1 The impact of English-medium instruction on university student performance Title: The impact of English-medium instruction on university student performance Date: Pre-Print February 2022 Arroyo-Barrigüete, J. L., López-Sánchez, J. I., Morales-Contreras, M. F., & Soffritti, M. (2024). The impact of English-medium instruction on university student performance. JOURNAL OF MULTILINGUAL AND MULTICULTURAL DEVELOPMENT, 45(6), 2315–2330. https://doi.org/10.1080/01434632.2022.2047193 Authors: Jose Luis Arroyo-Barrigüetea, Jose Ignacio López-Sánchezb, Manuel Francisco Morales- Contrerasc,d, Mirco Soffrittie. a Universidad Pontificia Comillas, Department of Quantitative Methods. Alberto Aguilera 23, 28015 Madrid, jlarroyo@comillas.edu. Spain. ORCID: https://orcid.org/0000-0002- 3660-3933 b Universidad Complutense de Madrid, Department of Management and Marketing. Campus de Somosaguas, 28223 Pozuelo de Alarcón (Madrid), Spain. jilopez@ucm.es. ORCID: https://orcid.org/0000-0002-4096-0992 c Corresponding Author. mfcontreras@comillas.edu d Universidad Pontificia Comillas, Department of Management, ICADE; Institute for Research in Technology (IIT), ICAI School of Engineering. Alberto Aguilera 23, 28015 Madrid, Spain. mfcontreras@comillas.edu. ORCID: https://orcid.org/0000-0002-1389- 3798 e IE Business School, Department of Operations & Decision Sciences. mirco.soffritti@gmail.com. ORCID: https://orcid.org/0000-0001-7593-2484 . https://doi.org/10.1080/01434632.2022.2047193 mailto:jlarroyo@comillas.edu https://orcid.org/0000-0002-3660-3933 https://orcid.org/0000-0002-3660-3933 mailto:jilopez@ucm.es https://orcid.org/0000-0002-4096-0992 mailto:mfcontreras@comillas.edu mailto:mfcontreras@comillas.edu https://orcid.org/0000-0002-1389-3798 https://orcid.org/0000-0002-1389-3798 mailto:mirco.soffritti@gmail.com https://orcid.org/0000-0001-7593-2484 2 The impact of English-medium instruction on university student performance Abstract During the last two decades universities around the world have increased the adoption of English-medium instruction (EMI) as a way to enhance internationalization and global competitiveness. EMI adoption presents a wide range of opportunities, but it also presents some challenges, being one of them the potential impact on students’ academic performance. This paper analyzes the impact of EMI on the academic performance of the students in a Spanish university. The objective is to extend previous research, that shows contradictory conclusions. In the first part of the paper, using a multiple linear regression model to control key covariates, we have compared the performance of 229 EMI Vs 635 Non-EMI students, corresponding to cohorts 2013-14 to 2017-18, considering the average grade in the 10 subjects of the first course. In the second part, we focus on the 2017-18 cohort (49 EMI Vs 116 Non-EMI students), carrying out a longitudinal study of its behavior during two academic years in four different subjects. The results show that there are no statistically significant differences in academic performance between EMI and non-EMI students, i.e., language of instruction does not play a relevant role in academic performance. Keywords. English-medium instruction, academic performance, higher education, business studies. Funding details. This work was supported by the Ministerio de Ciencia e Innovación de España under Grant ABANRED2020 Proyectos I+D+i «Retos Investigación» 2020 (Ref. PID2020- 116293RB-I00); and Santander - Universidad Complutense de Madrid 2020 under Grant Ref. PR108/20-10. Disclosure statement. The authors report there are no competing interests to declare. 3 Introduction Higher Education (HE) institutions are confronted to highly diverse and sometimes contradicting expectations coming from the students and societies: preparing the students for global challenges, helping them to achieve benefits and opportunities in the global labor market, contributing to the countries’ economic competitiveness and human resources development, enhancing the quality of research and education, and serving the society (Gregersen-Hermans & Lauridsen, 2021: p.21). Internationalization of HE is a broad, multifaceted, and varied concept, with different interpretations depending on the geographical area, disciplines or institutions (Green & Whitsed, 2014, p.5; De Wit et al., 2015, p.27; Gregersen-Hermans & Lauridsen, 2021, p.21), a central force in higher education, though its contours are unclear” (Altbach & Knight, 2007, p.303). It could be defined as “the intentional process of integrating an international, intercultural or global dimension into the purpose, functions and delivery of post-secondary education, in order to enhance the quality of education and research for all students and staff, and to make a meaningful contribution to society” (De Wit et al., 2015, p.29). We are living in a global world; thus, internationalization becomes a key driver for the development of universities (MECD, 2016), and it is commonly seen as an institutional strategy to fulfill the above- mentioned expectations (Gregersen-Hermans & Lauridsen, 2021, p.21). Academic literature offers different perspectives and analysis of this research topic, sometimes around specific areas as Business, Education or Health (Green & Whitsed, 2014), presenting cases and examples from countries in Europe, America, Asia, Africa and Oceania (Green & Whitsed, 2014; De Wit et al., 2015; Bowles & Murphy, 2020; Gregersen- Hermans & Lauridsen, 2021). 4 Internationalization of HE pursues some objectives, such as the enhancing the institution reputation (often measured by rankings), its visibility and competitiveness; the attraction and retention of talented students and scholars; economic profitability; demographic considerations; and a focus on students’ employability, productivity, earning power, increasing their experience and knowledge, and social engagement (European Commission, 2013; De Wit et al., 2015, p.27). A number of different universities worldwide have designed and implemented internationalization plans with the objectives of offering an international experience to different stakeholders, such as students, professors and staff, but a comprehensive internationalization strategy should be designed around the following three pillars: international student, professors, and staff mobility; the internationalization at home, and improvement of curricula and digital learning; and strategic cooperation, partnerships and capacity building (European Commission, 2013). For some decades, mobility, i.e., sending students or professors abroad or hosting foreign students or professors at home (MECD, 2016; Yang, Volet, & Mansfield, 2017; Zhang, 2018) has played a relevant role, especially in Europe, thanks to programs as ERASMUS or Marie Curie Fellowships. But the vast majority of learners are not mobile (Wächter, 2003), representing 80-90% according to European Commission (2013). Thus, new initiatives and approaches are necessary, such as internationalization at home (IaH) (Nilsson, 2003; Wächter, 2003; European Commission, 2013; De Wit et al., 2015) and internationalization of the curriculum (IoC) (Leask, 2009; Leask, 2013; Green & Whitsed, 2014; Gregersen-Hermans & Lauridsen, 2021). IaH has been defined as “any international activity with the exception of outbound student and staff mobility”, whose two main pillars are “an understanding of 5 internationalization that went beyond mobility and a strong emphasis on the teaching and learning in a culturally diverse setting” (Wächter, 2003), which requires an IoC (Leask, 2009). IoC was defined as “the incorporation of an international and intercultural dimension into the content of the curriculum as well as the teaching and learning arrangements and support services of a program of study” (Leask, 2009, p. 209). Leask (2013) proposes a five-stage model for the IoC, and states that it should be a planned, developmental, and cyclical process. Creating an internationalized learning opportunity for all students requires careful consideration and specific action, mainly from the upper and middle management teams responsible for education and study programs (Lauridsen, 2020). According to Green & Whitsed (2014) there is a convergence of the notions IaH, IoC and ‘comprehensive internationalization’, this last term defined as “a commitment, confirmed through action, to infuse international and comparative perspectives throughout the teaching, research and service mission of higher education” Hudzik (2011, p. 6). English as a Medium of Instruction (EMI) is directly related to IaH and IoC, as “the language of instruction is almost universally English” (Bowles & Murphy, 2020, p.8), with an exponential growth of English-taught programs around the world. “Other foreign languages [different than English] are virtually invisible in university policy documents” (Doiz & Lasagabaster, 2021, p.1, citing Kuteeva, Kaufhold & Hynninen, 2020). EMI application is in clear expansion in many universities (Costa & Coleman, 2013; Hultgren, 2014; Broggini & Costa, 2017), and offers enhanced academic and professional opportunities to compete with international students due to several reasons. Firstly, it helps the students to be prepared to meet the requirements of the global academic and labor markets (Beelen, 2011; Kim & Shin, 2014). As Coleman (2006: 11) points out, “the world will become 6 diglossic, with one language for local communication, culture and expression of identity, and another – English – for wider and more formal communication […]”. Secondly, offering EMI in the university has the potential of developing students’ sensitivity and capacity for appreciating cultural differences, as well as developing a sense of global citizenry. In fact, some authors consider that EMI in higher education (HE) not only could help to improve English language proficiency (Wu, 2006; Chang, 2010; Tatzl, 2011; Rose, Curle, Aizawa & Thompson, 2019), but also to develop internationally minded students, professors and staff members. Finally, promoting the use of English in HE supports the initiatives to enhance international visibility for the institution (Dafouz, 2018) by attracting foreign talents (Bowles & Murphy, 2020, p.8) and enhancing their connections within the international academic network. English is also the “the world language of academia” (Mauranen, Hynninen, & Ranta, 2010: 183), as it is commonly used in international conferences, journals and publications. Despite the above-mentioned opportunities of the use of EMI in HE, it could also be problematic from the points of view of the quality of education (English language proficiency, or training and mindset of lecturers) and the thread or marginalization of national languages and cultures (Bowles & Murphy, 2020, p.8-9). There are other challenges that will be addressed in the following section. One of these challenges refers to the potential impact of using a foreign language as English in the academic performance of the students. The research question that governs this study refers to the analysis of the differences in academic performance of university students enrolled in the degree of Business Administration in a Spanish university, depending on the language of instruction. This article is structured as follows: after the introduction, a literature review section reviews 7 the different challenges of EMI in HE, as well as the academic performance in EMI. Then, a methodology section presents the research methods used to conduct this study. Then, results are presented and discussed, as well as the research limitations; and finally, conclusions have been summarized. Literature Review Challenges of EMI in HE As a result of a study on HE in Europe in a Global Setting, the Ministers of Education of 47 countries signed in 2007 an agreement for the External Dimension of the Bologna Process, and a Mobility Strategy for Better Learning 2020. In addition, the Spanish government published an internationalization strategy for the Spanish universities 2015-2020 whose main objective was “consolidating a strong and internationally attractive university system that promotes the mobility of entry and exit of the best students, professors, researchers and administration and service personnel” (MECD, 2016, p.7). One key element was to increase the number of degrees offered in English, since one of the main weaknesses of the Spanish university system was precisely the "low rate of education taught in English and other foreign languages" (MECD, 2016, p.20). Global figures show a significant increase in the number of English-taught courses worldwide (Dearden 2015; Macaro et al. 2018). However, according to the report available from CRUE in the 2016/17 academic year, public universities in Spain with on-site courses offered only 7.5% of their degrees in a bilingual modality, a percentage that rose to 17.8% in the case of private universities (CRUE, 2018, p.115). In other words, HE in the Spanish university environment remains mostly monolingual, despite the above-mentioned advantages of EMI instruction. This has 8 also been reported in other countries as China, where limited EMI courses or programs are offered, and asymmetrical number of international degree students (Zhang, 2018). Nevertheless, it is important to highlight the fact that although the predominant view is that learning through a foreign language brings opportunities that will eventually payoff in a student’s professional career, not everyone seems to share such a perspective, and as Hultgren, Jensen & Dimova (2015:2) point out “in some corners of Europe, it has been met with fierce resistance”. Some researchers view EMI as a requirement that complicates the learning process (Airey, 2004; AlBakri, 2017), and, as Macaro, Curle, Pun, An, & Dearden (2018: 36) point out, “key stakeholders have serious concerns regarding the introduction and implementation of EMI despite sometimes recognizing its inevitability”. In fact, the use of EMI in HE presents different challenges that have been extensively discussed within the academic world, and they refer to the thread or marginalization of national languages and cultures and the quality of education (Bowles & Murphy, 2020, p.8-9). This thread of the multicultural perspective has been studied and discussed in different countries, such as Japan (Le Phan 2013; Tsuneyoshi, 2018; Aizawa & McKinley, 2020), South Korea (Piller and Cho, 2013; Williams, 2015), Turkey (Aslan, 2018), the Netherlands (Edwards, 2016; Gabriëls & Wilkinson, 2020) or Italy (Pulcini & Campagna, 2015; Costa & Mariotti, 2020). This contrasts with research carried out in Denmark (Haines, Kroese and Guo, 2020) where the adoption of English helps the integration and inclusion processes of international students into the learning community within the university. The quality of education is also considered a problematic issue about EMI adoption, and it refers to the English language proficiency, as well as the training and mindset of lecturers. Lack of language skills and competence conduct to “poor quality of communication and 9 interaction in the classroom, and limited learning efficiency” (Bowles & Murphy, 2020, p.9), becoming a worrying issue that has been reported by students or academic staff in research conducted in countries such as Indonesia (Pritasari et al., 2019), Japan (Aizawa & McKinley, 2020), South Korea (Lee & Lee, 2018), China (Wang, 2020), Malaysia (Kaur, 2020), and Ethiopia (Murphy & Mengistu, 2020), among others. Language policies at institutional and national level become necessary to address the above-mentioned challenges about quality of education and multicultural aspects. Moreover, there seems to be evidences about a “recurring and problematic gap between policies and actual practice, as well as a general lack of clarity within existing policies” (Bowles & Murphy, 2020, p.10). As Doiz & Lasagabaster (2021, p.1-2) point out “language learning objectives are overlooked in most higher education institutions”, and this is mainly due to three reasons: the different priorities and urgencies in EMI programs implementations; the assumption that language learning will be achieved just by implementing EMI; and the fact that “content teachers are reluctant to teach language in their classes” as they are not prepared “to deal with language problems”. Doiz & Lasagabaster (2021) present an overview of best practices in implementing EMI teaching experiences in the classroom about language issues, consciously or unconsciously, with the objectives of enhancing students’ EMI experience and reducing the barriers that students might find while learning content in a different language than their own. Thus, differences in use of EMI have been noted, and the vary depending on the country, due to different national language policies, implementation strategies and teaching traditions, or because of different levels of English proficiency among university students, lecturers and management in general (Costa & Coleman, 2013; Broggini & Costa, 2017; 10 Pritasari et al., 2019). As Dafouz, Camacho-Millano & Urquia (2014: 225) point out, three major concerns arise: “teacher and student views of EMI instruction, student foreign language proficiency and student academic competence”. We intend to focus our research in this third concern, as there is a lack of consensus in the related literature. For instance, the students’ limited language academic skills could be a barrier for an adequate implementation of EMI (Lee & Lee, 2018; Pritasari et al., 2019), provoking the frustration of lecturers involved in EMI university education (Hernandez-Nanclares & Jimenez-Munoz, 2015); thus, there could be a risk of producing a lack or reduction of course content comprehension (Chang, 2010; Byun, Chu, Kim, Park, Kim, & Jung, 2011; Tatzl, 2011). In fact, teacher and student attitudes have been broadly researched, and the findings can be grouped in three main areas, according to how positively they view their EMI experience. Some researchers, while recognizing the difficulties associated with EMI instruction, conclude that the advantages outweigh the inconveniences (Wu, 2006; Chang, 2010; Lasagabaster & Sierra, 2011; Tatzl, 2011; Costa & Coleman, 2013; Rose, Curle, Aizawa & Thompson, 2019; Rowland & Murray, 2020; Curle et al., 2020; Xie & Curle, 2020). A second group has an ambiguous position, where is not clear whether positive issues overcome the negative ones (Kim, 2011; Jensen & Thøgersen, 2011; Jiang, Jun Zhang & May, 2016; Costa & Mariotti, 2017). Finally, a third group of papers reports mainly negative perceptions, and often point to the difficulties associated with learning content in a different language other than L1 (Kim, Kweon & Kim, 2017; Belhiah & Elhami, 2015). The question is to what extent these perceptions on the part of students and teachers are really founded, that is, whether EMI instruction really has a negative impact on student performance, understanding this performance as being linked to the acquisition of specific 11 knowledge about different subjects. In this regard, several investigations have concluded that there are no significant differences in the academic performance of students who follow EMI instruction (Dafouz, Camacho-Millano, & Urquía, 2014; Dafouz & Camacho- Millano, 2016), while others point out that, paradoxically, EMI students may perform better than their Non-EMI peers (Hernandez-Nanclares & Jimenez-Munoz, 2015; Del Campo, Cancer, Pascual-Ezama, & Urquía-Grande, 2015). However, as it will be discussed next, these papers present some limitations, reason why the objective of this investigation is to try to overcome them through a quasi-experiment in which all relevant covariates have been controlled. Academic performance in EMI Doing a university course through EMI does not necessarily mean that students learn disciplines better or more thoroughly than they would in their own language. Academic performance is a difficult concept to define, since it can be done according to a number of different approaches. We will focus only on the immediate performance, where performance metrics will be based on some averaging measure of the students obtained grades. These metrics can be accepted as a first approximation although they are not perfect (Di Gresia, Porto, & Ripani, 2002), and this approach is adopted by the majority of the authors who analyze academic performance in terms of the language of instruction. In a recent paper Lin & He (2019) tested the relationship between the achieved academic performance and the language chosen as a medium of instruction (Chinese versus English), analyzing a sample of 498 Chinese students enrolled in a first-year course during the 2015 academic year. After discounting the obvious cultural and academic differences, the authors argued that the association between academic scores and bilingualism turns 12 out to be a statistically insignificant phenomenon. We believe that the limitations in Lin and He’s paper can be found in the fact that the students taking part in their experiment did not actually experience EMI education but bilingual instruction. The difference, as the authors themselves point out, is that “bilingual sessions use English version material” but “in a bilingual class, the instructor mainly uses L1 (Chinese) in lectures and interactions” (p. 6). In addition, authors focus on a specific subject, Fundamental Accounting, and therefore, their work is not able to detect whether a student’s performance may or may not improve over time as English proficiency naturally progresses through the degree. Similarly, Guo et al. (2018) used a quasi-experimental design in which, at the beginning of the semester, the language distribution was 50% English and 50% Chinese, gradually switching to 100% English toward the end of the semester. Hernandez-Nanclares & Jimenez-Munoz (2015), consider instead the data produced by a sample of students enrolled in a Spanish University where English is the medium of L2 instruction. The EMI data is contrasted with a control group of Spanish-taught students who were exposed to identical methodologies and took the same final exam in two different first-year subjects, World Economic History and World Economy. Their results show similar grades in both groups, but these researchers hinted that “those students taught through a foreign language generally learn and perform better” (pp. 11). Similarly, Dafouz, Camacho- Millano, & Urquía (2014) assess the statistic difference between the performance of a group of EMI and non-EMI students enrolled in a Business Administration degree at a large public Spanish university, whose scores in three different first-year subjects, Accounting, Finance and History, were recorded. Again, EMI and non-EMI groups obtained similar results in the three subjects. This work presents an important differential element with 13 respect to other similar research, which is the control of the quasi-experiment for the university access grade. However, these authors omit other covariates, which will be analyzed later. The research of Dafouz & Camacho-Millano (2016), on the subject of Financial Accounting I, also showed no statistical differences across EMI and non-EMI students. Finally, Del Campo, Cancer, Pascual-Ezama, & Urquía-Grande (2015), carried out an exercise in a Bachelor Degree in Business Administration, trying to isolate the effects of all the possible covariates, although with a very different methodological approach than the one in the present paper. Contrary to previous works, their results show significant differences between the final grades of the two students groups. However, this research presents the important limitation of working exclusively with 30 students, 15 EMI and 15 Non-EMI. More recent studies have also addressed this issue, and their results are shown in Table 1. In this research we try to overcome what, from our point of view, are the main limitations in most previous research: we have used a large sample, we have analyzed more than one subject, and, which is certainly the key element of this paper, we have controlled the quasi-experiment for all relevant covariates. Additionally, in the second part of the paper, we considered the performance of a constant sample of students analyzed in different subjects taught in different years. Our aim was to capture the possible variations in learning strategies and performance as students progress through their studies (Urquía-Grande, Camacho-Miñano, & Dafouz, 2018), as well as the possible differences existing between subjects. Table 1 shows a comparison of this paper with previous research already mentioned. 14 Table 1. Comparison of this paper with previous research Controlled by EMI Results (differences) Sample More than 1 subject More than 1 cohort University access grade Other covariates Dafouz, Camacho- Millano, & Urquía (2014) Yes No differences 316 Yes (3 subjects) No (2010-2011) Yes No Hernandez-Nanclares & Jimenez-Munoz (2015) Yes Slightly better performance in EMI students 654 Yes (2 subjects) Yes, 2 cohorts (2012–2014) No No Del Campo, Cancer, Pascual-Ezama, & Urquía-Grande (2015) Yes Better performance in EMI students 30 Yes (all the compulsory courses) No (2009-2010) Yes Yes Dafouz & Camacho- Millano (2016) Yes No differences 383 No Yes, 4 cohorts (2010–2014) No No Costa & Mariotti (2017) Yes Mixed effects depending on level & area) 572 Yes (4 subjects) No No No Guo et al. (2018) No No differences 47 No No (2017) No Yes Arco-Tirado et al. (2018) Yes Worse performance in EMI students 1288 Yes (all subjects) Yes, 2 cohorts (2011-2012) Yes Yes Li (2018) No Negative effect on learning 53 No No No Yes Lin & He (2019) No No differences 498 No No (2015) No Yes Lin & Lei (2021). Yes No differences 194 No No No Yes Current research Part 1 Yes No differences 864 Yes (10 subjects) Yes, 5 cohorts (2013-2018) Yes Yes Part2 Yes No differences 165 Yes (4 subjects) Yes Yes As Makel & Plucker (2014) point out, there is a shortage of replication in the Education Sciences, and this work aims precisely to fill this gap, extending the work of Dafouz, Camacho-Millano, & Urquía (2014), and Dafouz & Camacho-Millano (2016). Our research paper aims to deepen the above discussion about content learning for EMI in HE for non- English native speaking students. The main contribution of this paper is the design of the quasi-experiment (see Table 1), which controls the main covariates. So, our research question is formulated as follows: does the language of instruction have an impact on students’ academic performance? 15 Methods Data and covariates The study was carried out at medium-size private Spanish university. The university, founded in 1890, is composed of seven different schools, with over 13,000 students and around 1,700 lecturers, and offers 43 different undergraduate degrees and 21 official master’s and doctoral programs. Both lecturers and students are largely of Spanish nationality, and the teaching activity is developed mainly in a monolingual context. Nevertheless, the first EMI program in Business Administration was launched in 2013. In this sense, the students admitted to the EMI group must demonstrate an equivalent to a C1 level in English speaking and comprehension, according to the Common European Framework of Reference for Languages (CEFR), evaluated at an entry test provided by the University. The analysis developed in this paper is divided in two different parts. The first one in based on data from 2013, when the EMI business degree was launched, to 2017. So, considering a sample of 864 students enrolled in the Business Administration degree, the academic performance of EMI and Non-EMI groups at the end of the first course have been compared, considering the 10 subjects that constitute the mentioned course. In the second part, we have carried out a longitudinal study on two groups of students who were analyzed during two academic years in four different subjects, all belonging to the area of Quantitative Methods: Mathematics I, Mathematics II, Financial Mathematics, and Statistics. All students whose first language is other than Spanish and those repeating the academic year were removed from the data set1. 1 It should be noted that the four subjects of the EMI group considered in this paper were taught entirely in English, without codeswitching at any time, while this practice is not permitted by university regulations. 16 In both cases it is key to identify the covariates. As previously mentioned, the main contribution of this paper is to carry out a comparison by carefully controlling them. The first of these variables is the student’s gender, since some papers have reported a superior performance by female students in Business Administration related subjects when compared to their male counterparts (Martínez de Ibarreta et al., 2010; Durán-Santomil et al., 2016), although the current empirical evidence is not conclusive. The second covariate is the specialty studied in high school. In the Spanish university system, high school students can choose different branches of specialization. The science major is for students who wish to pursue a STEM-style degree, while the social sciences and humanities major is specifically designed and recommended for those who will choose a degree in business administration (among other degrees). Although the research on this issue remains scarce, some studies underline that in Business Administration-related degrees, paradoxically students who chose a science specialization tend to have a better academic performance than those who have opted for other specialties. It is puzzling that those students who have chosen the recommended major for the business administration degree (social sciences and humanities major) obtain worse results than those who chose a non-recommended major (science major). Arroyo-Barrigüete, Tirado, Mahillo-Fernández & Ramírez (2020) point out that students from science specialty generally obtain better results during the first year than those who studied the specialty of social sciences and humanities. Considering that previous studies provide some evidence of a potential effect of the specialty, we believe it relevant to control this variable. The third covariate is the resident origin of the student, i.e., whether the student comes from a different place from where he/she is going to study the degree, in this case Madrid. Beltrán Barco & La Serna Studzinski (2008) point 17 to a negative effect, due, among other factors, to adaptation to a new life situation, a result that coincides with the conclusions of Tejedor (2003). In contrast, Simón Pérez, Casado- Díaz, Castejón Costa, Driha & Martínez-Bernabéu (2018) do not observe significant differences in academic performance according to their type of residence. Due to this ambiguity, the corresponding variable has been incorporated. Finally, the fourth covariate is the pre-university performance. In the existing literature, pre-university performance has been identified as one of the most relevant variables to predict university performance (McKenzie & Schweitzer, 2001; Tejedor, 2003; Garbanzo-Vargas, 2007), since it somehow synthesizes both a student's aptitudes and skills, as well as their basic knowledge (Beltrán Barco & La Serna Studzinski, 2008). This condition was verified by comparing the grade earned in the university admission tests (EvAU - Evaluación para el Acceso a la Universidad). In the Spanish university system, the EvAU is the equivalent of the SAT, in the sense that the score obtained is used by public universities to select their future students. It is composed of the score obtained in a test, designed and evaluated by the Spanish Ministry of Education, and the average score in the two years of high school. Due to the need to control these four covariates, multiple regression analysis was used in the first part of the paper. Thus, including all students enrolled in the Business Administration degree during the period 2013 - 2018 (Table 2), we have used the academic performance of EMI and Non-EMI students at the end of the first year, considering the average grade in the 10 subjects of this course, as dependent variable. Table 2. Basic statistics of the sample (cohorts 2013-14 to 2017-18) Gender 2013- 14 2014 -15 2015 -16 2016 -17 2017 -18 2013-2018 Men Women 18 Non-EMI 143 123 135 118 116 635 46.3% 53.7% EMI 42 48 46 44 49 229 39.7% 60.3% Total 185 171 181 162 165 864 In the second part of the paper, the 2017-18 cohort has been specifically analyzed, carrying out a longitudinal study of its behavior during two academic years in four different subjects. This exercise complements the previous one, as it allows for the evaluation of the possible evolution of the students over time. We have proceeded in a different way because the sample is significantly smaller, analyzing, one by one, each covariate to verify that they are equal in both groups. Firstly, it has been verified that the percentage of men and women is similar in both groups (Table 3), that the percentage of students coming from a science specialty in high school is also similar, and that the same happens with the proportion of local students. In all cases two-proportions z-test show relatively high p-values, so that we can assume equality between both groups. Table 3. Basic statistics of the sample (cohort 2017-18), including gender, specialty in High School and local Vs non local students Gender High School Local student N Men Women p-value Science Others p-value Madrid Others p-value Non-EMI 116 44.8% 55.2% 1 37.9% 62.1% 0.641 71.6% 28.4% 0.245 EMI 49 44.9% 55.1% 32.7% 67.3% 81.6% 18.4% In relation to the fourth covariate, a comparison of the university entrance grades (EvAU) of both groups has been carried out. First, the hypothesis of normality of both EMI and non- EMI groups was checked by means of a Shapiro test, and then by verifying the requirement of homogeneity of variances. Specifically, a Brown-Forsythe Levene-type procedure was used (Brown & Forsythe, 1974), by considering the median as the function to compute the 19 center of each group, because this procedure generally provides a more robust test. Having verified both hypotheses, a t-test for equality of means (Table 4) was carried out, which confirmed that both groups are homogeneous in terms of their academic performance prior the university. Table 4. Independent samples T-test for EMI and non-EMI grades in university entrance tests (EvAU) University Entrance Tests Shapiro’s Test Levene’s Test t-test for Equality of Means Mean SD W p-value F p-value t df p-value Non-EMI 7.90 0.88 0.98 0.145 1.51 0.2211 0.38 163.00 0.702 EMI 7.95 0.76 0.98 0.452 Additionally, in the case of this longitudinal study, the specific choice of academic subjects significantly reduces any possible bias regarding the requirement level between the EMI and Non-EMI groups, i.e. the extent to which the two tracks place demands on students. On one hand, their belonging to the same area of knowledge and the fact of being taught by lecturers from the same department ensures homogeneity. On the other hand, while there are inevitable differences among lecturers, in each of the selected subjects a strong coordination is constantly implemented, which forces teachers to follow exactly the same evaluation system, to adopt similar intermediate evaluation tests and, most importantly, to use an identical final exam for all groups, where the language (Spanish versus English) is the only existing difference. This ensures an almost identical requirement level, at least within a given year. In the case of this paper, by using longitudinal data and, thus, by comparing the same final exam for all groups in each subject, we ensure that we are measuring performance under very similar conditions. Nevertheless, it is possible that 20 courses in mathematics and statistics could be less sensitive to medium of instruction than many other subjects. For that reason, in addition to the individual analysis of the four selected subjects, we have also compared EMI and Non-EMI groups considering the first course as a whole, which includes 10 subjects. We do not think that it makes sense to perform individual analyses because we cannot ensure an identical exigency level in all of them, but globally we think that the possible biases are compensated and that therefore the global comparison does make sense. As all the key covariates are controlled, a direct comparison can be performed. To summarize, the quasi-experiment described in this paper, both in part 1 and 2, allows us to control key covariates, so that we can evaluate whether university performance varies according to the chosen language of instruction. Statistical analysis Statistical analysis was carried out in R, a free software environment for statistical computing. In the first part of this paper, the models’ estimation has been carried out by using basic functions included in such programing environment (R Core Team, 2013) and the “RCurl” (Lang and CRAN team, 2019) and “car” (Fox and Weisberg, 2019) packages. We detected the existence of heteroskedasticity, and therefore, robust standard deviations have been used. Additionally, we have checked the absence of multicollinearity problems, verifying that variance inflation factors (VIF) are far below 10. Regarding the longitudinal study, two different tests for equality of means were used. Firstly, after the hypothesis of normality (via the Shapiro test) and homogeneity of means (via a Brown-Forsythe Levene-type procedure) were verified, a t-test was carried out. However, in most cases, even when the homogeneity of means is verified, it is not possible 21 to assume normality, so it is necessary to resort to the non-parametric contrast by Mann- Whitney-Wilcoxon (MWW). The four subjects used in the longitudinal study are offered in the Quantitative Methods department, and all of them have a workload of 6 ECTS (European Credit Transfer System). Mathematics 1 is taught during the first semester of the first year, and it deals with the topics of Introduction to Linear Algebra and Integral Theory. Mathematics 2, is taught during the second semester of the first year, and it focuses on Differential Calculus and Theory of Optimization. The subject of Statistics, is taught during the first semester of the second year, and it includes an introduction to Descriptive Statistics, Probability Theory and Inference. And finally, the subject of Financial Mathematics, is also taught during the first semester of the second year, deals with the Fundamentals of Financial Valuation. In all subjects, the final overall course grade in the first examination period, was considered as a measure of the academic performance of the students, evaluated on a 10-point scale (0 being the lowest and 10 the highest). Results and Discussion Global performance in the first course (2013-2017) As mentioned in the previous section, the variables considered for the regression model are the four covariates, as well as whether or not the student belongs to the EMI group. Specifically, this research has included the variable EvAU (grade earned in the university admission tests), Specialty in High School (1 for science, 0 for social sciences and humanities), Local Student (1 for students from Madrid, 0 for students from other places), Gender (1 for women, 0 for men) and belonging or not to the L1 group (1 for Non-EMI students, 0 for EMI students). The results shown in Table 5 indicate that the first three variables are highly significant. 22 Table 5. Regression analysis (2013 – 2017). Dependent variable: average grade in the 10 subjects of the first course Coef. Std. Error p_value EvAU 0.65 0.03 <.000 *** Specialty in High School 0.33 0.05 <.000 *** Local Student 0.35 0.05 <.000 *** Gender 0.05 0.05 0.30 Non-EMI 0.00 0.05 0.98 R-squared: 0.44 F-statistic: 123 (p-value: <.000) There is a clear positive correlation between EvAU and First course grade, which is also different according to the specialty studied in high school. Students from “science”, caeteris paribus perform better than their counterparts from other specialties, even with the same EvAU. The same goes for the place of origin: caeteris paribus, local students perform better than their counterparts. However, neither gender nor membership to the EMI group, which is the variable of interest in this paper, are significant. In fact, the p-value associated to this last variable is very close to 1, which means that after having controlled for EvAU, specialty in high school, and being a local or non-local student, the effect of studying the degree in English is completely irrelevant. In fact, taking into consideration that a relatively high sample of 864 students has been used, and that the final course grade includes 10 different subjects, this result, with such a high p-value, confers a considerable robustness to the conclusions. This is a result that, although consistent with that of previous studies (Dafouz, Camacho-Millano, & Urquia, 2014; Dafouz & Camacho-Millano, 2016; Guo et al., 2018; Lin and He, 2019; Lin & Lei, 2021), provides a more general view, since the sample is significantly larger and all covariates have been controlled. Regarding the paper of Del Campo, Cancer, Pascual-Ezama, & Urquia-Grande (2015), who do control for key 23 covariates, our results disagree with theirs. The reason might be, as the authors themselves admit, that their results are based on a very small sample (15 EMI Vs 15 Non- EMI students). However, the result is also different from that obtained by Arco-Tirado et al. (2018), which not only use a large sample, but also control for multiple covariates. This unexpected discrepancy should be studied in future research. Longitudinal study (cohort 2017) Firstly, we classified the sample students according to their level of performance: low performance (i.e. final grade between 0 and 4.99), medium performance (between 5 and 6.99), and high performance (between 7 and 10). It must be mentioned that the standardized grading system in Spain ranges from 1 to 10, with 5-10 representing a pass grade. Our findings, which are displayed in Table 6, point to similar distributions between EMI and Non-EMI students, except in the case of Financial Mathematics, where EMI students obtain, on average, a higher grade. In fact, this is the only subject in which the majority group is not of "medium performance", which reveals a difference with respect to the remaining three subjects. Table 6. EMI and non-EMI student distribution by level of achievement Low Performance Medium Performance High Performance N % N % N % Math. I Non-EMI 26 22.4% 64 55.2% 26 22.4% EMI 15 30.6% 27 55.1% 7 14.3% Math. II Non-EMI 16 13.8% 64 55.2% 36 31.0% EMI 5 10.2% 27 55.1% 17 34.7% Statistics Non-EMI 31 26.7% 58 50.0% 27 23.3% EMI 12 24.5% 26 53.1% 11 22.4% 24 Financial Math. Non-EMI 49 42.2% 33 28.4% 34 29.3% EMI 12 24.5% 16 32.7% 21 42.9% Performing the corresponding test for equality of means (Table 7)2, we can confirm that the means are not statistically different (5% confidence level) in any of the subjects. Neither can we find any observable pattern as students progress in their studies, since the results are similar in the considered subjects, which belong to different semesters and years: Mathematics I (first year, first semester), Mathematics II (first year, second semester), Financial Mathematics, and Statistics (second year, first semester). The three semesters considered are identical in terms of teaching load (30 ECTS), and very similar in terms of subject complexity. So, the lack of evolution in the grades suggests that there does not seem to be a negative effect of the language at the beginning of the studies. In other words, from the beginning of the studies, it does not seem that the performance of the EMI group is affected by the language, otherwise we would observe a gap with the Non-EMI students that would decrease as time goes by. Table 7. Test for Equality of Means for EMI and non-EMI final grades of the four considered subjects Shapiro’s Test Levene’s Test Test for Equality of Means Mean SD W p-value F p-value Statistic df p-value Math. I Non- EMI 5.79 1.57 0.99 0.560 0.19 0.660 -1.64 163 0.103 EMI 5.34 1.67 0.97 0.154 Math. II Non- EMI 6.05 1.52 0.99 0.376 0.78 0.378 2575.00 0.340 EMI 6.23 1.35 0.96 0.066 Statistics Non- EMI 5.70 1.79 0.98 0.097 0.24 0.626 2899.5 0.839 EMI 5.54 2.03 0.94 0.015 2 Regarding the subject of Business Mathematics I, a t-test was carried out. However, in the rest of subjects it is not possible to assume normality, so a non-parametric contrast of Mann-Whitney-Wilcoxon has been used. 25 Financial Math. Non- EMI 5.33 2.29 0.98 0.040 0.02 0.887 2357.5 0.084 EMI 5.87 2.47 0.91 0.002 Finally, comparing the overall performance of students in the first year of the degree, i.e. considering the 10 subjects of the first course, it can also be seen that there are no statistically significant differences (Table 8), although, as above mentioned, this grade is considerably higher than those obtained in the four subjects of Quantitative Methods that were analyzed in this paper. Table 8. Test for Equality of Means for EMI and non-EMI grades in the first course (across all subjects). First Course Grade Shapiro’s Test Levene’s Test t-test for Equality of Means Mean SD W p-value F p-value t df p-value Non-EMI 6.87 0.81 0.99 0.328 0.69 0.407 1.64 163 0.104 EMI 7.08 0.76 0.99 0.979 These results, again, coincide with those produced by Dafouz, Camacho-Millano, & Urquía (2014), Dafouz & Camacho-Millano (2016), Guo et al. (2018), Lin and He (2019) and Lin & Lei (2021). This confirms that, at least in quantitative subjects of a degree in Business Administration, the language of instruction has no influence on academic results. Conclusions Throughout this paper we analyzed the differences in academic performance of university students enrolled in the degree of Business Administration, depending on the language of instruction. In the first part, using regression analysis to control key covariates, we have compared the performance of 229 EMI Vs 635 Non-EMI students, corresponding to cohorts 2013-14 to 2017-18, considering the average grade in the 10 subjects of the first course. In the second part, we focus on the 2017-18 cohort (49 EMI Vs 116 Non-EMI students), 26 carrying out a longitudinal study of its behavior during two academic years in four different subjects. This exercise complements the previous one, as it allows for the evaluation of the possible evolution of the students over time. In both cases results show that there are no statistically significant differences among groups neither in any of the subjects nor in the global average grade. Additionally, since in the second part we adopted a longitudinal perspective on student academic performance over time, we were able to confirm that as students progress in their studies, no differences appear either. Bearing in mind that the three semesters considered are identical in terms of teaching load (30 ECTS), and very similar in terms of subject complexity, the lack of evolution in the grades suggests that there does not seem to be a negative effect of the language at the beginning of the studies. Additionally, the absence of differences with respect to the Non-EMI group remains in all cases. In other words, from the beginning of the studies, it does not seem that the performance of the EMI group is affected by the language, otherwise we would observe a gap with the Non-EMI students that would decrease as time goes by. In view of the results, and by taking into account the way in which this quasi-experiment was carried out, our conclusion is that the language of instruction does not play a relevant role in academic performance, at least in the degree of Business Administration, assuming that students have an adequate level of knowledge in the language for which they are going to receive instruction. In future research, we would consider it interesting to analyze to what extent this is due to the greater efforts that EMI students are expected to make in acquiring new knowledge in a non-native language, compared to a non-EMI group. It is relevant to analyze the additional effort required to achieve similar results than the Non- 27 EMI students, which probably is related with the more mature learning strategies of these undergraduates (Rivero-Menéndez, Urquía-Grande, López-Sánchez & Camacho-Miñano, 2018). Limitations The main limitation of this work is that, regarding the longitudinal study, the four considered subjects correspond to the area of Quantitative Methods and would be what Yang & Farley (2019: 14) call “less conceptually oriented (LC) subjects”. These authors found that non- native students had a disadvantage in more conceptually oriented subjects versus less conceptual ones, due to linguistic and psychological effects. For this reason, subjects of the first type should be included in the comparison in future research. However, in the first course of the degree there is a considerable amount of more conceptually oriented subjects, and as already indicated, there are also no differences in the overall performance of students when considering all of them. Another limitation of this paper is a potential pre- selection bias in the different cohort of students, due to a non-random sample of the population. Further replication across other universities will permit more confident generalizations beyond specific institutions. References Airey, J. (2004). Can you teach it in English? Aspects of the language choice debate in Swedish higher education. In Wilkinson, R (Ed.). Integrating content and language: Meeting the challenge of a multilingual higher education (97-108). Maastricht, Netherlands: Maastricht University Press. Aizawa, I. & McKinley, J. (2020). EMI challenges in Japan’s internationalization of higher education. In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 27-48). Palgrave MacMillan. AlBakri, S. (2017). Effects of English medium instruction on students’ learning experiences and quality of education in content courses in a public college in Oman. PhD thesis. University of Exeter, UK. 28 Altbach, P. G., & Knight, J. (2007). The internationalization of higher education: Motivations and realities. Journal of Studies in International Education, 11(3-4), 290–305. https://doi.org/10.1177/1028315307303542 Arco-Tirado, J. L., Fernández-Martín, F., Ramos-García, A. M., Littvay, L., Villoria, J., & Naranjo, J. A. (2018). A counterfactual impact evaluation of a bilingual program on students’ grade point average at a spanish university. Evaluation and Program Planning, 68, 81-89. https://doi.org/10.1016/j.evalprogplan.2018.02.013 Arroyo-Barrigüete, J. L., Tirado, G., Mahíllo-Fernández, I., & Ramírez, P. J. (2020). Predictors of performance in Business Administration degrees: the effect of the high- school specialty. Revista de Educación, 390, 129-154. https://doi.org/10.4438/1988- 592X-RE-2020-390-468 Aslan, M. (2018) The debate on English-medium instruction and globalisation in the Turkish context: a sociopolitical perspective, Journal of Multilingual and Multicultural Development, 39:7, 602-616, DOI: 10.1080/01434632.2017.1417413. Beelen, J. (2011). Internationalization at home in a global perspective: A critical survey of the 3rd Global Survey Report of IAU. Revista de Universidad y Sociedad del Conocimiento (RUSC), 8(2), 249-264. Belhiah, H., & Elhami, M. (2015). English as a medium of instruction in the Gulf: When students and teachers speak. Language Policy, 14(1), 3-23. https://doi.org/10.1007/s10993-014-9336-9 Beltrán Barco, A., & La Serna Studzinski, K. (2008). ¿Qué explica el rendimiento académico en el primer año de estudios universitarios? Un estudio de caso en la Universidad del Pacífico. Documento de discusión, DD/08/09. Centro de Investigación de la Universidad del Pacífico. Retrieved from: http://repositorio.up.edu.pe/handle/11354/342 [Accessed 28 May 2021]. Bowles, H. & Murphy, A.C. (Eds), (2020). English Medium Instruction and the Internationalisation of Universities, London: Palgrave Macmillan. Broggini, J. A., & Costa, F. (2017). A survey of English-medium instruction in Italian higher education. An updated perspective from 2012 to 2015. Journal of Immersion and Content- Based Language Education, 5(2), 238–264. https://doi.org/10.1075/jicb.5.2.04bro Brown, M. B., & Forsythe, A. B. (1974). Robust Tests for the Equality of Variances. Journal of the American Statistical Association, 69(346), 364-367. Byun, K., Chu, H., Kim, M., Park, I., Kim, S., & Jung, J. (2011). English-medium teaching in Korean higher education: Policy debates and reality. Higher Education, 62(4), 431-449. https://doi.org/10.1007/s10734-010-9397-4 Chang, Y. Y. (2010). English-medium instruction for subject courses in tertiary education: Reactions from Taiwanese undergraduate students. Taiwan International ESP Journal, 2(1), 55-84. https://doi.org/10.6706/TIESPJ.2010.2.1.3 Coleman, J. A. (2006). English-medium teaching in European Higher Education. Language Teaching, 39(1), 1–14. https://doi.org/10.1017/S026144480600320X Costa, F., & Coleman, J. A. (2013). A survey of English-medium instruction in Italian https://doi.org/10.1177/1028315307303542 https://doi.org/10.1016/j.evalprogplan.2018.02.013 https://doi.org/10.4438/1988-592X-RE-2020-390-468 https://doi.org/10.4438/1988-592X-RE-2020-390-468 http://repositorio.up.edu.pe/handle/11354/342 https://doi.org/10.1075/jicb.5.2.04bro https://doi.org/10.1007/s10734-010-9397-4 https://doi.org/10.6706/TIESPJ.2010.2.1.3 https://doi.org/10.1017/S026144480600320X 29 higher education. International Journal of Bilingual Education and Bilingualism, 16(1), 3- 19. https://doi.org/10.1080/13670050.2012.676621 Costa, F. & Mariotti, C. (2017). “Students’ Outcomes in English-Medium Instruction: Is there any Difference Related to Discipline?”, in Analisi linguistica e letteraria, Critical Issues in English −Medium Instruction at University, vol. 2, 361-371. Costa, F. & Mariotti, C. (2020). EMI Students’ “International Coexistence” at One Italian University. In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 229-258). Palgrave MacMillan. CRUE (2018). La Universidad Española en cifras 2016-2017. CRUE, Universidades Españolas. Retrieved from https://cutt.ly/ajeNYuk [Accessed 30 Dec 2020]. Curle, S., Yuksel, D., Soruç, A. & Altay, M. (2020). Predictors of English Medium Instruction academic success: English proficiency versus first language medium, System, Vol.95, December 2020, 102378, https://doi.org/10.1016/j.system.2020.102378. Dafouz, E. (2018). English-medium instruction and teacher education programs in higher education: ideological forces and imagined identities at work. International Journal of Bilingual Education and Bilingualism, 21(5), 540-552. https://doi.org/10.1080/13670050.2018.1487926 Dafouz, E., & Camacho-Miñano, M. M. (2016). Exploring the impact of English-medium instruction on university student academic achievement: The case of accounting. English for Specific Purposes, 44, 57-67. https://doi.org/10.1016/j.esp.2016.06.001 Dafouz E., Camacho M., & Urquia E. (2014). ‘Surely they can't do as well’: a comparison of business students’ academic performance in English-medium and Spanish-as-first- language-medium programs. Language and Education, 28(3), 223-236. https://doi.org/10.1080/09500782.2013.808661 De Wit, H., Hunter, F., Howard, L. & Egron-Polak, E. (Eds), (2015). Internationalisation of Higher Education Study. European Parliament, Directorate-General for Internal Policies. Policy Department B: Structural and Cohesion Policies. Retrieved from: https://www.europarl.europa.eu/RegData/etudes/STUD/2015/540370/IPOL_STU(2015) 540370_EN.pdf, accessed on 25/01/2022. Del Campo, C., Cancer, A., Pascual-Ezama, D., & Urquía-Grande, E. (2015). EMI vs. Non-EMI: Preliminary Analysis of the Academic Output within the INTE-R-LICA Project. Procedia-Social and Behavioral Sciences, 212, 74-79. Di Gresia, L. M., Porto, A., & Ripani, L. (2002). Rendimiento de los estudiantes de las universidades públicas argentinas. Documentos de Trabajo Nº 45. Retrieved from http://sedici.unlp.edu.ar/handle/10915/3476 [Accessed 4 June 2021]. Doiz, A. Lasagabaster, D. & Sierra, J. M. (2011). Internationalisation, multilingualism and English-medium instruction. World Englishes 30(3), 345-359. https://doi.org/10.1111/j.1467-971X.2011.01718.x Doiz, A. & Lasagabaster, D. (Eds.), (2021). Global Perspectives on Language Aspects and Teacher Development in English-Medium Instruction, London: Routledge. Durán-Santomil, P., Maside-Sanfiz, J. M., Rodeiro-Pazos, D., & Cantorna-Agra, S. https://doi.org/10.1080/13670050.2012.676621 https://cutt.ly/ajeNYuk https://doi.org/10.1080/13670050.2018.1487926 https://doi.org/10.1016/j.esp.2016.06.001 https://doi.org/10.1080/09500782.2013.808661 https://www.europarl.europa.eu/RegData/etudes/STUD/2015/540370/IPOL_STU(2015)540370_EN.pdf https://www.europarl.europa.eu/RegData/etudes/STUD/2015/540370/IPOL_STU(2015)540370_EN.pdf http://sedici.unlp.edu.ar/handle/10915/3476 https://doi.org/10.1111/j.1467-971X.2011.01718.x 30 (2016). Determinantes del rendimiento académico del alumnado de una asignatura de Contabilidad: el caso de la USC. REDU: Revista de Docencia Universitaria, 14(1), 151- 178. European Commission (2013). European higher education in the world. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions, COM (2013) 499 final, Brussels. European Commission (2017). Key Data on Teaching Languages at School in Europe – 2017 Edition. Eurydice Report. Luxembourg: Publications Office of the European Union. Fox J., & Weisberg, S. (2019). An {R} Companion to Applied Regression, Third Edition. Thousand Oaks CA: Sage. Gabriëls, R. & Wilkinson, R. (2020). Resistance to EMI in the Netherlands. In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 49-76). Palgrave MacMillan. Garbanzo Vargas, G. M. (2007). Factores asociados al rendimiento académico en estudiantes universitarios, una reflexión desde la calidad de la educación superior pública. Revista educación, 31(1), 43-63. https://doi.org/10.15517/REVEDU.V31I1.1252 Green, W. & Whitsed, C. (Eds.), 2014, Critical Perspectives on Internationalising the Curriculum in Disciplines: Reflective Narrative Accounts from Business, Education and Health (pp. 277-296). Rotterdam, Sense publisher. Gregersen-Hermans, J. & Lauridsen, K.M. (Eds) (2021). Internationalising Programmes in Higher Education: An Educational Development Perspective. New York, Routledge. Guo, H., Tong, F., Wang, Z., Min, Y., & Tang, S. (2018). English-vs. Chinese-medium instruction in Chinese higher education: A quasi-experimental comparison. Sustainability, 10(11), 4230. https://doi.org/10.3390/su10114230 Haines K., Kroese M., & Guo D. (2020). Language Usage and Learning Communities in the Informal Curriculum: The Student as Protagonist in EMI? In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 181- 204). Palgrave MacMillan. Hernandez-Nanclares, N. & Jimenez-Munoz, A. (2015): English as a medium of instruction: evidence for language and content targets in bilingual education in economics. International Journal of Bilingual Education and Bilingualism. 20(7), 883-896. https://doi.org/10.1080/13670050.2015.1125847 Hudzik, J. (2011). Comprehensive internationalization: From concept to action. Washington DC, NAFSA Association of International Educators. Hultgren, A. K. (2014). English language use at the internationalized universities of Northern Europe: Is there a correlation between Englishisation and world rank? Multilingua, 33(3-4), 389–411. https://doi.org/10.1515/multi-2014-0018 Hultgren, A. K., Jensen, C., & Dimova, S. (2015). English-Medium Instruction in European Higher Education: From the North to the South. In S. Dimova, A. K. Hultgren, & C. Jensen (Eds.). English-medium instruction in European higher education. Language and social https://doi.org/10.15517/REVEDU.V31I1.1252 https://doi.org/10.3390/su10114230 https://doi.org/10.1080/13670050.2015.1125847 https://doi.org/10.1515/multi-2014-0018 31 life, 4 (1-18). Berlin: De Gruyter Mouton. Jensen, C., & Thøgersen, J. (2011). Danish University lecturers' attitudes towards English as the medium of instruction. Ibérica, Revista de la Asociación Europea de Lenguas para Fines Específicos, 22, 13-33. Jiang, L., Jun Zhang, L. & May, S. (2016). Implementing English medium instruction (EMI) in China: teachers’ practices and perceptions, and students’ learning motivation and needs, International Journal of Bilingual Education and Bilingualism, 22(2). https://doi.org/10.1080/13670050.2016.1231166 Kaur J. (2020) Using English for Interaction in the EMI Classroom: Experiences and Challenges at a Malaysian Public University. In Bowles H., Murphy A.C. (Eds.). English- Medium Instruction and the Internationalization of Universities (pp. 129-154). Palgrave MacMillan. Kim, K. R. (2011). Korean professor and student perceptions of the efficacy of English- medium instruction. Linguistic Research, 28(3), 711-741. Kim, E. G., & Shin, A. (2014). Seeking an effective program to improve communication skills of non-English-speaking graduate Engineering students: The case of a Korean Engineering school. IEEE Transactions on Professional Communication, 57(1), 41–55. https://doi.org/10.1109/TPC.2014.2310784 Kim, E. G., Kweon, S. O., & Kim, J. (2017). Korean engineering students’ perceptions of English-medium instruction (EMI) and L1 use in EMI classes. Journal of Multilingual and Multicultural Development, 38(2), 130-145. https://doi.org/10.1080/01434632.2016.1177061 Kuteeva, M., Kaufhold, K. & Hynninen, N. (Eds.) (2020). Language Perceptions and Practices in Multilingual Universities. Cham: Palgrave Macmillan. Lang, D. T. &CRAN team (2019). Package ‘RCurl’. URL http://www.omegahat.net/RCurl Lauridsen K.M. (2020). It Does Not Happen by Osmosis: Creating an Internationalized Learning Opportunity for All Students Requires Careful Consideration and Specific Action. In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 205-228). Palgrave MacMillan. Leask, B. (2009). Using formal and informal curricula to improve interactions between home and international students. Journal of Studies in International Education, 13, 205- 221. Leask, B. (2013). Internationalising the curriculum in the disciplines – imagining new possibilities. Journal of Studies in International Education, 17 (2), 103-118. https://doi.org/10.1177/1028315312475090. Le Phan, H. (2013). Issues surrounding English, the internationalization of higher education and national cultural identity in Asia: A focus on Japan. Critical Studies in Education, 54(2), 160–175. Li, M. (2018). The effectiveness of a bilingual education program at a Chinese university: a case study of social science majors. International Journal of Bilingual Education and Bilingualism, 21(8), 897-912. https://doi.org/10.1080/13670050.2016.1231164 https://doi.org/10.1080/13670050.2016.1231166 https://doi.org/10.1109/TPC.2014.2310784 https://doi.org/10.1080/01434632.2016.1177061 https://doi.org/10.1177/1028315312475090 https://doi.org/10.1080/13670050.2016.1231164 32 Lin, T., & Lei, J. (2021). English-Medium Instruction and Content Learning in Higher Education: Effects of Medium of Instruction, English Proficiency, and Academic Ability. SAGE Open, 11(4), 21582440211061533, https://doi.org/10.1177/21582440211061533 Lin T., & He Y. (2019). Does Bilingual Instruction Impact Students’ Academic Performance in Content-Based Learning? Evidence from Business School Students Attending Bilingual and L1 Courses. Sustainability, 11(1), 262. https://doi.org/10.3390/su11010263 Lee, K. & Lee, H. (2018) Korean graduate students’ self-perceptions of English skills and needs in an English-medium instruction context, Journal of Multilingual and Multicultural Development, 39:8, 715-728, https://doi.org/10.1080/01434632.2018.1438442 Macaro, E., Curle, S., Pun, J., An, J., & Dearden, J. (2018). A Systematic Review of English Medium Instruction in Higher Education. Language Teaching, 51(1), 36-76. https://doi.org/10.1017/S0261444817000350 Makel, M. C., & Plucker, J. A. (2014). Facts are more important than novelty: Replication in the education sciences. Educational Researcher, 43(6), 304-316. https://doi.org/10.3102/0013189X14545513 Martínez de Ibarreta, C., Rúa-Vieites, A., Redondo-Palomo, R., Fabra-Florit, M. E., Núñez-Partido, A., & Martín-Rodrigo, M. J. (2010). Influencia del nivel educativo de los padres en el rendimiento académico de los estudiantes de la ADE. Un enfoque de género. AEDE, Asociación de Economía de la Educación. 5(1), 1273-1294. Mauranen, A., Hynninen, N., & Ranta, E. (2010). English as an academic lingua franca: The ELFA project. English for Specific Purposes, 29(3), 183-190. https://doi.org/10.1016/j.esp.2009.10.001 McKenzie, K. & Schweitzer, R. (2001). Who succeeds at university? Factors predicting academic performance in first year Australian university students. Higher education research & development, 20(1), 21-33. https://doi.org/10.1080/07924360120043621 MECD, Ministerio de Educación, Cultura y Deporte (2016). Estrategia para la internacionalización de universidades españolas 2015-2020. Edición 2016. Secretaría General de Universidades. Subdirección General de Documentación y Publicaciones. Murphy A.C. & Mengistu M.S. (2020). The Challenges of Internationalization in EMI Tertiary Education in Ethiopia. In Bowles H., Murphy A.C. (Eds.). English-Medium Instruction and the Internationalization of Universities (pp. 77-102). Palgrave MacMillan. Nilsson, B. (2003). Internationalization at home from a Swedish perspective: The case of Malmö. Journal of Studies in International Education, 7(1), 27-40. https://doi.org/10.1177/1028315302250178 Pritasari, A., Reinaldo, H. & Watson, C. W. (2019). English-medium instruction in Asian business schools: a case study, Journal of Multilingual and Multicultural Development, 40:1, 1-13, DOI: 10.1080/01434632.2018.1458855 R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL: http://www.R-project.org/ Rivero-Menéndez, M. J., Urquía-Grande, E., López-Sánchez, P., & Camacho-Miñano, M. https://doi.org/10.1177/21582440211061533 https://doi.org/10.3390/su11010263 https://doi.org/10.1080/01434632.2018.1438442 https://doi.org/10.1017/S0261444817000350 https://doi.org/10.3102/0013189X14545513 https://doi.org/10.1016/j.esp.2009.10.001 https://doi.org/10.1080/07924360120043621 https://doi.org/10.1177/1028315302250178 33 M. (2018). Motivation and learning strategies in accounting: Are there differences in English as a medium of instruction (EMI) versus non-EMI students?. Revista Española de Contabilidad, 21(2), 128-139. https://doi.org/10.1016/j.rcsar.2017.04.002 Rose, H., Curle, S., Aizawa, I., & Thompson, G. (2019). What drives success in English medium taught courses? The interplay between language proficiency, academic skills, and motivation. Studies in Higher Education, 45(11), 2149-2161. https://doi.org/10.1080/03075079.2019.1590690 Rowland, L. & Murray, N. (2020) Finding their feet: lecturers’ and students’ perceptions of English as a medium of instruction in a recently-implemented Master's programme at an Italian university, Journal of Multilingual and Multicultural Development, 41:3, 232-245, DOI: 10.1080/01434632.2019.1614186. Simón Pérez, H. J., Casado Díaz, J. M., Castejón Costa, J. L., Driha, O. M. & Martínez- Bernabéu, L. (2018). Efecto del tipo de alojamiento y el tiempo de desplazamiento sobre el rendimiento académico del alumnado universitario. In Memorias del Programa de Redes-I3CE de calidad, innovación e investigación en docencia universitaria: Convocatoria 2017-18 (1849-1855). Instituto de Ciencias de la Educación. Tatzl, D. (2011). English-medium masters’ programmes at an Austrian university of Applied Sciences: Attitudes, experiences and challenges. Journal of English for Academic Purposes, 10, 252-270. https://doi.org/10.1016/j.jeap.2011.08.003 Tejedor, F. J. (2003). Poder explicativo de algunos determinantes del rendimiento en los estudios universitarios. Revista española de pedagogía, 224, 5-32. Tsuneyoshi, R. (2018). The internationalization of Japanese education: ‘International’ without the ‘Multicultural’. Educational Studies in Japan: International Yearbook, 12, 49– 59. Urquía-Grande, E., Camacho-Miñano, M. M., & Dafouz, E. (2018). ¿Hacia una nueva dinámica de clase? Análisis comparativo de las preferencias de aprendizaje del alumnado universitario en titulaciones bilingües. Papeles de Europa, 31(1), 41-55. https://doi.org/10.5209/PADE.61488 Wächter, B. (2003). An introduction: Internationalisation at home in context. Journal of Studies in International Education, 7(5), 5–11. Wang Y. (2020). The Role of English in Higher Education Internationalization: Language Ideologies on EMI Programmes in China. In Bowles H., Murphy A.C. (Eds.). English- Medium Instruction and the Internationalization of Universities (pp. 103-128). Palgrave MacMillan. Williams, D.G. (2015). A systematic review of English Medium Instruction (EMI) and implications for the South Korean higher education context. English Language Teaching World Online. Special Issue on CLIL, 1–23. Wu, W. S. (2006). Students’ attitudes toward EMI: Using Chung Hua University as an example. Journal of Education and Foreign Language and Literature, 4, 67-84. Xie, W. & Curle, S. (2020). Success in English medium instruction in China: Significant indicators and implications. International Journal of Bilingual Education and Bilingualism, https://doi.org/10.1016/j.rcsar.2017.04.002 https://doi.org/10.1080/03075079.2019.1590690 https://doi.org/10.1016/j.jeap.2011.08.003 https://doi.org/10.5209/PADE.61488 34 1-13. https://doi.org/10.1080/13670050.2019.1703898 Yang, Y., Volet, S., & Mansfield, C. (2017). Motivations and influences in Chinese international doctoral students’ decision for STEM study abroad, Educational Studies 44(3), 264-278. https://doi.org/10.1080/03055698.2017.1347498 Yang, H. H., & Farley, A. (2019). Quantifying the impact of language on the performance of international accounting students: A cognitive load theory perspective. English for Specific Purposes, 55, 12-24. https://doi.org/10.1016/j.esp.2019.03.003 Zhang, Z. (2018) English-medium instruction policies in China: internationalisation of higher education, Journal of Multilingual and Multicultural Development, 39:6, 542-555, DOI: 10.1080/01434632.2017.1404070. https://doi.org/10.1080/13670050.2019.1703898 https://doi.org/10.1080/03055698.2017.1347498 https://doi.org/10.1016/j.esp.2019.03.003 Introduction Academic performance in EMI Methods Data and covariates Statistical analysis Results and Discussion Global performance in the first course (2013-2017) Longitudinal study (cohort 2017) Conclusions Limitations References