Students’ self-regulatory processes in content and language integrated learning: a vignette-based microanalytic study

ABSTRACT

Despite the global adoption of Content and Language Integrated Learning (CLIL) and extensive research on the effectiveness (product-oriented) and classroom interactions (process-oriented) of CLIL, existing studies have overlooked the motivational beliefs and metacognitive strategies of CLIL students. Hence, there is a need for a participant-oriented approach to better understand students’ perspectives and enhance CLIL’s success. Drawing on the social cognitive model of self-regulated learning (SRL), this study aimed to explore how students plan, monitor, and evaluate their learning in CLIL contexts by employing a novel vignette-based microanalytic assessment approach to empirically capture students’ SRL processes. Participants were 108 junior secondary school students from three schools in Hong Kong. Data were analysed using structuring qualitative content analysis. Findings revealed that students employed diverse strategies to manage the demands of language and content learning, with variations in self-efficacy beliefs influencing their approaches. Furthermore, some learners faced challenges in maintaining effective self-regulation. This study argues that improving self-regulatory processes and enhancing CLIL’s success requires prioritising assessment, considering the state of ‘others’, and fostering deeper learning beyond CLIL. This study contributes to the empirical, methodological, and pedagogical understanding of students’ SRL in CLIL settings and informs future research on student-focused aspects in CLIL.

KEYWORDS:

• Content and language integrated learning
• metacognitive strategies
• motivational beliefs
• self-regulatory processes
• English medium instruction
• bilingual education

Introduction

Content and Language Integrated Learning (CLIL) has been rapidly expanding worldwide, promoting the simultaneous learning of language and content subjects. Defined as ‘any pedagogical approach that integrates the instruction of content and second/foreign languages’ (Linares and Morton 2017), CLIL emphasises using language as a window to knowledge and ensuring that content learning helps students understand and acquire language (Llinares and Evnitskaya 2021). Originating in Europe to foster multilingualism (Cenoz, Genesee, and Gorter 2014), it has been widely adopted internationally, especially in non-European countries, due to its perceived dual outcomes of acquiring content knowledge and developing second language (L2) proficiency.

Initially adopting a product-oriented approach, CLIL research predominantly evaluated CLIL’s effectiveness in terms of language and content learning outcomes but yielded inconclusive results (Graham et al. 2018; Lo and Lo 2014). Subsequently, researchers adopted process-oriented approaches by exploring classroom interactions to examine learning processes (e.g. An, Macaro, and Childs 2021; Lo and Macaro 2015; Pun and Macaro 2019). However, these approaches often overlook students’ perspectives, which can provide valuable insights for enhancing CLIL teaching and learning effectiveness and improving overall educational outcomes (An and Thomas 2021). Therefore, scholars have called for understanding students’ perspectives to enhance their ownership of the learning process (Buckingham, Fernández Álvarez, and Halbach 2023; Coyle 2013; Fung and Lo 2023; Lo 2024). Lo (2024) advocated a participant-oriented (or person-oriented) approach to CLIL research, referring to examining CLIL from the students’ perspective including individual difference variables such as attitudes, beliefs, motivation, and learning strategies. Further, most existing studies are confined to the classroom context, although class time is limited and most learning occurs outside the classroom (Zhou and Thomas 2023). Therefore, it is essential to place students at the centre of research by investigating their goals, expected outcomes, and strategic planning and monitoring their progress through self-regulated learning (SRL).

Self-regulation refers to the self-generated thoughts, feelings, and actions that individuals use to achieve their learning goals (Zimmerman and Moylan 2009). It involves metacognitive processes, such as planning, monitoring, and evaluating progress, along with motivational factors, including intrinsic interest, outcome expectations, self-efficacy beliefs, and goal orientations. These elements interact dynamically within a cyclical model of forethought, performance, and self-reflection, enabling students to continually adapt and improve their learning processes. While self-regulation has been widely studied in both content and L2 learning, scrutinising it in CLIL, which is an amalgamation of content and language, has been lacking. Understanding learners’ thoughts, feelings, and behaviours during the forethought, performance, and self-reflection phases can enhance the CLIL teaching, learning, and assessment processes.

Drawing on Zimmerman and Moylan’s (2009) social cognitive SRL model, this study aimed to examine CLIL students’ motivation constructs (intrinsic value and outcome expectations) and metacognition (knowledge, awareness, and regulation of thinking) in their learning processes. This study seeks to address gaps in previous research by focusing on students’ motivational beliefs and metacognitive strategies, providing a more comprehensive understanding of their SRL in CLIL contexts. This study makes empirical, methodological, and pedagogical contributions to CLIL, laying a robust foundation for advancing CLIL research and providing evidence-based guidance to CLIL teachers and policymakers.

Bilingual education in Hong Kong

With the increasing interconnectedness of content and language learning worldwide, the use of various terms has caused confusion. CLIL is an umbrella term for programmes that integrate L2 and content, such as immersion- and content-based instruction (Cenoz, Genesee, and Gorter 2014; Linares and Morton 2017). Some researchers (e.g. Dalton-Puffer et al. 2014; Lasagabaster and Sierra 2010) have considered this substitution problematic. As no fixed models dictate how CLIL should evolve, its implementation varies significantly (Coyle 2013). Therefore, it is important to describe how bilingual education is operationalised in Hong Kong secondary education, which is the context of this study, and the terminology.

Compared with primary and tertiary education, Hong Kong secondary education has seen notable shifts in Medium of Instruction (MoI) practices over the decades, influenced by political transitions and educational reforms. Before 1997, the colonial government allowed secondary schools to choose their MoI, leading to a preference for English as a Medium of Instruction (EMI) schools. Post-handover, the Special Administrative Region government mandated Chinese as the MoI from primary 1 to secondary 3 (years 1–9) in 1998, facing strong opposition for depriving students of exposure to English. Consequently, 114 schools that met the specific criteria were allowed to remain EMI schools. Chinese as a Medium of Instruction (CMI) schools could offer up to 25% of lessons in English and choose their MoI from secondary 4 (year 10) onwards. From 2010–2011, a ‘fine-tuning’ policy eliminated the bifurcation of schools into CMI and EMI and granted more autonomy to secondary schools, resulting in diverse bilingual education models (Lo and Lin 2019): mostly EMI (high-ranking, English for most subjects), mostly CMI (lower proficiency, Chinese for most subjects with some English activities), MoI switching (Chinese at junior levels, English at senior levels), EMI by class (EMI classes for diverse student groups, such as non-Chinese-speaking students), and EMI by subject (specific subjects in English, usually mathematics and science).

Research on bilingual programmes at the secondary level in Hong Kong has used several terms, predominantly CLIL and EMI. This study adopted ‘CLIL’ as the overarching term, except when referring to specific literature for the following reasons. First, ‘integrating language learning and content learning’ is key focus of the ongoing renewal of the school curriculum for primary and secondary schools in Hong Kong (Curriculum Development Council 2017, 7), emphasising a dual focus on language and content. Second, although EMI aims to deliver the same benefits as CLIL, it has not been explicitly defined to reflect these commonalities (Rose et al. 2023). The defining characteristic of CLIL, which integrates content and language learning, is sufficiently broad to encompass the various models in Hong Kong (Lo and Lin 2019).

The waves of CLIL research

The first and second waves: product and process-oriented research

CLIL research can be broadly categorised as either product-oriented or process-oriented (Dalton-Puffer 2011; Lo 2020). Product-oriented research indicates that teaching subjects in an L2 setting positively affects the development of some L2 language skills (e.g. Lo and Murphy 2010; San Isidro and Lasagabaster 2019; Vraciu and Marsol 2023; Xanthou 2011) and does not harm content learning for high-performing or self-selected students but may do so for lower academic performers (e.g. Fernández-Sanjurjo, Fernández-Costales, and Arias Blanco 2019; Fung and Yip 2014; Piesche et al. 2016). San Isidro and Lasagabaster (2019) found that CLIL students in Spain showed greater progress in L2 English without negatively affecting content learning. In Cyprus, Xanthou (2011) observed significant improvement in L2 vocabulary among CLIL students, while both CLIL and non-CLIL students showed significant content knowledge gains. Lo and Lo's (2014) review revealed that students in Hong Kong were more proficient in L2 and performed better in affective variables (i.e. self-concept, motivation, and learning strategies); however, their comprehension of content subjects deteriorated. Fung and Yip (2014) compared the academic achievements and motivations of students in physics modules delivered through EMI and CMI and found that both low- and high-achieving students performed better when taught in L1 Chinese than L2 English.

Given these inconclusive findings, researchers have called for process-oriented research that examines the actual teaching and learning processes in CLIL lessons. An, Macaro, and Childs (2021) observed classroom interactions in EMI high schools in China and found that native English-speaking teachers did not enhance student participation compared with non-native English-speaking teachers, and classroom interactions remained teacher-dominated with minimal student participation. CLIL classrooms in Hong Kong have also been found to be teacher-centred with limited student interaction (Lo and Macaro 2015; Pun and Macaro 2019). Lo and Macaro (2015) found teacher–student interactions in both early and subsequent CLIL settings to be predominantly teacher-dominated. Pun and Macaro (2019) observed that teachers in subsequent EMI schools used higher-order questions and were more interactive when L1 was used more often. Given the monological nature of CLIL classrooms and the limited student interactions observed, Lo (2020) argued that ‘process-oriented’ research raises concerns about CLIL teacher preparation and professional development, specifically regarding how teachers are prepared, both psychologically and pedagogically, for CLIL teaching. Bruton (2013) asserted that students’ attitudes and preferences should be impartially surveyed. Although teachers play a crucial role in CLIL, other key actors, especially students, should be given equal importance.

The third wave: participant-oriented research

Investigating students’ perspectives for improving EMI/CLIL teaching and learning has recently gained importance (An and Thomas 2021; Coyle 2013; Fung and Lo 2023; Lo 2024; Otwinowska and Foryś 2017). This approach complements process-oriented classroom interaction studies, as global CLIL practice has shifted towards more student-centred approaches (Dalton-Puffer, Hüttner, and Llinares 2022). An and Thomas (2021) highlighted that students’ beliefs about interactions could explain their classroom observations. They found that while students recognised the benefits of interaction in both science and language learning, they were less inclined to engage extensively because of linguistic challenges and a preference for teacher-led instruction. Discrepancies between the students’ broader beliefs and their immediate classroom actions were noted, with many students feeling that the interactions took valuable time away from learning new content.

Coyle (2013) demonstrated that learners could perceive the need for changes in pedagogic practices through their CLIL experiences and contended that the quality and nature of learning experiences must be understood from learners’ perspectives to ensure that secondary school students succeed in their language learning endeavours. Fung and Lo (2023) also emphasised the importance of considering learners’ perspectives in teacher-dominated EMI classrooms by examining students’ strategy use when comprehending teacher input.

Building on studies focusing on students, Lo (2024) called for a participant-oriented approach, possibly representing the third wave of CLIL research, that aims at understanding students’ perceived strengths, challenges, and opportunities. As CLIL/EMI lessons have been found to be teacher-centred, students need to actively employ multiple self-regulated strategies, often outside the classroom (Zhou and Rose 2021; Zhou and Thomas 2023). Therefore, holistically understanding CLIL students’ self-regulatory practices, including in classroom and autonomous settings, is crucial.

Importance of self-regulation in CLIL settings

Zimmerman and Moylan’s (2009) self-regulation model, grounded in Bandura’s (1986) triadic social cognitive theory, has garnered significant empirical support (Panadero 2017). This theory posits that learning occurs through dynamic and reciprocal interactions among individuals, the environment, and their behaviours within a social context (Bandura 1986). Zimmerman and Moylan (2009) proposed a social cognitive SRL model that focused on the intersection of motivational constructs, including intrinsic value and outcome expectations, and metacognitive processes, such as strategy use and self-monitoring. It involves self-generated thoughts, feelings, and actions aimed at achieving educational goals and comprises three cyclical phases: forethought, performance, and self-reflection. The forethought phase involves goal setting and strategic planning driven by self-efficacy, outcome expectations, task interest, and goal orientation. The performance phase focuses on self-control methods, such as self-instruction, imagery, and time management, along with self-observation techniques. The self-reflection phase includes self-judgement and self-reaction, completing the cycle and influencing future learning efforts.

Recent studies have underscored the role of self-regulation in enhancing the effectiveness of learning environments in CLIL settings (Hu and Gao 2020; Jaekel 2020; Lo 2024; Ruiz de Zarobe and Smala 2020). Jaekel (2020) examined language learning strategies (LLS) and their impact on language proficiency among year 9 students in Germany by comparing CLIL and EFL classes. No significant differences in LLS use were found between the two groups. However, LLS use was negatively associated with language proficiency whereas self-efficacy strongly predicted higher language proficiency. This suggests that enhancing self-efficacy may benefit language learners more than focusing solely on strategy use.

To effectively navigate the demands of content and language acquisition, students in CLIL programmes must develop robust SRL strategies. Salili and Lai (2003) showed that CMI students in Hong Kong used more strategies than EMI students. Low-ability students employed fewer strategies; exhibited lower self-efficacy, positive attitudes, and motivation; and experienced higher anxiety. Hu and Gao (2020) revealed that high-achieving bilingual first-year secondary students effectively used resources, such as dictionaries and notes, while underachievers struggled owing to their lower English proficiency and lack of strategic resource use.

These studies suggest that enhancing self-regulatory skills in CLIL is crucial. However, studies on how self-efficacy and (metacognitive processes) intersect in managing the complex dual demands of linguistic and cognitive tasks are limited. A comprehensive grasp of all three cyclical phases elucidates language learners’ dynamic self-regulatory processes. Moreover, fostering self-regulation skills in CLIL contexts supports students’ transition from secondary to tertiary education, ensuring that they are better prepared for the challenges of EMI in higher education (Evans and Morrison 2011; Zhou and Rose 2021; Zhou and Thomas 2023). Drawing on Zimmerman and Moylan’s (2009) three-phase social cognitive model, this study aimed to answer the following research questions (RQs):

• RQ1: How do students plan and strategise their learning in CLIL, and how do their self-efficacy beliefs affect their expectations and outcomes? (forethought)

• RQ2 What strategies do students use during learning and when facing challenges? (performance)

• RQ3: How do students use reflection for successful and unsuccessful learning in CLIL, and how does it impact their future learning? (reflection)

Methods

Research context and participants

Data were collected from three secondary schools in Hong Kong, each employing a different CLIL instruction model. Descriptions of the CLIL schools in this study are provided below.

School A is a government-aided co-educational grammar school, classified as a ‘mostly EMI’ school (Lo and Lin 2019). Most subjects are taught in English, except Chinese and Chinese History. The school has used EMI since its establishment and has maintained an English-only policy for both language (L2 English) and content subject lessons. It is a top-ranked school in Hong Kong and boasts a high university admission rate.

School B is a government-aided co-educational school operating under an ‘EMI-by-class’ model, where EMI is applied in all subjects except Chinese-related ones within specific classes at each grade level. Only the existing EMI classes at the school were included in this study. Students in these classes have higher English proficiency and academic performance than their non-EMI peers at the same school.

School C is a Direct Subsidy Scheme co-educational grammar school also adopting the ‘EMI-by-class’ model. Schools B and C are middle-ranked schools based on the academic performance of their Secondary 1 students and university admission rates of their graduates. Both schools promote the frequent use of English in language and content classes during instruction and classroom interactions but allow the strategic and occasional use of L1 Chinese for concept clarification and classroom management in content classes. However, English is the only language allowed in language classes.

A total of 108 students participated in the study (n = 36, School A; n = 56, School B; n = 16, School C); 62 male and 46 female students aged between 12 and 15 years (M = 13.5, SD = 1.0). All participants were junior secondary school students who were selected for two reasons. First, discussions with school teachers revealed that senior secondary schools (years 10–12) had hectic schedules and were unable to participate. Second, this population aligns with Pun and Macaro’s (2019) call for research in Hong Kong to ‘establish the actual problem in the early years of middle school (i.e. lower secondary)’ (12). Schools A, B, and C included students from years 7 and 8, 7–9, and 7 and 9, respectively. This selection was based on the schools’ convenience. Most students (n = 99) spoke Cantonese while nine students spoke Mandarin as their L1. All participants were recruited from their existing, intact EMI classes, that is, the classes were preexisting and not specifically formed or altered for this study.

Materials

Vignette-based SRL microanalytic assessment

The recent interdisciplinary diversification of self-regulation research in applied linguistics is not reflected in the use of diverse research designs or advances in self-regulation elicitation methodologies (Rose et al. 2018; Teng and Zhang 2022). There have been calls (e.g. Rose 2015; Zhang, Thomas, and Qin 2019) for more qualitative studies that delve deeper into learners’ processing capacity for strategy use. Consequently, the vignette methodology was adopted, commonly used in social science research, combining it with a microanalytic assessment grounded in the SRL cyclical model, to investigate students’ self-regulation and strategy use in CLIL.

Vignettes are vivid scenarios that ask participants to imagine themselves engaging in or observing depicted situations that simulate real events or problems. Rooted in mental imagery research (Kosslyn, Thompson, and Ganis 2006), the vignette methodology can examine situation-specific dynamic variables, such as emotions and cognition, in language learning by creating a scenario based on a textual or (audio)visual stimulus (Goetze 2023). The strengths of the vignette methodology are its cost-effectiveness, versatility, and potential to enhance methodological consistency. It has recently been employed in emotion-related research due to its ability to simulate real-life situations and elicit authentic responses (Bielak and Mystkowska-Wiertelak 2020; Goetze 2023) but limited to this focus.

While providing a vignette mirroring a task scenario is useful for eliciting students’ strategy use, it is crucial to incorporate methodologies rooted in SRL to capture the dynamic and context-specific nature of SRL in L2 studies, such as microanalytic assessments. SRL microanalytic assessments examine learners’ behaviours and cognition in real time as they engage in tasks (Cleary 2011; Zimmerman and Moylan 2009). Assessments administered to individual students may require them to concurrently respond to context-specific questions as they apply self-regulatory processes to tasks. However, microanalytic assessment is almost absent in L2 studies (Teng and Zhang 2022), possibly due to its cost-intensiveness and challenges in eliciting information that may be difficult to observe (and sensitive). Therefore, vignette methodology was combined with a microanalytic assessment to understand participants’ motivational beliefs and metacognitive processes in response to the scenario. This approach aimed to reduce complexity by eliciting information about the quality of self-regulatory processes, rather than merely their frequency. By requiring learners to describe their reactions to hypothetical but realistic scenarios, this approach provides a nuanced understanding of how students prepare for, monitor, and evaluate their learning in CLIL contexts.

The design of the vignette-based microanalytic assessment (henceforth, the Assessment) first referred to Gkonou and Oxford’s (2016) Managing Your Emotions for Language Learning (MYE), which also adopted the vignette methodology to elicit open-ended strategies from participants regarding their emotions in response to a series of scenarios. The Assessment included 1 situation and 12 questions capturing how students plan, monitor, and adjust their learning strategies in response to academic tasks in CLIL.

In this vignette, students were instructed to imagine that they were dissatisfied with their recent science learning in the first term and wanted to improve by the end of each of the following two terms. Science was selected since it is the most common CLIL subject in Hong Kong (Kan, Kirkpatrick, and Law 2011; Lo and Lin 2019). Participants were asked to reflect on their strategies in response to the situation. The microanalysis was divided into three parts according to the three-phase cyclical SRL processes (Zimmerman and Moylan 2009). In the forethought phase, students specified their learning goals, the reasons behind them, and their expected outcomes. They also reflected on whether they anticipated success from planning their studies. During the performance phase, the participants described their learning processes and how they dealt with language difficulties. In the evaluation stage, students assessed their learning and how they would improve their future learning. To ensure clarity, the Assessment was piloted with six junior secondary students who did not participate in the study. All the instructions were presented in Chinese and English. Students were allowed to respond in either language, with the Assessment expected to take approximately 30 minutes.

Data collection and analysis

After obtaining approval from the Institution’s Research Ethics Committee, the first researcher sent invitations to schools in 2022. Three schools agreed to participate in this study. Written consent was obtained from the school principals, parents, and students. Subsequently, a Qualtrics link to the Scenario Assessment was shared with the participants.

Assessment responses were entered into a spreadsheet and analysed using NVivo 11 software. Data in Chinese were translated into English by the researcher and checked by an external reviewer, a postgraduate majoring in applied linguistics, to ensure clarity and accuracy. Content analysis was applied following Kuckartz and Rädiker’s (2023) structuring qualitative content analysis to identify the different levels of categories through several rounds of data inspection. The first author carefully reviewed the data. The main categories were developed based on Zimmerman and Moylan’s (2009) cyclical phase model of self-regulation, integrating metacognitive processes and key measures of motivation (i.e. task analysis and self-motivation beliefs in the forethought phase, self-control and self-observation in the performance phase, and self-judgement and self-reaction in the self-reflection phase). The researcher reviewed each question and response and assigned student responses to categories as coded segments. Each coded segment was extracted (e.g. ‘achieving 70 marks’) l such that it could be understood from its original context (Kuckartz and Rädiker 2023). After compiling all the coded text segments, subcategories were developed based on Zimmerman and Moylan’s (2009) sources for each main category (e.g. goal setting and strategic planning under task analysis). The researcher further segmented the coded text from the main categories into the newly defined subcategories, ensuring that each unit of meaning was coded into only one subcategory and placed into a thematic matrix (Table 1).

Table 1. Thematic matrix of Vignette Microanalytic SRL Assessment.

Phases	Main Category	Sub-category	Source		Thematic Summary
			Data segments	n
Forethought Phase	Task analysis	Goal setting
		Strategic planning
	Self-Motivation Beliefs	Self-efficacy
		Outcome expectations
		Task Interest/Value
		Goal orientation
Performance Phase	Self-Control	Task strategies
		Self-instruction
		Imagery
		Time management
		Environmental Structuring
		Help-seeking
		Interest incentives
		Self-consequences
	Self-observation	Metacognitive monitoring
		Self-recording
Self-reflection Phase	Self-Judgement	Self-evaluation
		Causal attribution
	Self-reaction	Self-satisfaction/affect
		Adaptive/defensive

The matrix included categories and subcategories, coded segments, and thematic summaries. Thematic summaries incorporated paraphrases of the text in the researchers’ own words, distilling individual statements to their essence and summarising them in relation to the research question. This study may not be representative of all CLIL students or predict individual strategic behaviours. However, group studies such as this research highlight the potential to account for different learning trajectories of similar learners over time (Thomas, Bowen, and Rose 2021), providing a broad picture of the metacognitive processes and motivational constructs among CLIL students.

Results

Forethought phase: planning and beliefs

The forethought phase included students’ task analysis processes and sources of self-motivation before engaging in the learning tasks. Students’ CLIL task analysis processes were firstly analysed, which involved setting goals and devising effective learning methods suitable for the task. Students’ self-efficacy beliefs (to what extent they believe they will succeed) and outcome expectations were analysed to reveal what CLIL students perceived and the results they hoped to obtain before engaging in the CLIL learning process.

Goal setting and strategic planning

Most students (n = 67) set clear, quantifiable goals accompanied by detailed plans. Primary goals included ‘getting at least 70% in the test’ (S28) and ‘passing the test or exam’ (S66). This focus on grades suggests that learners viewed performance as a key indicator of academic progress. Students employed strategic planning methods, such as ‘understanding the key concepts in the textbook by summarising each chapter in my own words’ (S23), ‘regularly reviewing my notes to ensure that I remember the concepts’ (S49), ‘reworking the problems until I can solve them’ (S2), and ‘attempting past papers’ (S92).

Some students (n = 13) set goals specific to content subject knowledge acquisition. They viewed science as a content subject that required a thorough understanding of progress in CLIL. Accordingly, they planned to understand ‘all the information in the science textbook’ (S4) and ‘grasp the key scientific theories and principles’ (S61). They preferred understanding the concepts over rote learning to ‘master the basic concepts to understand advanced topics later’ (S57). Planned strategies included ‘learning one to two knowledge points every day’ (S85), creating ‘detailed revision schedules to systematically cover all topics’ (S34), and taking ‘organised notes’ (S45).

Some participants (n = 12) perceived language-directed goals as advantageous. While learning science in English, these students believed that better language skills were crucial for comprehending and excelling in the subject content, particularly when noting the importance of vocabulary. Their goals included ‘reciting the science vocabulary’ (S79) and ‘mastering the grammar in reading passages’ (S40). To enhance their language skills, students planned to use certain strategies, such as ‘flashcards to memorise scientific terms’ (S72) and ‘a comprehensive list of important vocabulary’ (S54).

Peer influence and social motivation were significant factors for some participants (n = 9), with goals, such as ‘getting good grades to stay in the same class as my friends’ (S95) and ‘aiming for better results to keep pace with my classmates’ (S77). CLIL subjects appeared to heavily influence students’ results. However, not all students preferred to set goals. For example, some participants expressed, ‘I don’t set goals because I don’t want to pressure myself’ (S83), and ‘I don’t prefer to set learning goals to avoid disappointment’ (S75). These responses indicate that emotional resilience and motivational strategies may affect goal-setting.

Self-efficacy beliefs

CLIL students’ self-efficacy beliefs reflected varying degrees of confidence and preparedness regarding their studies. While many found structured plans empowering, others contended with doubts and unpredictability, indicating that perceptions of self-efficacy can significantly affect students’ confidence in achieving academic success.

Participants (n = 69) expressed positive expectations of success after making clear study plans, which helped them ‘stay disciplined and focused’ (S25), fostering a strong belief in their eventual success. Students expressed that, ‘When I have a study plan, I can track my progress, which motivates me’ (S42), and ‘Having a timetable makes me feel more organised’ (S13).

However, a smaller group of participants (n = 22) lacked confidence in their abilities due to previous unsuccessful attempts despite making plans and the belief that unexpected events may derail their plans. They expressed sentiments, such as, ‘I know that I cannot achieve my goals because I have failed before’ (S58), and ‘Something unexpected may affect my plan, so I prefer to be flexible with my studies’ (S67). These students often felt that rigid plans could not accommodate the complexities of their daily lives.

Some students (n = 9) displayed uncertainty regarding their potential for success. While they recognised the benefits of planning, they doubted their ability to effectively execute these plans, as evident in the following statements: ‘Maybe, I don’t know. I don’t usually have any plans before studying; I just study when I can’ (S73), and ‘I don’t expect success because I don’t know how to organise my schedule’ (S81). These expressions indicate that some students faced challenges in translating planning into action despite understanding its potential advantages.

Outcome expectations

In the short term, students (n = 45) were driven by the immediate goal of achieving top marks on their CLIL science exams and securing high ranks, reflecting a desire for academic excellence and competitive success. Participants expressed goals, such as ‘scoring 75–85% marks’ (S34) or ‘achieving at least 90%’ (S7), often accompanied by ambitions to be ‘in the top 10 in the class’ (S82) and ‘in the top 5 in their form’ (S55). Additionally, the prospect of gaining recognition and validation from family, teachers, or through scholarships was an expected outcome, translating into specific goals, such as ‘being in the top three of the year and winning a scholarship’ (S11).

Long-term expectations shifted towards personal growth and the enhancement of knowledge and skills, with students (n = 27) viewing the learning process as a means for personal enrichment rather than mere immediate academic success. For instance, participants aimed to ‘learn more words in English to improve their study efficiency’ (S45) and ‘gain a deeper understanding of the subject’ (S21). Meanwhile, some students wanted to achieve satisfactory outcomes that met their general standards, such as ‘good results’ (S1) and ‘a good outcome’ (S27), indicating decent performance without precise expectations. Furthermore, some CLIL students exhibited a lack of specific expectations or uncertainty about their desired outcomes, indicating that they might not have fully considered their goals.

Performance phase: strategies and techniques

During this phase, students employed various strategies to overcome the challenges encountered during the learning process, which helped them maintain their focus, effectively manage their study time, and seek the necessary assistance.

One primary strategy used by students was task strategies, which involved developing a systematic process to address the specific components of a task. When facing language difficulties, students (n = 13) frequently used online tools to translate science texts from English into Chinese. This allowed them to understand the material in their L1, reducing the time and effort required to grasp complex concepts. S55 shared that, ‘Google Translate is extremely useful. I can obtain the Chinese meaning of words without spending considerable time trying to comprehend the English text as a whole’. Task strategies also included note-taking for condensing information into manageable chunks, facilitating studying and recalling the concepts (n = 8). For example, students mentioned, ‘I note down difficult words in my vocabulary notebook so that I can revise them later’ (S71), ‘I maintain a list of important vocabulary in a notebook for regular review’ (S80), and ‘I write down important words in a notebook for easy revision’ (S9). These examples demonstrate how notetaking aids vocabulary and content knowledge retention, fostering a positive emotional connection and engagement with the subject matter delivered through EMI and providing a tangible record of areas needing further attention (Yuan 2024).

Help-seeking was a prevalent strategy (n = 35), reflecting students’ recognition of the importance of external support. Help-seeking involves soliciting assistance while learning or performing tasks. When faced with language or content difficulties, many students opted to ask for help. For instance, S13 stated, ‘I ask my teacher for help’, highlighting the role of teachers as a primary source of support. S34 mentioned, ‘I ask my friends and teachers for help in understanding complex concepts’, indicating a collaborative approach. S56 shared, ‘I ask my elder sister because she is in Secondary 4 and has more experience’, demonstrating reliance on familial support for academic guidance. Content teachers were cited as a crucial source of help when encountering language challenges; however, none of the students mentioned their language teachers. Rather, they turned to content teachers for help with both content and/or language challenges. In out-of-school contexts, students watched online videos to self-direct their learning (n = 6). This facilitated the learning process and resolved their linguistic and content difficulties. S86 explained, ‘I search for online tutorial videos and take notes to ensure that I understand the concepts’. Additionally, S48 shared, ‘I use educational applications on my tablet to review the concepts and attempt the practice exercises’. These examples show how students actively seek knowledgeable individuals, both inside and outside the classroom, who can provide the guidance they need to successfully navigate academic challenges.

Self-instruction played a crucial role in managing academic challenges in CLIL settings by promoting self-reliance and study techniques (n = 15). Self-instruction refers to the use of self-guided methods for understanding and mastering new information. When encountering challenges in content and language, students often face difficulties in contextually deriving meaning. Here, self-instruction helps students use contextual clues to infer meanings and enhance their comprehension skills. For instance, students mentioned various strategies, such as attempting to understand the text by ‘reading the content near the word’ (S44), guessing the meaning before searching it in the dictionary to ‘test their understanding’ (S18), and ‘deducing the meaning of the word based on the context, then checking the dictionary if necessary’ (S67). Self-instruction is often associated with technology-enhancing strategies. S22 mentioned using the Internet for assistance: ‘I search for anything that I don’t know on Google’. This blend of self-instruction and digital resources helps students become independent and resourceful.

Time management is another vital approach for balancing study schedules in CLIL. Students (n = 9) described their strategies for accomplishing learning tasks on schedule, including setting specific task goals, estimating the time requirements, and monitoring their progress. For instance, S89 mentioned, ‘I set a specific time each day to review my notes and ensure that I remember the details’. S103 added, ‘I look up the meaning of the words and memorise them'. Effective time management allows students to allocate their study time efficiently, ensuring that all tasks are completed. Thereby, students ensure that they address gaps in their knowledge and keep pace with their studies without feeling overwhelmed.

Metacognitive monitoring involves informal mental tracking of performance processes and outcomes. Students (n = 6) often engaged in metacognitive monitoring to ensure that they understood the language and content of their subjects. Participants shared, ‘I review all the concepts until I remember what they mean in English’ (S47), and ‘I repeatedly read the book until I understand all the concepts’ (S29). These practices ensure that students do not move on from a topic until they are confident in their understanding, thus reinforcing their learning and aiding retention.

Self-Reflection phase: judging and reflecting

In this phase, students responded to scenarios with reactions and judgements about their learning outcomes, providing insights into how they assessed their performance and made informed decisions to enhance their future learning experiences.

Self-judging CLIL performance

Participants reflected on their learning experiences by comparing their outcomes with their expectations and previous performance levels, providing causal attributions that link their success or failure to specific causes. Generally, students expressed satisfaction and a sense of achievement. Students (n = 58) who received satisfactory results often felt that their planning and monitoring efforts could lead to positive outcomes, as seen in comments, such as ‘feel productive’ (S97) and ‘feel very happy and excited to achieve the marks I want’ (S100). These reflections indicate a sense of accomplishment and satisfaction associated with the positive emotions linked to their achievements. Students felt relieved and confident about their learning processes, attributing their success to effective planning stages. They credited their improved performance to strategic planning efforts, such as ‘feeling more confident in learning English after planning’ (S99). These attributes can significantly affect students’ motivation and future efforts, reinforcing the importance of recognising the factors that influence their learning outcomes.

Conversely, students (n = 27) reported defensive decisions in response to negative self-reactions. They experienced feelings of tiredness, pressure, and a lack of satisfaction, such as feeling ‘hopeless and helpless’ (S64), ‘I am a failure’ (S99), and deciding to ‘give up studying science and choose other subjects’ (S2). This negative self-evaluation can significantly undermine their motivation to continue striving and their willingness to persevere. Some described feelings of hopelessness or failure, which suggest that their planning and monitoring strategies did not meet their expectations or needs. They attributed their challenges to external factors, ‘Sometimes I feel tired because I don’t get enough time to sleep’ (S16), indicating the physical and mental toll of their current study habits. Internal factors could relate to a lack of engagement or commitment to their study strategies, such as ‘I didn’t plan or monitor my learning’ (S4) and ‘I feel nothing about my experience’ (S53). These responses suggest that students may not fully recognise the potential impact of structured planning and monitoring on their academic performance.

However, some students (n = 7) displayed a growth mindset, reflecting a desire to adapt and improve. For instance, they shared, ‘I don’t feel satisfied, and I would love to improve my time management skills’ (S98), and ‘I think that I can do better’ (S23). These statements demonstrated an awareness of their potential and commitment to self-improvement.

Adaptive and defensive decisions

Students made adaptive (continuing or modifying strategies to enhance learning) and defensive decisions (avoiding further efforts to prevent dissatisfaction) regarding their future learning in CLIL. Many participants (n = 74) exhibited a willingness to make adaptive changes to improve their learning and actively sought ways to enhance their strategies and outcomes. They recognised areas for improvement and proposed specific actions to address them, including ‘spending more time on studies after school’ (S24) and ‘having a timetable to ensure better time management’ (S53). Participants understood the importance of setting goals and creating structured plans to maintain consistent study habits.

However, some students (n = 14) displayed defensive reactions, indicating a reluctance to adapt their strategies due to perceived shortcomings. They were reluctant to change their current study habits or acknowledge the need for improvement. Statements, such as ‘I feel that I am a failure. I will continue to be a failure’ (S59), and ‘No, I wouldn’t like to do that (change) to improve my future content learning’ (S115), signal a defensive stance. These students may feel that their existing methods are sufficient or may resist change because of a lack of confidence or fear of additional effort. Some responses indicated a dismissal of the value of planning and monitoring, suggesting a fixed mindset, which could impede their academic growth and limit their potential.

Discussion

This study is among the first to holistically provide empirical evidence of CLIL students’ strategic learning behaviour by adopting Zimmerman and Moylan’s (2009) social cognitive SRL model. It examined how students strategically set goals and plans before studying (forethought phase), strategies for tackling language and content challenges during studying (performance phase), and self-judgement with further strategies employed (reflection phase). The results are particularly novel as they offer a comprehensive view of students’ goals, motivational beliefs, and metacognitive skills, a perspective largely absent in previous quantitative studies, thus providing a more holistic understanding that expands upon earlier research. Moreover, this study is pioneering in that it adopted the vignette methodology, combining with microanalytic assessment, contributing to studies on self-regulation, cognition, motivation, and emotion.

Rethinking the learning–teaching–assessment cycle in CLIL

Findings revealed that students set specific goals for their CLIL studies, involving strategic planning. Self-efficacy beliefs varied; most students expressed confidence because of clear plans, whereas some doubted their success because of past failures or unpredictable events. Short-term expectations focused on achieving high exam scores and class rankings, while long-term expectations emphasised personal growth and enhanced knowledge. Goal planning and strategies in CLIL often rely heavily on assessment, a blind spot and an under-researched area in CLIL (Lo and Fung 2020; Massler, Stotz, and Queisser 2014). Examination success in Confucian-influenced societies, such as Hong Kong, mainland China, Japan, and Korea, is seen as a major purpose of education, an opportunity for social mobility, and a means of ‘gaining face’ for individuals and families (Carless and Lam 2014). Although Hong Kong has long debated its MoI policy and exam-oriented culture (Carless 2011), listening to learners is crucial for making CLIL learning ‘successful’ (Coyle 2013). This requires planning, learning, teaching, and assessment to meet learners’ needs. Findings indicated that students prioritised assessment as their primary goal; therefore, this paper argues for revisiting the learning–teaching–assessment cycle in CLIL.

Teachers and policymakers should strive to balance the dominant CLIL approach of Assessment of Learning (AoL) with both Assessment as Learning (AaL) and Assessment for Learning (AfL). Coyle (2013) emphasised that learner achievement in CLIL involves ‘organic processes that cannot be exclusively captured through summative means’ (248). While AoL focuses on evaluating learner achievement and AfL utilises formal and informal assessment activities, we encourage putting more emphasis on AaL, which blurs the boundaries between assessment and learning, making assessment ‘not merely an adjunct to teaching and learning but offers a process through which pupil involvement in assessment can feature as part of learning’ (Dann 2002, 153). It relates to developing pupils’ engagement in and response to CLIL students' self-assessment, focusing on exploring self-regulation, self-efficacy, metacognition, and feedback as dimensions of both assessment and learning. Teachers can support AaL in CLIL by developing and assessing students’ self-regulatory practices, such as planning, performance, and reflection. This approach encourages students to actively question their learning processes, identify areas of improvement, and set personal learning goals. While summative assessment remains prevalent in societies with deep-rooted educational cultures, especially in East Asia, formative and continuous assessment should be encouraged at both the school (school-based assessment) and policy levels (public examinations).

From self-regulated to other-regulated learning for ‘successful’ CLIL

To foster self-regulation in CLIL, it is essential to promote students’ learning autonomy by rethinking the roles of others in SRL. In formal language learning contexts, strategic learners may not be self-regulated or autonomous, as suggested in the literature (Bowen and Thomas 2022; Thomas, Bowen, and Rose 2021). Our study found that neither were all learners strategic nor did they wish to be self-regulated. CLIL requires students to manage content and L2 simultaneously, and many students are not fully aware of these challenges or how to address them. Therefore, effective scaffolding is crucial for optimising self-regulatory practices and meeting the linguistic and cognitive demands of CLIL.

Various resources can support self-regulatory practice, with teachers acting as transformative figures in the classroom and supplemented by resources, such as L1 and L2 symbols, images, and videos. Adopting Thomas and Rose’s (2019) concept of the inevitable stabilisation of regulation in an other-determined state, Lin’s (2015) Multimodalities-Entextualization Cycle can guide CLIL educators in employing multilingual and multimodal resources to support and enhance self-regulatory practices (see Lo 2024). From a translanguaging perspective, Tai and Li (2021) also argue that co-learning in EMI classrooms through mobilising various resources promotes equity in knowledge construction, as the teacher is no longer the sole holder of knowledge in the class. Instead, teachers and students learn from each other and engage in the joint construction of knowledge. Outside school, parents and private tutors play similar roles, with digital tools having enormous potential to significantly enhance students’ self-regulation, making them more autonomous and strategic (Lai and Hwang 2023). For instance, learning apps with self-regulatory features can help students set goals, track their progress, and reflect on their learning strategies. Future research should explore the role of AI in facilitating CLIL, nurturing self-regulated behaviour among students, and predicting learning outcomes (cognitive, emotional, and motivational).

The concept of other-determined state can be extended to teacher collaboration or CLIL teacher self-regulation, where language and content teachers must collaborate to achieve better learning, teaching, and assessment outcomes. Co-regulation could be an area for further research on CLIL teacher education and professional development.

Self-reflective learners: towards deeper learning in CLIL

This study showed that CLIL students assessed their performance through self-judgement and self-reaction, which significantly impacted their motivation and learning strategies. Positive self-evaluations often lead to satisfaction and increased motivation, with students attributing success to effective planning and monitoring. Conversely, negative self-evaluations can result in feelings of failure and demotivation. Bandura (1986) considered self-reflection an important personal attribute that contributes to one’s ability to positively alter their thinking and behaviour; however, only a few studies have investigated self-reflective practice in CLIL literature, focusing specifically on teachers (e.g. Lazarević 2022).

Recent CLIL research has focused on deeper learning through which students can apply their learning to new situations (Coyle and Meyer 2021). Among all the constructs in self-regulation, self-reflection is considered the core predictor of continued interest and motivation for learning, with the potential to impact progression in learning and overcoming new challenges (Zimmerman and Moylan 2009). Mezirow (2009) argued that the transformation into deeper learning occurs in the reflective process and the sharing of one’s learning process with others. In CLIL contexts, it is paramount for students to be self-reflective, not by summative means but through cognition, emotion, and decision-making. Students should recognise and address their perceived difficulties in linguistic nuances and content knowledge with the support of experienced experts. CLIL students’ cognitive processes are also moderated by their affectivity (Otwinowska and Foryś 2017). While celebrating students with a positive self-evaluation and growth mindset, those with a negative self-evaluation should be provided more scaffolding. In Hong Kong, students are occasionally trained to perform well on tests while only developing a limited or partial understanding of what has been assessed (Carless 2011). Maximising students’ self-reflective practice in both cognition and emotions can enhance CLIL learning, teaching, and assessment, leading to deeper learning that extends beyond CLIL in plurilingual and pluricultural classrooms (Coyle and Meyer 2021).

Conclusion

This study investigated students’ self-regulated practices in CLIL using the social cognitive SRL model. CLIL has been widely adopted at all educational levels with the aim of developing both content knowledge and L2 proficiency. Optimising self-regulation and CLIL requires a deep understanding and consideration of how students think and act. However, self-regulation does not rely solely on learners’ individual efforts; navigating the complexities of the dual linguistic and cognitive demands of CLIL necessitates practical support both in the school environment and through digital resources. Not all students are naturally autonomous, so nurturing self-regulation skills is essential. Using the socio-cognitive model of SRL, future research can explore different phases of SRL and the relationships between the constructs in different contexts to gain a broader understanding of the motivational beliefs and metacognitive strategies of CLIL students.

Acknowledgments

I would like to express my gratitude to all the participating schools, teachers, and students in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Alfred W. T. Lo

Alfred W. T. Lo is an Oxford-Wolfson-Marriott doctoral researcher of Applied Linguistics and East Asian Studies at the University of Oxford, UK, as well as a junior visiting scholar of English Language Education at the Education University of Hong Kong, Hong Kong SAR, China. His research interests include Content and Language Integrated Learning, English Medium Instruction, self-regulation/self-regulated learning, Global Englishes (and its pedagogical implications), and transcultural communication.