questionnaire in Basque (the teaching language for all subjects except English and Spanish), as the students participating in the non-CLIL pilot group had some comprehension problems with the items in the English version. In contrast, the CLIL group filled it out easily, which is why it was decided that these students would complete it in English after some minimal changes were incorporated to the initial questionnaire. Next, the piloting of the Basque version for the EFL group was undertaken and some minimal changes made. The participants eventually completed the questionnaire in class at the end of the academic year in June.

The questionnaire was divided into three sections: the first one dealt with personal information such as gender, age, L1, or parental educational background (items 1–5). The second section (items 6–31) represented the bulk of the questionnaire and was focused on the scales under consideration in this article. The items were followed by a traditional five-point Likert scale which ranged from 1 = strongly agree to 5 = strongly disagree. Within the second section, seven items aimed at measuring the students’ motivation towards subject content learning were included. These items were preceded by a heading which explained that, in this particular section of the questionnaire, the participants had to focus on content learning in their CLIL classes. The third and last part of the questionnaire included three open questions in which the CLIL students were asked about the advantages and disadvantages of studying subjects in English, what they liked most and least in their CLIL classes and, finally, they were invited to make any additional comments about the CLIL programme. Since this last open part of the questionnaire tackled the CLIL approach, the non-CLIL group’s questionnaire did not include these questions. The statistical analysis of the close-ended items was carried out by means of the SPSS (IBM SPSS Statistics 21).

First-year students completed the questionnaire in three consecutive years: time 1 in the first year of secondary education, time 2 in the second year, and time 3 in the third year. Their older peers completed the survey in two consecutive years: time 1 in the third year of secondary education and time 2 in the fourth year. The latter were surveyed only twice (in grades 3 and 4) because afterwards some of them continued their studies through different options (vocational training or baccalaureate) and different schools, which would impede any follow-up study.

RESULTS

Since CLIL aims to integrate both language and content learning, this section will focus on the two parts of the questionnaire that deal with students’ attitudes and motivation towards learning English, and towards the CLIL classes. First of all, we addressed the items related to the English language and conducted factor analyses to reduce the number of variables in the questionnaire in order to obtain the underlying factors of the items included in the research instrument. An exploratory factor analysis was carried out because the scales were based on two different studies (Gardner 1985; Schmidt and Watanabe 2001) undertaken in two contexts very different to ours: the BAC is a bilingual community in which English represents a FL, and is considered an L3 to most. This is unlike Canada, where English is commonly present as an L1 or L2. Moreover, a new item, ‘I am learning English to understand films, videos, music, games, chat on the Internet’, was included that had not been used in previous research, but which we considered to be important for secondary education students at the present time. As a result of this, differences could emerge for the variables most likely to load onto each factor. Therefore, we could not take for granted the number of factors that we would encounter and which variables would load onto some of the factors. In fact, the three items originally encompassed in the instrumental orientation factor (Gardner 1985) were divided into two factors in our data in each survey phase: instrumental orientation and global integration (this latter factor consisted of Gardner’s third instrumental item and the new item mentioned above). The exploratory factor analysis also allowed us to have lower cut-offs of factor loadings.

Table 3:

Cronbach alpha coefficients for each scale (English class)

	Cronbach alpha
Scales	Time 1	Time 2	Time 3
Intrinsic motivation	.827	.837	.782
I really enjoy learning English
I enjoy my English class
I enjoy using English in the class
I don’t like learning English (reverse coded)
I often feel bored when I study for my English class (rev. coded)
Instrumental orientation	.787	.805	.871
Studying English is important for me because I’ll need it for my future studies
Studying English is important for me because it will be useful in getting a job
Global integration	.504	.487	.433
Studying English is important for me because people will respect me more if I speak English
I am learning English to understand films, videos, music, games, chat on the Internet
Interest in foreign languages/cultures	.701	.748	.661
Studying English is important for me because I can meet and talk with more people
I would like to learn different foreign languages
I enjoy meeting and talking with people from other countries and other cultures
Studying foreign languages is an important part of my education
Anxiety	.737	.714	.738
I feel nervous when I have to speak in English in my English class
I don’t worry about making mistakes when speaking in front of my English class (rev. coded)
I feel more tense and nervous in my English class than in my other classes
I always feel that the other students speak English better than I do
Motivational strength	.654	.664	.817
I work hard in my English class even when I don’t like what we are doing
Even when course materials are not interesting, I always finish my work
I put my best effort into learning English in my English language class
Parental encouragement	.581	.621	.649
My parents try to help me with my English
My parents think I should dedicate more time to English
My parents really encourage me to study English
My parents show considerable interest in anything to do with my English class at school
My parents have stressed the importance English will have for me when I leave school
Overall mean Cronbach Alpha	.684	.696	.707

Separate analyses were performed for each survey phase. This would allow us to find out how many factors there were by applying the varimax rotated factor matrix. The 25 items were gathered in seven factors or dimensions in the three survey phases. As Table 2 shows, the seven-factor solution encompassed easily interpretable clusters of variables that essentially matched those proposed in the original questionnaires (Gardner 1985; Schmidt and Watanabe 2001). The seven factors coincided in the three phases, which underscores the stability of the solution. By adding the seven factors in each phase, the total common variance or communality is obtained, which was 60.484 per cent of the total variance in time 1, 63.118 per cent in time 2, and 69.267 per cent in time 3.

Table 2:

Factor analysis of the items (English class)

Variables		Time 1							Time 2							Time 3
		1	2	3	4	5	6	7	1	2	3	4	5	6	7	1	2	3	4	5	6	7
Intrinsic motivation	I really enjoy learning English	.68								.63							.51
	I enjoy my English class	.81								.69							.67
	I enjoy using English in the class	.59								.55							.55
	I don’t like learning English (reverse coded)	.59								.56							.50
	I often feel bored when I study for my English class (rev. coded)	.72								.71							.72
Instrumental orientation	Studying English is important for me because I’ll need it for my future studies				.73								.83					.87
	Studying English is important for me because it will be useful in getting a job				.75								.77					.85
Global integration	Studying English is important for me because people will respect me more if I speak English							.71							.78							.81
	I am learning English to understand films, videos, music, games, chat on the Internet							.78							.66							.62
Interest in foreign languages/ cultures	Studying English is important for me because I can meet and talk with more people			.44					.65												.75
	I would like to learn different foreign languages			.75					.62												.51
	I enjoy meeting and talking with people from other countries and other cultures			.55					.84												.64
	Studying foreign languages is an important part of my education			.67					.68												.48
Anxiety	I feel nervous when I have to speak in English in my English class		.82								.81					.78
	I don’t worry about making mistakes when speaking in front of my English class (rev. coded)		.56								.74					.25
	I feel more tense and nervous in my English class than in my other classes		.80								.71					.80
	I always feel that the other students speak English better than I do		.69								.58					.84
Motivational strength	I work hard in my English class even when I don’t like what we are doing						.80							.76					.82
	Even when course materials are not interesting, I always finish my work						.76							.78					.79
	I put my best effort into learning English in my English language class						.52							.61					.75
Parental encouragement	My parents try to help me with my English					.48						.63								.71
	My parents think I should dedicate more time to English					.56						.50								.51
	My parents really encourage me to study English					.74						.70								.64
	My parents show considerable interest in anything to do with my English class at school					.61						.69								.70
	My parents have stressed the importance English will have for me when I leave school					.51						.58								.58

Factor 1 encompasses five items and is labelled Intrinsic motivation. The first four items were also included in the original cluster by Schmidt and Watanabe (2001), but the factor analyses also gathered together the item ‘I often feel bored when I study for my English class (reverse coded)’ in this cluster in the three phases. Although it was part of the Motivational strength factor in Schmidt and Watanabe’s study, our results cogently support its incorporation in the Intrinsic motivation factor. Factor 2 received salient loadings from two items which are linked to the instrumental value of studying English, which is why this factor is labelled Instrumental orientation (as was the case in Gardner’s original study, 1985). These two items focus on the importance of English for future studies and employment.

The two items loading onto factor 3 concern the importance of English as a tool to integrate into the global community (to be respected by others and understand films, games or chat on the Internet) and therefore was termed Global integration. As mentioned above, studies confirm that the role of English as lingua franca spreads the concept of integrativeness to a much wider spectrum (Lamb 2004; Coetzee-Van Rooy 2006) that goes well beyond the English-speaking countries and their speakers. This global trend highlights the growing gap between global English and other world languages (Dörnyei and Csizér 2002; Dörnyei et al. 2006), which led us to choose Global integration as the label.

Factor 4 shows salient loadings from four items which have to do with the students’ Interest in FLs and cultures, as they revolve around their appreciation of FLs and their wish to meet people from other countries. Factor 5 received salient loadings from the five items connected to students’ degree of anxiety while in the English class, a factor that will consequently be referred to as Anxiety. The three items loading onto factor 6 relate to students’ Motivational strength, whereas the last factor has to do with the perceived support provided by parents when it comes to studying the English language, which is why this factor is labelled Parental encouragement, as it is based on the same scale proposed by Gardner (1985). Parents are not physically present at schools, but their figures pervade the whole educational process and, since we are focusing our study on secondary education, this scale could be considered essential.

The information about the internal consistency of the items comprised in each of the seven factors and in each phase can be observed in Table 3. The scales display good reliability, as confirmed by the mean coefficient of the scales across the three time phases: 0.684 in time 1, 0.696 in time 2, and 0.707 in time 3, which is not too high but can still be regarded as acceptable for short scales ranging from two to five items (Dörnyei et al. 2006). It has to be pointed out that the lower Cronbach alpha values obtained in the Global integration category indicate that this scale would have benefited from a larger number of items.

Once the underlying factors were identified, we gathered the descriptive statistics of both CLIL and non-CLIL participants in the two age groups and in the three phases of the survey, so that the data could be compared across time by means of repeated measures analysis of variance (ANOVA) analyses (see Tables 4 and 5).

Table 4:

Descriptive information about each scale and ANOVA statistics across time (English class): younger students

CLIL students	School year	Mean	SD	F	Sequencea	Non-CLIL Students	School year	Mean	SD	F	Sequencea
Intrinsic motivation	1st year	2.30	0.63	9.671***	1 > 2 & 3	Intrinsic motivation	1st year	3.14	0.86	1.575	ns
	2nd year	2.56	0.66				2nd year	2.92	0.76
	3rd year	2.70	0.67				3rd year	2.97	0.89
Instrumental orientation	1st year	1.22	0.36	7.653**	1 > 2 & 3	Instrumental orientation	1st year	1.50	0.73	0.009	ns
	2nd year	1.44	0.56				2nd year	1.51	0.66
	3rd year	1.53	0.63				3rd year	1.50	0.62
Global integration	1st year	3.22	0.91	2.156	ns	Global integration	1st year	3.25	0.82	1.000	ns
	2nd year	3.17	0.72				2nd year	3.10	0.91
	3rd year	3.00	0.86				3rd year	3.01	0.80
Interest in foreign languages/cultures	1st year	2.00	0.56	2.249	ns	Interest in foreign languages/cultures	1st year	2.11	0.49	0.137	ns
	2nd year	2.17	0.59				2nd year	2.12	0.63
	3rd year	2.09	0.63				3rd year	2.16	0.67
Anxiety	1st year	3.27	0.75	1.664	ns	Anxiety	1st year	3.63	0.70	7.144**	3 > 1
	2nd year	3.41	0.63				2nd year	3.25	0.98
	3rd year	3.35	0.77				3rd year	3.05	1.01
Motivational strength	1st year	2.21	0.66	6.962**	1 > 2 & 3	Motivational strength	1st year	2.34	0.78	2.688	ns
	2nd year	2.51	0.60				2nd year	2.37	0.76
	3rd year	2.51	0.70				3rd year	2.67	1.09
Parental encouragement	1st year	2.45	0.67	3.519*	1 > 2	Parental encouragement	1st year	2.73	0.76	3.151	ns
	2nd year	2.68	0.68				2nd year	2.41	0.62
	3rd year	2.58	0.63				3rd year	2.54	0.65

^aPost hoc Bonferroni comparison. Numbers refer to the phases of the survey. ‘ns’ indicates non-significant differences between values, while ‘>‘ denotes significant differences.

*p < .05; **p < .01; ***p < .001.

Table 5:

Descriptive information about each scale and ANOVA statistics across time (English class): older students

CLIL students	School year	Mean	SD	F	Sequencea	Non-CLIL students	School year	Mean	SD	F	Sequencea
Intrinsic motivation	3rd year	2.49	0.73	0.346	ns	Intrinsic motivation	3rd year	2.88	0.63	0.632	ns
	4th year	2.54	0.80				4th year	2.96	0.74
Instrumental orientation	3rd year	1.59	0.63	0.041	ns	Instrumental orientation	3rd year	1.70	0.69	0.315	ns
	4th year	1.61	0.80				4th year	1.65	0.63
Global Integration	3rd year	3.00	0.78	0.002	ns	Global integration	3rd year	3.53	0.85	0.553	ns
	4th year	3.00	0.92				4th year	3.43	0.72
Interest in foreign languages/cultures	3rd year	2.01	0.65	4.498*	2 > 1	Interest in foreign languages/cultures	3rd year	2.43	0.68	0.003	ns
	4th year	1.88	0.64				4th year	2.42	0.87
Anxiety	3rd year	3.34	0.86	9.173**	2 > 1	Anxiety	3rd year	3.18	0.86	0.637	ns
	4th year	3.13	0.80				4th year	3.14	0.88
Motivational strength	3rd year	2.56	0.71	1.149	ns	Motivational strength	3rd year	2.30	0.66	2.067	ns
	4th year	2.47	0.73				4th year	2.47	0.65
Parental encouragement	3rd year	2.54	0.68	1.691	ns	Parental encouragement	3rd year	2.51	0.56	1.246	ns
	4th year	2.46	0.63				4th year	2.40	0.62

^aPost hoc Bonferroni comparison. Numbers refer to the phases of the survey. ‘ns’ indicates non-significant differences between values, while ‘ > ’ denotes significant differences.

*p < .05; **p < .01; ***p < .001.

The analysis of how the different affective factors changed over the three-year period in the case of the younger students and the two-year period in the case of the older students indicates that the non-CLIL students suffered hardly any significant variation in both age groups. As a matter of fact, significant differences were only found among the younger students in the anxiety scale, since they appeared less anxious in grade 1 (M = 3.63) than in grade 3 (M = 3.05). In the case of the older students, no significant difference was detected across time in any of the seven scales. Therefore, and contrary to expectations, no significant decline in non-CLIL students’ motivation was found.

The younger CLIL students did undergo significant variations in four of the scales. Thus, they were more intrinsically motivated, more instrumentally oriented and showed greater motivational strength in grade 1 than in grades 2 and 3. Similarly, their perceived parental encouragement was higher in grade 1 than in grade 2, although no differences were found in grade 3. Analysis of the older CLIL students’ means revealed that the only significant differences in the two-year span were found in the interest in FLs and cultures and the anxiety scales: in grade 4 students were more interested in everything foreign and they also felt more anxious when speaking English. These changes may be the result of the age factor, as we will discuss later in our conclusions.

Since we intended to examine the effect of CLIL both on students’ motivation to learn the FL and on the CLIL classes themselves, their attitudes towards this approach were also considered. Once again, we first performed a factor analysis to shed light on the underlying factors of the part of the questionnaire focused on the CLIL context. Table 6 shows two factors in the three phases of the study, both factors coinciding with the intrinsic motivation and motivational strength categories described above, and the items included in each factor being the same. The common variance or communality was 62.801 per cent in time 1, 62.991 per cent in time 2, and 59.728 per cent in time 3. Table 7 shows that the scales display reasonably good reliability coefficients in each of the three survey phases, the mean being 0.684 in time 1, 0.698 in time 2, and 0.690 in time 3.

Table 6:

Factor analysis of the items (CLIL classes)

Scales		Time 1		Time 2		Time 3
		1	2	1	2	1	2
Intrinsic motivation	I like learning the subjects in English	.78		.88		.71
	I enjoy these classes	.81		.84		.83
	I enjoy using English in these classes	.82		.87		.78
	I often feel bored when I study for my classes taught in English (rev. coded)	.63		.34		.50
Motivational strength	I work harder in these classes than in others even when I don’t like what we are doing		.77		.91		.53
	Even when course materials are not interesting, I always finish my work		.59		.34		.83
	I put my best effort into learning the subjects taught in English		.56		.60		.86

Table 7:

Cronbach alpha coefficients for each scale (CLIL classes)

	Cronbach alpha
Scales	Time 1	Time 2	Time 3
Intrinsic motivation	.797	.779	.732
I like learning the subjects in English
I enjoy these classes
I enjoy using English in these classes
I often feel bored when I study for my classes taught in English (rev. coded)
Motivational strength	.572	.618	.648
I work harder in these classes than in others even when I don’t like what we are doing
Even when course materials are not interesting, I always finish my work
I put my best effort into learning the subjects taught in English

As can be seen in Tables 8 and 9, there were no significant differences across time among both the younger and the older participants, the means even being very similar in the two time phases (T1 and T2) they share. The fact that the younger students were selected, whereas the older ones were not, did not have any significant impact on either their intrinsic motivation or their motivational strength.

Table 8:

Descriptive information about each scale and ANOVA statistics across time (CLIL classes): younger students

CLIL students	School year	Mean	SD	F	Sequencea
Intrinsic motivation	1st year	2.51	0.63	2.242	ns
	2nd year	2.65	0.69
	3rd year	2.70	0.60
Motivational strength	1st year	2.56	0.66	1.389	ns
	2nd year	2.71	0.65
	3rd year	2.63	0.59

^aPost hoc Bonferroni comparison. “ns” indicates non-significant differences between values.

Table 9:

Descriptive information about each scale and ANOVA statistics across time (CLIL classes): older students

CLIL students	School year	Mean	SD	F	Sequencea
Intrinsic motivation	3rd year	2.49	0.65	2.492	ns
	4th year	2.61	0.69
Motivational strength	3rd year	2.53	0.64	3.379	ns
	4th year	2.68	0.65

^aPost hoc Bonferroni comparison. “ns” indicates non-significant differences between values.

CONCLUSIONS

In this article we intended to analyse CLIL and non-CLIL students’ motivation over time, as longitudinal studies in CLIL (particularly studies concerning motivation) are scant. One of the most striking results found has to do with the fact that the non-CLIL students’ motivation to learn English does not decline among younger students in the three-year period or among the older students in the two-year period under scrutiny in this study. Thus, counter to the results obtained in previous studies undertaken in diverse contexts which show a downward tendency in students’ motivation to learn FLs as they move to higher grades (Chambers 1999; Davies and Brember 2001; Madrid 2002; Williams et al. 2002; Dörnyei et al. 2006; Lorenzo et al. 2009; Fernández and Terrazas 2012), our study reveals that motivation is maintained over time in non-CLIL classes.

Secondly, and contrary to our expectations, the means of the non-CLIL students were in fact very similar to those of their CLIL counterparts in the third phase of the study (see Table 4). This could be attributed to the fact that, as predicted by Dörnyei et al. (2006), the hegemonic position of English leads students to assign an enormous symbolic value to this language, to the point that non-CLIL students are currently willing to make the effort to learn a language that opens up a great deal of different career opportunities. In fact, the instrumental-orientation means of both the CLIL and the non-CLIL groups are almost identical in time 3 among the younger participants (M = 1.53 and M = 1.50, respectively) and in time 2 among the older peers (M = 1.61 and M = 1.65, respectively). They are the lowest (which means higher motivation in our Likert scale) when all the scales are considered. Since the two items included in the instrumentality scale have to do with the promotion focus (Dörnyei 2010), it could be concluded that, as a result of students’ widespread awareness of the role that English may play in their future careers, students’ ideal L2 self is currently very strong irrespective of the approach (CLIL versus non-CLIL).

It would be interesting to carry out similar longitudinal studies in other contexts because, if similar results were obtained, it would mean that the hegemonic role of English as the current lingua franca helps to sustain students’ motivation irrespective of the type of approach involved in its teaching. Whereas students’ motivation towards other FLs decreases over time, English may have become the exception due its role as world language (Dörnyei et al. 2006). Additionally, this seems to apply to all the affective dimensions (intrinsic motivation, instrumental orientation, cultural understanding, interest in FLs and cultures, anxiety, motivational strength, and parental encouragement) under scrutiny in this article, as both age groups in non-CLIL programmes showed no significant changes throughout our study. Moreover, the status of English as the main lingua franca throughout the world and the subsequent globalization of culture have brought about the need to acknowledge the presence of a new scale in motivation, namely, global integration. The results for this scale are maintained in both groups in time, highlighting its importance across all our students regardless of the group they belonged in.

Conversely, it was expected that the CLIL approach would help to sustain (and even strengthen) students’ motivation, but this was not the case. Once again, and contrary to our expectations and what is stated in the literature, CLIL did not help to maintain motivation and there was a motivational decline over time in some of the affective dimensions of the younger students. Thus, CLIL students were more intrinsically motivated, more instrumentally oriented, and showed a higher motivational strength in the first year than in the subsequent academic years, but it seems that the initial high motivation starts to wane once CLIL is not a novelty anymore and becomes normal practice (Dalton-Puffer and Smit 2013). However, this does not affect both age groups in the same way. In fact, the younger students who were selected to take part in the programme are the ones who became less enthusiastic as time went by, whereas the positive effect lingered in the case of the older students who were not selected to participate in the CLIL experience. The latter remained more intrinsically motivated, more instrumentally oriented, and showed greater global integration and interest in FLs and cultures. Thus the confluence of age and being non-selected seems to play a role in maintaining their motivational stance, although older students also seem to become more anxious when it comes to speaking English in their CLIL classes.

These results may be closely linked to the difficulty of the subjects delivered in CLIL, as subjects become more complex and cognitively demanding in later grades. Therefore, we conclude that CLIL does not equally affect the different dimensions under study (our second research question). However, higher levels of anxiety do not necessarily have negative consequences for language learning. In fact, in their study Marcos-Llinás and Juan-Garau (2009) revealed that greater anxiety does not necessarily result in poor course achievement. In fact, there are two kinds of anxiety (Alpert and Haber 1960; see Marcos-Llinás and Juan-Garau 2009 for more references): debilitative and facilitative. The former impedes language learning and ‘learners who suffer from debilitative anxiety may feel fear or insecurity and even suffer from poor performance and withdrawal from the foreign class (Gardner 1985; Steinberg and Horwitz 1986; MacIntyre and Gardner 1991)’ (Marcos-Llinás and Juan-Garau 2009: 96). Facilitative anxiety, on the other hand, is like ‘an increase in drive level’ which improves performance and may be conducive to language learning because it ‘motivates students to learn the foreign language and to perform better overall’ (Marcos-Llinás and Juan-Garau 2009: 104). In our study, we did not take into account how anxiety affects students’ academic performance or their learning achievement, an issue that should be addressed in future research in CLIL settings.

CLIL has sometimes been accused of being an elitist approach (Bruton 2011), but our results indicate that, as far as affective factors are concerned, its effects are more positive in the case of non-elitist (non-selected) groups of students—in our case, the older group. The selection criterion did not have any significant impact on either students’ intrinsic motivation or their motivational strength and, therefore, it can be concluded that the selection criterion’s alleged influence wanes in CLIL. In addition, CLIL helps to sustain students’ positive motivation over time when it comes to learning the subject content (our third research question), as their means are closer to the point of the scale that shows a positive stance (see Tables 8 and 9) and no significant decline is observed.

To sum up, and bearing in mind the specificities of the context under scrutiny, our results can be summarized in three main conclusions: (i) the relationship between English and motivation in non-CLIL contexts is not as negative as previous studies confirm; (ii) CLIL does not have long-term positive effects on students’ motivation towards English language learning; (iii) motivation to learn the subject matter is maintained in CLIL classes. However, a word of caution should be expressed as our results need to be borne out in other contexts. Likewise, future research studies could be complemented by incorporating a qualitative perspective that would help to delve into the different motivational dimensions considered in this article.

SUPPLEMENTARY DATA

Supplementary material is available at Applied Linguistics online.

ACKNOWLEDGEMENTS

The results presented in this article are part of the following research projects: FFI2012-34214 (Spanish Ministry of Economy and Competitiveness), IT311-10 (Basque Government), and UFI11/06 (UPV/EHU). We would also like to thank the school management, the staff, and the students who participated in the study.