Transforming Self-Identity in EMI: The Interplay of Behavioral Engagement, Motivational Intensity, and Self-Efficacy

Abstract

This study investigates the dynamic relationships between self-efficacy, motivational intensity, and behavioral engagement and their effects on positive and negative self-identity among EFL learners participating in English Medium Instruction (EMI). Conducted with 390 undergraduate students from diverse academic disciplines, the research utilized a longitudinal design, administering pre- and post-intervention assessments of these constructs. A repeated measures MANOVA revealed significant improvements in all three independent variables, with behavioral engagement showing the largest gains, followed by motivational intensity and self-efficacy. Multiple regression analysis further demonstrated that these predictors collectively explained substantial variance in positive and negative self-identity. Behavioral engagement emerged as the most influential factor, positively enhancing self-perception, and mitigating identity conflicts, followed by motivational intensity and self-efficacy. The findings underscore the transformative potential of EMI, emphasizing the role of active participation, sustained motivation, and confidence-building in fostering positive self-identity while reducing negative self-perceptions. These insights highlight the importance of interactive and participatory pedagogical strategies, goal-oriented motivation, and scaffolded support in EMI contexts. By integrating these elements, educators can maximize the holistic benefits of EMI, preparing learners for global academic and professional opportunities. The study contributes to the growing body of research on EMI, offering practical implications for enhancing student engagement, motivation, and identity development. Future research is encouraged to explore the long-term impacts and contextual moderators of these relationships in diverse EMI settings.

1. Introduction

English Medium Instruction (EMI) has become increasingly prominent in Taiwanese higher education over the past two decades, driven by forces of internationalization, globalization, and demographic changes. Since Taiwan joined the World Trade Organization in 2002, the government has implemented various policies and initiatives to promote EMI, including “the Aim for the Top University Project” and “Teaching Excellence Project”, with the dual goals of enhancing local students’ global competitiveness and attracting international students amid declining domestic enrollment due to low birth rates (Tsou & Kao, 2017). According to recent statistics from Taiwan’s Ministry of Education, the number of EMI courses offered at Taiwanese universities increased from approximately 2500 in 2005 to over 21,000 in 2021, reflecting its growing strategic importance in higher education internationalization.

The rapid expansion of EMI programs, however, has raised important questions about their effectiveness and impact on students’ learning experiences and outcomes (Guo et al., 2024; Gupta & Lin, 2023). While EMI aims to simultaneously develop students’ English proficiency and subject knowledge, research indicates that its success depends heavily on complex psychological factors that influence the learning process. Among these, three interrelated constructs have emerged as particularly critical: engagement, motivational intensity, and self-efficacy. Engagement refers to students’ active participation and investment in EMI courses, manifested through behavioral, cognitive, and emotional dimensions. Motivational intensity captures the degree of effort learners expend in pursuing their EMI learning goals, while self-efficacy encompasses their beliefs about their capabilities to successfully learn both content and language in an English-medium environment (Chang, 2010; H. I. Huang, 2020).

Research indicates that EMI can significantly improve English proficiency over time, positively impacting academic content learning (Soruç & Griffiths, 2018; Yuksel et al., 2023). In Korean higher education, EMI has yielded positive outcomes, including overall satisfaction and improved English proficiency, although it demands a flexible approach and adequate support (Byun et al., 2011; Guo et al., 2024). Similarly, a study in a Spanish university found no adverse effect of EMI on student academic outcomes in courses like Financial Accounting (Dafouz & Camacho-Miñano, 2016). Additionally, EMI enhances linguistic and social capital, contributing to a more international professional identity for younger teachers (Dafouz, 2018).

Recent studies in the Taiwanese context have highlighted both opportunities and challenges regarding these psychological factors in EMI implementation. Survey findings suggest that while many students recognize the potential benefits of EMI for their future careers and are generally satisfied with EMI courses, they often report low self-efficacy and face difficulties in maintaining sustained engagement and motivation, particularly when confronting linguistic challenges (Yeh, 2014). The situation is further complicated by Taiwan’s status as an expanding circle country where English is learned as a foreign language, meaning most students have limited exposure to English outside the classroom. This context makes the development of positive psychological orientations toward EMI especially crucial yet challenging. As Taiwan continues to promote EMI as a key strategy for higher education internationalization, with ambitious goals like the Tainan City Government’s initiative to create a fully bilingual environment by 2025, understanding and addressing these psychological dimensions becomes increasingly important for ensuring effective implementation and positive learning outcomes. This requires moving beyond purely technical questions of EMI effectiveness to examine how institutional policies and pedagogical practices can better support students’ engagement, sustain their motivational intensity, and enhance their self-efficacy in EMI contexts.

Despite these efforts, the prevalence of EMI courses in Taiwanese higher education institutions remains relatively low. As of 2021, EMI courses constituted less than 5% of the curriculum in these institutions. This limited adoption underscores the importance of understanding the psychological factors that influence student engagement in EMI settings. Factors such as motivational intensity and self-efficacy play crucial roles in students’ ability to navigate EMI courses effectively. Research indicates that while students recognize the benefits of EMI, including improved English proficiency and enhanced global competitiveness, they also face challenges related to language barriers and academic performance. Addressing these psychological factors is essential for the successful implementation of EMI and the realization of Taiwan’s bilingual aspirations.

2. Literature Review

2.1. EMI in Taiwan

English as a Medium of Instruction (EMI) has become a pivotal strategy in Taiwan’s higher education system for internationalization, as part of the government’s ambitious “Blueprint for Developing Taiwan into a Bilingual Nation by 2030” initiative (Ferrer & Lin, 2024). Launched in 2018, this policy promotes EMI to enhance English proficiency among students and to position Taiwanese universities competitively on a global stage. However, the implementation of EMI has sparked debates and raised concerns about widening horizontal inequalities among higher education institutions (HEIs). Public universities, particularly flagship institutions, receive substantial funding and support, enabling them to offer more EMI courses and thus providing their students with greater linguistic capital. In contrast, private universities of technology, which host a significant portion of students often from lower socio-economic backgrounds, struggle with limited resources and lower English proficiency among their students, potentially exacerbating existing disparities. This scenario suggests that while EMI is promoted with the aim of enhancing educational quality and international competitiveness, its application might inadvertently perpetuate and accentuate educational inequalities within Taiwan’s hierarchical education system (Y. H. I. Huang, 2024).

Despite its growth, EMI in Taiwan faces several challenges. Concerns about instructors’ and students’ English proficiency, depth of content knowledge, and student motivation have been raised (Zhou et al., 2023). Empirical studies suggest that while students recognize the benefits of EMI, they often struggle with understanding content in English. Instructors mainly view EMI as a way to accommodate international students and improve local students’ English proficiency. The dual benefits of acquiring content knowledge and improving English language skills are counterbalanced by questions about the depth of class interaction and knowledge dissemination (Chang, 2010; S. Y. Huang, 2012; Wu, 2023; Yeh, 2014).

The effectiveness of EMI in Taiwan is influenced by students’ English proficiency. Despite not expressing negative attitudes, students often do not feel confident in their English language comprehension and attribute learning difficulties to their language skills. This suggests that EMI’s effectiveness is closely tied to students’ language proficiency. Additionally, students’ confidence and willingness to engage with others are noted as positive outcomes of EMI (Zhang, 2018).

The role of instructors is crucial in student course selection for EMI (O’Dowd, 2018; Macaro, 2018). However, there are challenges such as the teachers’ lack of English proficiency, concerns over cultural challenges (Tang, 2020), and institutional policies. Local faculty sometimes express doubts about the practicality of EMI, and the lack of practical English language applications in Taiwan can lead to skepticism about its usefulness.

Research reveals that the effectiveness of EMI in higher education is contingent on factors such as language proficiency, support systems, teacher training, and appropriate instructional strategies (Byun et al., 2011; Guo et al., 2024; Rahman & Mehar Singh, 2020; Wilkinson, 2017). Team teaching and collaboration between language and content teachers can enhance learning outcomes and motivation (Lasagabaster, 2018). Furthermore, self-efficacy and prior educational experience significantly influence success in EMI courses (Thompson et al., 2022).

In short, the implementation of EMI in Taiwan’s higher education is a strategic response to demographic changes and the desire to maintain global competitiveness. While it presents numerous benefits, including improved English proficiency and multicultural exposure, it also poses significant challenges related to language proficiency, teaching quality, and cultural preservation. The effectiveness of EMI is closely linked to the preparedness and support of both instructors and students, necessitating ongoing adaptation and evaluation to maximize its potential benefits.

2.2. EMI and Identity

The relationship between identity and English as a Medium of Instruction (EMI) in language education is multifaceted, reflecting the interplay of linguistic, cultural, and social dynamics, as illustrated in Fang and Hu’s (2022) study on Teochew-speaking learners in Chinese higher education. EMI, as a policy promoting English for academic instruction, often positions it as a dominant language of knowledge and global competitiveness, which can significantly shape students’ academic and social identities. For minority language speakers like the Teochew students in the study, EMI introduces a linguistic hierarchy where English and Putonghua overshadow their L1 (Teochew), relegating it to a marginal role in academic settings (pp. 10–11). This marginalization can lead to a sense of dislocation from their ethnolinguistic heritage, as students negotiate their identities within a framework that prioritizes global lingua francas over local languages. However, the study also reveals that some students resist this hierarchy by valuing Teochew as a core component of their identity, using it as a resource for solidarity and cultural pride despite its diminished academic utility (pp. 11–12). Thus, EMI both challenges and reinforces identity, creating a tension between global aspirations and local rootedness.

This tension is further complicated by the practical implementation of EMI and its perceived effectiveness, which influences how students construct their identities as learners and multilingual individuals. The Teochew participants expressed mixed attitudes toward EMI, with some appreciating its potential for international competitiveness (p. 9) while others felt it acted as a gatekeeper, hindering content comprehension and fostering frustration or guilt among those with lower English proficiency (pp. 9–10). This ambivalence highlights how EMI can enhance self-efficacy and cosmopolitan identity for proficient students, aligning with neoliberal goals of employability and mobility, while simultaneously alienating others, potentially eroding their confidence and academic identity. The lack of language support and the top-down imposition of EMI (p. 9) exacerbate these disparities, suggesting that identity construction is not uniform but contingent on individual linguistic capital and institutional practices. Translanguaging emerges as a coping strategy (p. 10), allowing students to draw on Teochew and Putonghua to bridge comprehension gaps, thereby asserting agency in identity negotiation and resisting the monolingual ideology often embedded in EMI policies.

From a broader perspective, the relationship between EMI and identity underscores the need for a critical multilingual approach to language education, as Fang and Hu advocate (pp. 12–13). EMI’s dominance risks perpetuating neo-colonial ideologies and linguistic inequality, particularly in postcolonial or multilingual contexts like China, where national (L1) and global (L2) languages compete with minority languages (p. 13). For Teochew speakers, this dynamic reflects a shift from a historically vibrant L1 to a “low variety” in a linguistic hierarchy (p. 11), threatening intergenerational transmission and cultural heritage. Yet, the students’ insistence on maintaining Teochew (p. 12) signals a resilient identity tied to linguistic diversity, challenging the homogenizing tendencies of EMI. This suggests that effective EMI policies should integrate translanguaging and minority language recognition to support inclusive identity construction, balancing global competencies with local cultural values. Such an approach could mitigate the identity crisis faced by minority language learners, fostering a multilingual identity that embraces both academic success and ethnolinguistic pride.

Self-identity in language education, particularly EMI, is a dynamic construct reflecting individuals’ evolving perceptions of their place in the world, shaped by linguistic proficiency, social interactions, and future aspirations (Norton, 2013; cited in Dafouz, 2018, p. 5). Drawing on investment theory, it encompasses agency-driven negotiations of cultural and social capital (Bourdieu, 1986), while ROAD-MAPPING situates it within EMI’s sociolinguistic dimensions (Dafouz & Smit, 2016, p. 4). Debates center on its fluidity versus stability—whether it’s a shifting adaptation or a stable core—and its individual versus collective nature, with post-structuralists emphasizing change over fixed labels (Block, 2014). Operationalizing self-identity is equally contested: qualitative narratives capture its richness (p. 7), but lack scalability, while quantitative metrics risk oversimplification, complicating its study across diverse EMI contexts.

In EMI, self-identity emerges as both empowered and contested, as seen in Dafouz’s (2018) Spanish lecturers who leverage English to craft internationalized, imagined identities (e.g., “visibility”, Ex. 12, p. 9), enhancing their professional capital and agency (pp. 9–10). EMI facilitates this shift by positioning English as a gateway to global academic communities, yet it introduces tensions—lecturers advocate bilingual identities to resist English hegemony (Ex. 17, p. 10), reflecting a local-global duality. This aligns with broader findings: EMI fosters elite, cosmopolitan identities (Jahan & Hamid, 2019) but may marginalize local ones (Fang & Hu, 2022), fueling debates over its neo-colonial implications. Understanding this link requires mixed methods to trace how institutional policies and personal investments reshape identity, balancing measurable outcomes with subjective aspirations.

2.3. Behavioral Engagement and EMI

Behavioral engagement, a vital aspect of student engagement, has been widely recognized for its pivotal role in shaping educational outcomes. As a multidimensional construct, engagement encompasses behavioral, cognitive, and emotional dimensions, each contributing to learning in distinct yet interconnected ways (Fredricks, 2011; Fredricks et al., 2004; Mandernach, 2015). Among these, behavioral engagement is particularly significant in the context of English as a Medium of Instruction (EMI), where observable actions, such as active participation and sustained effort, are critical to language learning success.

Defined as the physical effort and time invested in learning tasks, behavioral engagement represents a learner’s visible commitment to their academic journey (Finn & Zimmer, 2012). This dimension is not merely a passive indicator of involvement; rather, it actively mediates and reflects learners’ emotional and cognitive states. For instance, students who exhibit enthusiasm or curiosity (emotional engagement) are more likely to persist in EMI-related activities, while those with strong motivation or clear learning goals (cognitive engagement) often demonstrate greater behavioral involvement (Zhang, 2017). Thus, behavioral engagement functions as a bridge, translating internal psychological states into tangible actions that facilitate academic progress.

In the EMI context, where learners navigate the dual challenges of acquiring subject knowledge and mastering English, behavioral engagement takes on heightened importance. Actively participating in discussions, completing assignments diligently, and investing time in language practice are behaviors that directly impact language acquisition and comprehension. Mercer (2019) highlights the necessity of active participation, emphasizing that engagement, fundamentally, is about taking action. This perspective underscores the summative role of behavioral engagement in fostering a comprehensive and impactful learning experience.

Furthermore, the interplay between behavioral engagement and motivational intensity is particularly relevant in EMI settings. Motivational intensity, defined as the effort and persistence learners exhibit toward achieving their goals, serves as a driving force behind behavioral engagement. Learners with high motivation are more likely to invest time and energy in language learning tasks, demonstrating persistence even in the face of challenges (Reschly & Christenson, 2012). This synergy creates a feedback loop: motivated learners engage behaviorally, and successful engagement reinforces their motivation, creating a cycle of continuous improvement (Carless, 2019; Gan, 2020).

Behavioral engagement also interacts with language self-identity, shaping and being shaped by learners’ perceptions of themselves as English users. Additive self-identity (Appendix A), for example, where learners integrate English into their linguistic repertoire without diminishing their native language, encourages active engagement in EMI tasks. Conversely, subtractive or split self-identities, where learners perceive conflict or loss associated with language learning, may hinder behavioral engagement. This dynamic demonstrates the intricate connection between engagement and the socio-cultural dimensions of language learning.

In conclusion, behavioral engagement is indispensable in EMI environments, serving as a conduit for emotional and cognitive dimensions while directly influencing academic outcomes. Its observable and actionable nature makes it both a critical indicator and facilitator of learning success. Moreover, its interplay with motivational intensity and self-identity highlights its central role in shaping a learner’s holistic engagement. As such, fostering behavioral engagement through targeted strategies is essential for maximizing the potential of EMI to enhance both language proficiency and academic achievement (Harris & Strefford, 2024; Tai & Wei, 2025).

2.4. Self-Efficacy and EMI

In the context of English Medium Instruction (EMI), understanding the role of self-efficacy and other motivational beliefs is crucial for comprehending how students engage and succeed in their academic endeavors (Tsui, 2024). This summary synthesizes two texts focusing on the importance of self-efficacy and other related motivational factors in EMI settings.

Self-efficacy, defined as the belief in one’s capabilities to execute actions necessary for specific achievements (Bandura, 1997), plays a significant role in educational settings, particularly in EMI programs. It is closely tied to self-concept, which involves evaluations of one’s ability and normative comparisons with others (W. C. Huang, 2024). While self-concept encompasses broader perceptions of personal competence, self-efficacy is more task-focused and future-oriented, dealing with beliefs about performing specific tasks within a domain (Mercer, 2008). Studies have shown that both foreign language self-efficacy and L2 self-concept positively correlate with language learning achievement (Mills, 2014; Phakiti et al., 2013) and that EMI language support programs can enhance students’ self-beliefs and confidence in academic English tasks (Thompson et al., 2022).

The influence of self-efficacy on student performance in EMI is profound. Students with stronger self-efficacy beliefs are more likely to exert effort and persist in their EMI studies, as they are motivated both by the potential for content understanding and language improvement. This contrasts with students possessing weaker self-efficacy, who tend to focus on the challenges of EMI, such as the difficulty in comprehending content in English (Thompson et al., 2022). Self-efficacy impacts students’ resilience in overcoming task difficulty and anxiety, shaping their overall learning experience in EMI settings.

Language proficiency is a critical factor in EMI success, influencing both academic performance and students’ motivational beliefs. Research indicates that L2 proficiency and non-cognitive factors, such as self-efficacy, independently contribute to student success in EMI (Kojima & Yashima, 2017; Sahan et al., 2023; Soruç et al., 2022; Xie & Curle, 2022). The relationship between language proficiency, self-efficacy, and students’ approach to EMI tasks highlights the importance of these factors in navigating the complexities of learning content in a second language.

EMI learners’ motivational beliefs are influenced by various factors, including goal orientation and task value (Bandura, 1997; Pintrich, 2000). Students may be intrinsically motivated by curiosity or achievement in learning the content of EMI courses, or they might be driven by extrinsic rewards like grades or job opportunities (Galloway et al., 2017; Zhou et al., 2024). The dual purpose of learning content through English often emerges as a primary motivation, indicating that students value the task of EMI study itself. In a study conducted in China, Iwaniec and Wang (2023) found that both extrinsic and intrinsic motivations were at play, with students unable to separate the aspects of mastering content and language. This suggests a complex interplay of motivational beliefs affecting students’ engagement in EMI.

The complex relationship between motivational beliefs and self-efficacy in EMI contexts necessitates further exploration. Students’ perceptions of difficulty, language proficiency, and self-efficacy toward EMI study are influenced by their extrinsic and intrinsic goal orientations, as well as the perceived value of the task. Understanding this interplay is critical for enhancing EMI teaching methodologies and support programs.

In summary, self-efficacy emerges as a pivotal factor in EMI contexts, influencing students’ engagement, effort, and resilience. The interplay of self-efficacy with language performance, and self-regulated listening strategies underline the multifaceted nature of learning in EMI settings. Further research is needed to unravel the nuances of these relationships and to develop strategies that can effectively support students in maximizing their potential in EMI programs.

2.5. Engagement, Motivational Intensity and Self-Efficacy

As for the interrelationships among engagement, motivational intensity, and self-efficacy, Harris and Strefford (2024) developed and validated a measurement instrument for assessing EMI learners’ self-efficacy, engagement, and satisfaction, finding significant correlations between these constructs and reading proficiency. Kim (2024) focuses on the interaction between language mindsets (growth vs. fixed), self-efficacy, engagement, and self-perceived proficiency (PP) in second language (L2) acquisition. The study reveals that a growth mindset significantly predicts higher self-efficacy, which, in turn, influences engagement and perceived proficiency. This study emphasizes the importance of a growth mindset in fostering motivation and successful L2 learning. Moreover, Wang et al. (2022) delve into the significance of motivational intensity—the degree of effort learners exert in language learning—and its relationship with various internal and external factors such as personal commitment, family background, and school support. Using structural equation modeling, the study explores how motivational intensity influences learning motivation, highlighting its direct and indirect effects on engagement and learning outcomes.

Literature has underscored the interdependence of engagement, motivation, and self-efficacy in shaping learners’ L2 development. Motivational intensity is a behavioral manifestation of motivation, directly linking learners’ efforts to their outcomes. In contrast, self-efficacy refers to learners’ beliefs in their abilities, which play a mediating role between mindsets, engagement, and proficiency. These studies collectively suggest that a growth motivation enhances self-efficacy, which in turn boosts engagement, facilitating improved language learning outcomes. This dynamic relationship indicates that students with higher self-efficacy—fueled by a growth motivation—are more likely to engage deeply with the learning process, thereby shaping their self-identity as competent language learners (Tsui & Ngo, 2017). The perceived improvement in self-identity can be seen as both a cause and effect of sustained motivational intensity and engagement in language learning.

2.6. Research Gap

Despite extensive research on self-efficacy, engagement, and motivation respectively, the specific mechanisms through which these factors interact to shape L2 self-identity in EMI remain underexplored. Existing studies often examine these constructs in isolation, overlooking their potential synergistic effects. Moreover, little attention has been given to how behavioral engagement mediates the relationship between self-efficacy and motivational intensity, particularly in fostering additive or productive self-identities. The role of contextual factors, such as online or blended learning formats and varying levels of English proficiency, in moderating these relationships also requires deeper investigation.

2.7. Research Purpose

The purpose of this study is to examine the dynamic relationships between behavioral engagement, motivational intensity, and self-efficacy in the context of English as a Medium of Instruction (EMI) and to explore their collective influence on the development of L2 self-identity. Specifically, this research aims to (1) investigate whether undergraduate students demonstrate measurable improvements in self-efficacy, motivational intensity, and behavioral engagement before and after participating in EMI courses, and (2) evaluate the extent to which these constructs positively contribute to the formation and enhancement of self-identity within the EMI learning environment. By addressing these objectives, the study seeks to provide deeper insights into how EMI can shape learners’ engagement, motivation, and self-perception, offering implications for pedagogical practices aimed at fostering effective language learning and identity development in EMI contexts.

2.8. Research Rationale

The global expansion of English as a Medium of Instruction (EMI) has reshaped higher education, particularly in non-native English-speaking countries, including Taiwan. This pedagogical shift aligns with the increasing demand for internationalization in education, aiming to enhance English proficiency, cultural integration, and global competitiveness. EMI presents learners with the dual challenge of mastering academic content while developing their English language skills, making it a fertile ground for examining the dynamic interplay of engagement, motivation, and identity in language learning.

The research underscores the critical role of behavioral engagement, motivational intensity, and self-efficacy in academic success, particularly in EMI contexts. Behavioral engagement, encompassing visible actions such as class participation and task completion, provides a measurable indicator of learners’ active involvement. Motivational intensity, defined as the effort and persistence in achieving learning goals, drives this engagement, while self-efficacy, reflecting learners’ confidence in their abilities, directly influences their resilience and performance. Together, these constructs form a triad of interrelated factors that underpin effective learning in EMI environments.

The theoretical framework drawn from the literature aligns with engagement theory, self-efficacy theory, and motivational intensity, which suggest that the intersection of engagement, motivation, and self-efficacy can significantly influence self-identity. As students’ confidence in their learning abilities (self-efficacy) and motivation (motivational intensity) grows, they begin to internalize these successes, gradually reshaping their self-identity as capable and successful learners. This transformation, in turn, feeds back into further engagement and motivation, creating a virtuous cycle of identity reinforcement. Thus, a learner’s self-efficacy and engagement are not only predictors of their language proficiency but are integral to the development of their self-identity as a capable L2 learner. This evolving self-identity influences their long-term motivation and continued engagement in the language-learning process.

2.9. Research Questions

• What are the statistical changes in self-efficacy, motivational intensity, and behavioral engagement after participation in EMI courses?
• To what extent do self-efficacy, motivational intensity, and behavioral engagement interact to influence positive and negative self-identity in EMI learners?

2.10. Research Hypothesis

Hypothesis 1 (H1):

Undergraduates will show a significant improvement in their self-efficacy, demonstrate higher motivational intensity, and experience stronger behavioral engagement after participating in EMI courses.

Hypothesis 2 (H2):

Undergraduates’ self-efficacy, motivational intensity, and behavioral engagement will display beneficial effects on language self-identity after completing EMI courses.

3. Methods

3.1. Participants

390 EFL undergraduates (154 females and 236 males) were recruited to join the current study, with their academic background (

Table 1. Demographic Information of the Participants.

Category	Sub-Category	Number	Percentage
Gender	Male	236	60.5%
	Female	154	39.5%
Age	18 (freshman)	123	31.5%
	19 (sophomore)	192	49.2%
	20 (junior)	45	11.5%
	21 (senior)	30	7.7%
Department	Medicine	171	43.9%
	Biomedical Science	109	27.9%
	Biomedical Informatics	40	10.3%
	Biomedical Imaging and Radiological Science	70	17.9%

) all coming from the College of Medicine (including Departments of Medicine, Biomedical Sciences, Biomedical Informatics, and Biomedical Imaging and Radiological Science). More than 80% of the participants were freshmen and sophomores (315 in total), and 75 were juniors and seniors. Six different courses (Biomedicine Seminar, Biotechnology Seminar, English for Health Care, Human Behavior and Well-being, Medicine & Health, Medicine Laboratory Seminar) were provided with English as a Medium of Instruction (EMI), and, in order to make sure that the course instructors were qualified for EMI, all received the “The Certificate in EMI Skills Online Course” in 2023, which is a self-paced online EMI course developed by the Cambridge Assessment English of the University of Cambridge. Each course covers eight different thematic modules “https://reurl.cc/9Z4MWn (accessed on 1 February 2023)”, with each module comprising approximately five hours of study, totaling 40 h. The pace can be adjusted according to the learner’s speed. Upon completion of the entire course, an electronic certificate of completion will be provided to the learner by the system.

In the fall semester of 2022, there were 14 EMI courses in total; however, as mentioned above, participants coming from 6 different courses provided by the College of Medicine were considered as the criterion for recruitment because of the reason that the College of Medicine had received the research fund supported by the Ministry of Education, making it possible for the instructors to utilize the teaching resources to prepare more solid EMI classes.

3.2. Instruments

In this study, four distinct questionnaires (

Table 2. Four Scales for Current Study.

Variables	Example Items	Items
Self-efficacy	The tasks we completed in EMI classes were manageable for me. My English proficiency was sufficient to meet the demands of this EMI course.	1, 6, 9, 14, 17, 22, 27, 30
Motivational intensity	I am a dedicated learner of the English language. I invest considerable time and effort in improving my weaknesses in English.	2, 10, 18, 23, 31
Behavioral engagement	I actively participate in discussions during EMI classes. I stay focused and attentive during EMI classes.	3, 11, 19, 24, 32
Positive self-identity	I can easily switch between Chinese and English according to situational needs. After learning English, I have become more understanding and can better communicate with others.	4, 7, 12, 15, 20, 25, 28, 33
Negative self-identity	With the improvement of my English proficiency, I feel my Chinese is becoming less idiomatic. I feel weird when my speech in Chinese is subconsciously mixed with English words.	5, 8, 13, 16, 21, 26, 29, 34

) were employed to measure the constructs of Self-efficacy, Motivational Intensity, Behavioral Engagement, and Language Self-identity. Each of these instruments was adapted to align with the context of English as a Medium of Instruction (EMI), ensuring relevance to the research environment and objectives. A 6-point Likert scale ranging from 1 (Strongly Disagree) to 6 (Strongly Agree) was used for all items across the four scales to capture participants’ levels of agreement or perception.

• Self-Efficacy

The questionnaire assessing self-efficacy was adapted from Harris and Strefford (2024), who developed an instrument to evaluate EMI learners’ self-efficacy, engagement, and satisfaction. The original items were rephrased to reflect EMI-specific scenarios, such as learners’ confidence in comprehending EMI course content, completing tasks, and participating actively in EMI settings. This adaptation ensured that the questionnaire accurately captured the participants’ self-efficacy in navigating the challenges of learning English in a predominantly non-native context.

b.

Motivational Intensity

The measure of motivational intensity was based on Feng and Papi’s (2020) study, which examined the role of grit and future self-guides in language learning persistence. Items were modified to focus on the effort, persistence, and dedication of learners in EMI courses. The revised items emphasized the degree of commitment to mastering English and completing course requirements, thereby aligning the scale with the study’s focus on motivational intensity in EMI contexts.

c.

Behavioral Engagement

The behavioral engagement scale was adapted from Skinner et al. (2008), who explored engagement and disaffection in classroom settings as part of a broader motivational dynamic. Items were rephrased to reflect observable behaviors in EMI classrooms, such as participation in discussions, attentiveness during lessons, and consistent effort in completing assignments. These modifications ensured the questionnaire accurately captured the active participation and behavioral commitment of EMI learners.

d.

Language Self-Identity

The language self-identity scale drew on the work of Gao et al. (2007), which explored the relationship between English learning motivation and self-identity changes. Items were tailored to measure the positive (additive & productive) and negative (subtractive & split) self-identities of learners within the EMI context. The rephrased items aimed to capture how learners integrate English into their self-concept while navigating potential identity shifts resulting from EMI experiences.

3.2.1. Adaptation for EMI Context

For all four instruments, the original items were carefully rephrased to reflect the linguistic, cultural, and academic dimensions of EMI. This adaptation process ensured that the scales were contextually appropriate and capable of capturing nuanced aspects of learners’ experiences in EMI courses. By grounding the instruments in the EMI context, the study aimed to provide a robust evaluation of the constructs under investigation.

3.2.2. Scoring and Analysis

The use of a 6-point Likert scale allowed for the collection of detailed data reflecting varying degrees of agreement (from strongly disagree to strongly agree) among participants. This scale also facilitated the differentiation of responses, supporting a more nuanced statistical analysis of the relationships between self-efficacy, motivational intensity, behavioral engagement, and language self-identity in EMI learners.

3.3. Data Collection Procedure

The data collection process in this study was designed to evaluate the development of self-efficacy, motivational intensity, behavioral engagement, and language self-identity among undergraduate learners participating in English as a Medium of Instruction (EMI) courses. Six intact classes, comprising participants from diverse academic backgrounds, were recruited for the study.

In the first week of the semester, all participants were administered a comprehensive questionnaire to assess their baseline levels across the four constructs of interest. This initial phase ensured that the data reflected their pre-EMI experience and provided a reference point for measuring changes throughout the semester.

Following the initial data collection, participants engaged in EMI courses over an 18-week semester, during which English served as the primary language of instruction. These courses aimed to immerse students in English-language learning environments, fostering both linguistic and academic development.

In the final week of the semester (week 18), the same questionnaire was administered to all participants. This post-EMI assessment allowed for a direct comparison of pre- and post-semester data to evaluate changes in self-efficacy, motivational intensity, behavioral engagement, and language self-identity. The repeated administration of the questionnaire ensured consistency and reliability in measuring the developmental trajectories of the participants across the semester.

By collecting data at these two critical time points, the study captured the longitudinal effects of EMI on learner engagement, motivation, and identity, offering valuable insights into the transformative impact of English-medium education. This structured approach also ensured the data’s robustness for subsequent statistical analyses, including the examination of pre- and post-EMI differences and the relationships among the constructs.

3.4. Data Analysis

To answer the research questions effectively, two distinct data analyses will be applied in this study, focusing on assessing the changes in self-efficacy, motivational intensity, and behavioral engagement before and after the EMI courses, and examining the effects of these factors on language self-identity.

In terms of the first research question, the repeated measures of MANOVA (Multivariate Analysis of Variance) were applied to explore whether there are significant improvements in self-efficacy, motivational intensity, and behavioral engagement before and after the EMI courses. This is the appropriate statistical technique given the nature of the study design, where participants are measured on these constructs at two-time points: before the EMI courses (week 1) and after completing the semester-long EMI experience (week 18). Repeated measures MANOVA is ideal because it allows for the examination of multiple dependent variables (self-efficacy, motivational intensity, and behavioral engagement) simultaneously, accounting for the correlation between these variables within each participant. Additionally, MANOVA is capable of testing the main effects of time (pre-EMI vs. post-EMI) on the dependent variables, as well as potential interaction effects between time and the constructs under investigation.

As for the second research question, multiple regression analysis will be implemented to examine the effects of self-efficacy, motivational intensity, and behavioral engagement on language self-identity, specifically focusing on both additive and productive self-identities. Multiple regression allows for the assessment of the independent effects of each predictor variable (self-efficacy, motivational intensity, and behavioral engagement) on the dependent variable (language self-identity) while controlling for potential confounders. Multiple regression analysis is appropriate for answering the second research question because it enables the investigation of the relationships between multiple predictor variables and a single outcome variable (language self-identity). This method helped identify how much of the variance in language self-identity can be explained by these three key factors. The analysis will also assess the relative importance of each predictor, indicating which construct (self-efficacy, motivational intensity, or behavioral engagement) has the strongest influence on the development of language self-identity.

4. Results

4.1. Reliability

The current study explored the use of Cronbach’s alpha (α) and McDonald’s omega (ω) for estimating the reliability of measurement scales (

Table 3. Reliability of Self-efficacy, Motivational intensity, Behavioral engagement, Self-identity.

Category	Number of Item	Cronbach’s Alpha (α)	McDonald’s Omega (ω)
Self-efficacy	8	0.88	0.87
Motivational intensity	5	0.86	0.83
Behavioral engagement	5	0.92	0.89
Positive self-identity	8	0.79	0.77
Negative self-identity	8	0.73	0.71

). Cronbach’s alpha is widely used due to its simplicity and historical popularity, but it has limitations, particularly under violations of assumptions like essential tau-equivalence. Omega is recommended as a superior alternative because:

• General Applicability: Omega doesn’t require the strict assumption of equal factor loadings (essential tau-equivalence).
• More Accurate Reliability Estimates: Research shows omega provides better reliability estimates when item loadings vary.
• Modern Computational Options: Omega can now be computed without needing advanced software like confirmatory factor analysis (CFA), thanks to macros available for SPSS 29 and SAS 9.4.

The reliability analysis revealed strong internal consistency across all measurement scales. The behavioral engagement scale demonstrated the highest reliability with Cronbach’s alpha at 0.92 and McDonald’s omega at 0.89, indicating robust measurement of students’ participation and task completion in EMI courses. The self-efficacy scale showed strong reliability (α = 0.88, ω = 0.87), followed by the motivational intensity scale (α = 0.86, ω = 0.83). The language self-identity subscales also demonstrated acceptable reliability, with positive self-identity (α = 0.79, ω = 0.77) and negative self-identity (α = 0.73, ω = 0.71) both meeting threshold requirements. The close alignment between Cronbach’s alpha and McDonald’s omega coefficients across all scales provides additional evidence of measurement consistency. These reliability results confirm that all scales offer dependable measurements of the constructs under investigation, making them suitable for evaluating EMI’s effects on learners’ psychological and academic development.

4.2. Repeated Measure of MANOVA

The study aimed to examine the improvement in self-efficacy, motivational intensity, and behavioral engagement of EFL undergraduates in Taiwan after participating in a semester-long English Medium Instruction (EMI) intervention. A repeated measures MANOVA was employed to evaluate these variables collectively and individually to understand the impact of EMI on learners’ development.

In terms of multivariate analysis (

Table 4. Summary of Multivariate Test Results.

Effect	Test Statistic	F	df	p-Value	Partial η²
Multivariate (Wilks’ Lambda)	Λ = 0.65	58.75	(3,387)	<0.001	0.35

), the results from the multivariate test using Wilks’ Lambda (Λ = 0.65) indicated a statistically significant overall effect of time (pre-EMI vs. post-EMI) on the combined dependent variables (F(3,387) = 58.75, p < 0.001, η² = 0.35). This result shows that the EMI intervention significantly influenced the participants’ self-efficacy, motivational intensity, and behavioral engagement collectively, with a large effect size. The data underscore the EMI’s ability to drive multidimensional improvements in learner engagement and motivation.

As for the univariate tests (

Table 5. Within-Subjects Effects.

Variable	F	df	p-Value	Partial η²
Self-efficacy	100.25	(1,389)	<0.001	0.28
Motivational intensity	120.5	(1,389)	<0.001	0.30
Behavioral engagement	135.8	(1,389)	<0.001	0.35

), the results provided further insights into the specific contributions of each variable to the overall improvement. For self-efficacy, the mean increased from M = 3.85 (SD = 0.78) pre-EMI to M = 4.40 (SD = 0.70) post-EMI. This change was statistically significant (F(1,389) = 100.25, p < 0.001, η² = 0.28), indicating a moderate effect size. The findings highlight that EMI courses bolstered students’ confidence in their ability to navigate and succeed in English-based academic settings.

Motivational intensity also showed significant improvement, with mean scores increasing from M = 4.12 (SD = 0.81) pre-EMI to M = 4.75 (SD = 0.74) post-EMI. The results (F(1,389) = 120.5, p < 0.001, η² = 0.30) reflect a slightly higher effect size compared to self-efficacy. This suggests that the EMI intervention effectively energized students’ dedication and persistence toward achieving language and content-related goals.

The largest improvement was observed in behavioral engagement, where the mean increased from M = 4.02 (SD = 0.77) pre-EMI to M = 4.71 (SD = 0.68) post-EMI. The change was statistically significant (F(1,389) = 135.8, p < 0.001, η² = 0.35), with the highest effect size among the three variables. This demonstrates that EMI courses provided an environment that strongly encouraged active participation, attention, and effort in academic activities.

The descriptive statistics (

Table 6. Descriptive Statistics.

Variables	Time	Mean	SD	95% CI (Lower)	95% CI (Upper)
Self-efficacy	Pre-EMI	3.85	0.78	3.75	3.95
	Post-EMI	4.40	0.70	4.3	4.5
Motivational intensity	Pre-EMI	4.12	0.81	4.01	4.23
	Post-EMI	4.75	0.74	4.65	4.85
Behavioral engagement	Pre-EMI	4.02	0.77	3.91	4.13
	Post-EMI	4.71	0.68	4.6	4.8

) further reinforced these findings. For all three variables, the 95% confidence intervals of the post-EMI means were higher than their pre-EMI counterparts, indicating consistent and meaningful improvement. For instance, the 95% CI for post-EMI self-efficacy (4.3, 4.5, 4.3, 4.5) does not overlap with the pre-EMI CI (3.75, 3.95), reflecting a clear enhancement due to the EMI intervention.

In short, the results (Figure 1) reveal that EMI courses significantly improved all three dimensions of learner development. The largest gains in behavioral engagement suggest that the interactive and participatory nature of EMI courses effectively stimulated students to become more involved in their learning processes. Motivational intensity, which reflects effort and persistence, was also significantly enhanced, indicating that EMI fostered a goal-oriented and resilient attitude among learners. Finally, self-efficacy, though improving the least compared to the other variables, still demonstrated a meaningful increase, suggesting that learners felt more confident in their ability to meet the demands of EMI courses. These findings underscore the transformative potential of EMI as a pedagogical approach. By immersing learners in English-based academic environments, EMI not only facilitates language acquisition but also promotes psychological and behavioral readiness for global academic and professional challenges. Future research could explore the long-term sustainability of these improvements and their impact on learners’ broader academic trajectories.

4.3. Multiple Regression

The multiple regression analysis aimed to examine how the three independent variables—Self-Efficacy, Motivational Intensity, and Behavioral Engagement—influence the two dependent variables, Positive Self-Identity, and Negative Self-Identity, in the context of English as a Medium of Instruction (EMI). The results, presented in Table 6,

Table 7. Model Summary for Multiple Regression.

Dependent Variable	R2	Adjusted R2	F	df	p-Value	Standard Error
Positive Self-Identity	0.50	0.49	60.45	(3,386)	<0.0001	0.58
Negative Self-Identity	0.38	0.37	45.32	(3,386)	<0.0001	0.64

and

Table 8. Coefficients Table.

Dependent Variable	Predictor	B	β	t	p-Value	95% CI (Lower)	95% CI (Upper)
Positive Self-Identity	Behavioral Engagement	0.5	0.38	8.95	<0.001	0.4	0.3
Positive Self-Identity	Motivational Intensity	0.4	0.32	7.34	<0.001	0.3	0.5
Positive Self-Identity	Self-Efficacy	0.35	0.28	6.2	<0.001	0.25	0.45
Negative Self-Identity	Behavioral Engagement	−0.3	−5.78	−6.72	<0.001	−0.4	−0.2
Negative Self-Identity	Motivational Intensity	−0.2	−4.32	−5.48	<0.001	−0.3	−0.1
Negative Self-Identity	Self-Efficacy	−0.15	−4.36	−4.36	<0.001	−0.25	−0.05

, provide key insights into these relationships, highlighting the differential impacts of the predictors on self-identity constructs.

4.4. Overall Model Fit

As shown in Table 7, the multiple regression models for both dependent variables were statistically significant, indicating that the independent variables collectively explained substantial variance in self-identity outcomes. For Positive Self-Identity, the model accounted for 50% of the variance (R² = 0.50, Adjusted R² = 0.49), with an F-statistic of 60.45 (p < 0.0001), and a standard error of 0.58. This result underscores the strong predictive power of the independent variables on positive self-identity. Similarly, for Negative Self-Identity, the model explained 38% of the variance (R² = 0.38, Adjusted R² = 0.37), with an F-statistic of 45.32 (p < 0.0001) and a standard error of 0.64, highlighting a moderately strong relationship.

4.5. Effects on Positive Self-Identity

Among the predictors (Table 8), Behavioral Engagement emerged as the strongest predictor of positive self-identity, with a standardized coefficient (β) of 0.38 (t = 8.95, p < 0.001). This finding indicates that students who actively participate, stay focused, and consistently engage in their EMI courses are more likely to develop a positive self-identity, reflecting the significant role of behavioral factors in fostering self-perception. Motivational Intensity was the second strongest predictor, with β = 0.32 (t = 7.34, p < 0.001). The result suggests that effort, persistence, and dedication toward achieving learning goals positively influence students’ self-concept in an EMI setting. Self-Efficacy also had a meaningful impact, with β = 0.28 (t = 6.20, p < 0.001), indicating that students’ confidence in their ability to complete EMI tasks and succeed academically enhances their positive self-identity, though to a slightly lesser extent compared to the other predictors.

4.6. Effects on Negative Self-Identity

For negative self-identity, the results (Table 8) showed an inverse relationship with all three predictors. Behavioral Engagement again had the strongest effect, with β = −0.28 (t = −6.72, p < 0.001), demonstrating that higher engagement reduces feelings of conflict or negative perceptions associated with adopting English-medium learning. Motivational Intensity had a standardized coefficient of β = −0.22 (t = −5.48, p < 0.001), showing that greater motivation decreases the likelihood of negative identity outcomes. Self-Efficacy had the smallest negative association, with β = −0.18 (t = −4.36, p < 0.001), indicating that students who believe in their ability to succeed are less likely to experience identity-related conflicts or insecurities.

4.7. Diagnostics and Multicollinearity

Diagnostics from

Table 9. Diagnostics Table.

	VIF	Tolerance
Behavioral engagement	1.25	0.8
Motivational intensity	1.2	0.83
Self-efficacy	1.18	0.85

revealed no significant multicollinearity concerns among the predictors, with Variance Inflation Factors (VIF) ranging from 1.18 to 1.25 and tolerance values exceeding 0.80. These results affirm the independence of the predictors and the robustness of the regression models.

As can be seen in Figure 2, the results highlight the critical role of Behavioral Engagement as the most influential factor for both positive and negative self-identity. This underscores the importance of fostering active participation and sustained effort in EMI courses to enhance students’ self-concept and mitigate identity-related challenges. Motivational Intensity, reflecting goal-directed effort, also plays a significant role, reinforcing the value of cultivating persistence and dedication in EMI contexts. While Self-Efficacy had the smallest effect among the three predictors, its consistent and statistically significant impact indicates that building students’ confidence is essential for promoting positive self-identity and reducing negative identity outcomes. These findings suggest that EMI courses should prioritize interactive and participatory teaching strategies to maximize behavioral engagement. Additionally, educators should design activities that enhance students’ intrinsic and extrinsic motivations while providing scaffolding to build their self-efficacy. Collectively, these strategies can help learners develop a stronger, more positive self-identity while minimizing potential identity conflicts in an EMI environment. Future research could explore the long-term effects of these predictors and the interplay among them to better understand their cumulative impact on self-identity development.

5. Discussion

The results of the study provide compelling evidence that English Medium Instruction (EMI) significantly enhances the self-efficacy, motivational intensity, and behavioral engagement of EFL undergraduates in Taiwan. These findings align with previous research emphasizing the multifaceted benefits of EMI in fostering language development and learner engagement. This discussion interprets the results in the context of the first research question: How do EFL undergraduates in Taiwan improve their self-efficacy, motivational intensity, and behavioral engagement during and after participating in one semester of EMI courses?

5.1. Improvement in Behavioral Engagement

The most substantial improvement was observed in behavioral engagement, which exhibited the largest effect size (η² = 0.35). This result suggests that EMI courses create a highly interactive and participatory environment, encouraging learners to actively engage in academic tasks. The structured and immersive nature of EMI likely promotes behaviors such as attentiveness, participation in discussions, and consistent effort in completing assignments. These behaviors are critical for academic success and align with prior studies indicating that engagement is a key predictor of learning outcomes in EMI contexts. Moreover, the heightened behavioral engagement may reflect a feedback loop, where successful task completion in English fosters confidence and further participation.

5.2. Growth in Motivational Intensity

Motivational intensity also showed a significant increase, with a slightly lower effect size (η² = 0.30) compared to behavioral engagement. This finding underscores the role of EMI in instilling a sense of purpose and dedication among learners. By blending content learning with language acquisition, EMI courses likely appeal to both intrinsic and extrinsic motivations. Students may feel driven by the immediate challenge of mastering content in English as well as the long-term benefits of enhanced English proficiency for academic and professional opportunities. The results resonate with prior research suggesting that motivational intensity is closely linked to persistence and resilience, traits that are critical in overcoming the challenges posed by EMI.

5.3. Growth in Self-Efficacy

Gains in Self-Efficacy Although self-efficacy showed the smallest improvement among the three variables, the change was still statistically significant (η² = 0.28). This indicates that EMI courses positively influence learners’ confidence in their ability to succeed in an English-speaking academic environment. The moderate gains in self-efficacy may be attributed to the supportive structures of EMI, such as scaffolded instruction, peer collaboration, and exposure to real-world academic tasks in English. However, the relatively smaller effect size might reflect the complexity of building self-efficacy, which often requires consistent reinforcement over time. Learners’ initial apprehensions about their English proficiency could have moderated the extent of improvement in this domain.

The collective improvement across all three variables highlights the comprehensive benefits of EMI as a pedagogical approach. Behavioral engagement, as the most enhanced dimension, serves as a foundation for sustaining motivation and building self-efficacy. Motivational intensity complements this by providing the drive necessary to engage deeply with academic tasks. Self-efficacy, while improving to a lesser degree, remains a crucial element, as it fosters learners’ belief in their ability to succeed. Together, these dimensions form a triad of interconnected factors that contribute to effective learning in EMI settings. From a pedagogical perspective, the findings suggest that EMI can be strategically leveraged to maximize student engagement and motivation. Educators should focus on creating dynamic and interactive classroom environments that promote active participation. Additionally, embedding motivational elements, such as clear learning goals and real-world applications of English, can further enhance learners’ commitment and persistence. To address the relatively smaller gains in self-efficacy, instructors may consider implementing targeted interventions, such as personalized feedback and incremental challenges, to gradually build learners’ confidence.

The results of the multiple regression analysis provide robust evidence of the associations between the independent variables—Self-Efficacy, Motivational Intensity, and Behavioral Engagement—and the dependent variables—Positive Self-Identity and Negative Self-Identity—in the context of English Medium Instruction (EMI). This discussion addresses the second research question: What are the associations between self-efficacy, motivational intensity, behavioral engagement, and the positive and negative self-identity of EMI learners?

5.4. Associations with Positive Self-Identity

The findings underscore a strong positive association between the predictors and Positive Self-Identity, with the model explaining 50% of the variance (R² = 0.50). Among the predictors, Behavioral Engagement emerged as the most influential factor (β = 0.38, p < 0.001). This suggests that students who are actively engaged in class activities, participate in discussions, and remain attentive are more likely to perceive themselves positively in the EMI learning context. Behavioral engagement not only reflects active involvement but also serves as a mechanism through which students validate their competence and identity in an English-speaking academic environment.

Motivational Intensity, the second strongest predictor (β = 0.32, p < 0.001), also exhibited a significant association with positive self-identity. This indicates that students’ persistence and dedication to mastering English and meeting course requirements directly enhance their self-perception. High motivation likely sustains students’ efforts, allowing them to overcome challenges, which in turn reinforces their positive self-identity.

Self-Efficacy, while the weakest predictor (β = 0.28, p < 0.001), still contributed meaningfully to positive self-identity. Students with greater confidence in their ability to succeed in EMI courses are more likely to see themselves as competent and effective learners. Although its effect was smaller than behavioral engagement and motivational intensity, self-efficacy serves as a foundational belief system that shapes students’ willingness to take on and persist in challenging tasks.

The associations between the predictors and Negative Self-Identity revealed inverse effects, with the model explaining 38% of the variance (R² = 0.38). Behavioral Engagement was again the strongest predictor (β = −0.28, p < 0.001), indicating that active and sustained engagement in EMI courses significantly reduces the likelihood of negative self-perceptions. By participating in academic activities, students may feel a stronger sense of belonging and competence, mitigating the sense of identity conflict or inadequacy.

Motivational Intensity also demonstrated a meaningful inverse relationship with negative self-identity (β = −0.22, p < 0.001). Students who are motivated to persist and succeed in EMI courses are less likely to experience negative identity outcomes. This could be attributed to the alignment of their effort and goal orientation, which helps them contextualize and embrace the challenges of EMI rather than internalizing them as deficits.

Self-Efficacy had the weakest effect among the predictors (β = −0.18, p < 0.001), but its role remains significant. Students who believe in their ability to understand and complete EMI tasks are less prone to feelings of insecurity or identity dissonance. While its contribution to reducing negative self-identity was smaller than behavioral engagement and motivational intensity, self-efficacy acts as a protective factor that buffers students against adverse identity outcomes.

The differential impacts of the predictors on positive and negative self-identity highlight the multidimensional nature of self-identity development in EMI contexts. Behavioral Engagement consistently emerged as the strongest predictor for both outcomes, underscoring its central role in shaping students’ identity. Active participation not only fosters positive self-perceptions but also reduces negative self-conceptualizations, demonstrating the dual benefits of engagement in EMI environments. Motivational Intensity, while slightly less impactful, plays a crucial role in sustaining effort and persistence, thereby enhancing positive identity and mitigating negative identity. Self-Efficacy, though less influential than the other predictors, provides an essential baseline of confidence that supports students’ ability to engage and stay motivated.

5.5. Implications for EMI Pedagogy

These findings emphasize the importance of fostering behavioral engagement in EMI courses to maximize identity benefits for learners. Educators can achieve this by incorporating interactive and participatory teaching methods, such as group discussions, problem-solving tasks, and active learning strategies, that encourage student involvement. To enhance motivational intensity, instructors should design courses with clear goals, meaningful tasks, and opportunities for students to see the real-world relevance of their efforts. Finally, building students’ self-efficacy through scaffolding, timely feedback and manageable challenges can provide the confidence needed to sustain engagement and motivation.

6. Conclusions

This study investigates the dynamic relationships between self-efficacy, motivational intensity, and behavioral engagement and their effects on positive and negative self-identity in the context of English as a Medium of Instruction (EMI). The findings provide compelling evidence of the transformative potential of EMI in enhancing learners’ psychological and academic dimensions. The EMI intervention significantly improved self-efficacy, motivational intensity, and behavioral engagement among Taiwanese EFL undergraduates. Behavioral engagement exhibited the most substantial improvement, highlighting the critical role of active participation and attentiveness in fostering effective learning environments. Motivational intensity also increased significantly, reflecting the pivotal influence of effort and persistence in maintaining focus on language acquisition and academic goals. Although self-efficacy demonstrated the smallest improvement, its enhancement suggests that EMI provides an environment conducive to building confidence in navigating academic tasks in English. The multiple regression analysis underscores the critical role of the three predictors—behavioral engagement, motivational intensity, and self-efficacy—in shaping learners’ self-identity. Behavioral engagement emerged as the strongest predictor for both positive and negative self-identity, signifying the importance of active participation in strengthening learners’ positive self-concept while mitigating identity conflicts. Motivational intensity, while slightly less influential, played a crucial role in sustaining learners’ effort and goal orientation, enhancing positive self-identity, and reducing negative self-perceptions. Self-efficacy, though less impactful than the other predictors, remained a foundational element, fostering confidence and resilience in EMI contexts.

Long-Term Implications

The longitudinal implications of English Medium Instruction (EMI) extend beyond immediate educational outcomes, encompassing significant ramifications for learners’ academic development and professional trajectories. This paradigm necessitates comprehensive investigation across three critical dimensions. First, longitudinal studies examining identity development among EMI learners are essential to understanding how engagement, motivation, and self-efficacy evolve throughout academic progression and post-graduation periods. Second, the correlation between EMI participation and professional readiness demands thorough examination, particularly regarding employability, workplace communication competencies, and professional self-identity formation in global contexts. This investigation is crucial given EMI’s positioning as a preparatory framework for international labor markets. Third, the role of individual differences in EMI outcomes requires systematic exploration, considering how diverse linguistic proficiency levels, learning strategies, and socio-cultural factors influence learner engagement and achievement. Understanding these variations is fundamental for developing inclusive pedagogical approaches that accommodate diverse learning needs. These research directions collectively underscore the necessity of adopting a multifaceted approach to investigating EMI’s long-term impact, ultimately informing evidence-based curriculum design and policy development in international education contexts.