Developing Intercultural Communicative Competence: A Cornerstone in EMI in-service Training Programmes in Higher Education

Globalisation and technological advances have contributed greatly to increasing mobility in higher education (HE). One of the main stakeholders in this process are lecturers, for whom internationalisation is often understood as simply the use of a lingua franca— usually English—in their teaching (then known as English medium instruction, EMI) to the detriment of other key factors such as intercultural competence. The present study focuses on the analysis of a specific in-service training module intended to develop the intercultural communicative competence (ICC) of a group of 21 Spanish lecturers working at two public universities. Adopting an empirical approach, we specifically aim to contrast the participants’ beliefs about their own ICC before and after taking the training course. Adopting a mixed-method approach, two data gathering processes were used prior to and after the training: questionnaire and self-reflection report. The results show that before the training participants were willing to learn but mostly unaware of the importance of ICC. Once trained, many of them seem to have become aware that EMI goes beyond language issues and that there is a crucial need to work on the intercultural component to fully develop HE internationalisation. Furthermore, this study highlights the need for major changes in in-service training courses as well as in language policies for EMI in order to give ICC the presence it should have.


Introduction
Lecturers are one of the main stakeholders in the steady process of internationalisation that universities are undergoing. However, for many of these lecturers, internationalisation (despite its complexity) is merely understood as the use of English as a lingua franca in their teaching. As a result, many institutions merely demand stakeholders (teachers, students, staff) to certify a specific level of English (this varies across institutions) to conduct their international lessons in English (aka English medium instruction, EMI) and pay no heed to other key factors such as intercultural communicative competence (ICC) (see Dafouz and Smit 2020).
In fact, as research has proven (Laffier et al. 2017;Murray 2016;Zhou and Austin 2017, among others), linguistic competence in the language of instruction is insufficient in internationalised classrooms, and both lecturers and students need not only to be intercultural aware but also intercultural competent (Deardorff and Arasaratnam-Smith 2017). Keeping the above in mind, the present study focuses on the analysis of specific training intended to develop the ICC of lecturers from different disciplines engaged in this new teaching scenario. It adopts a mixedmethod approach combining both a quantitative and qualitative analysis.
By adopting an empirical approach, we aim to contrast the participants' beliefs about their intercultural communicative competence before and after doing the training. More specifically, our aim is to look into the participants' initial perceptions of their competence to see what their starting point is. Second, we looked at what the participants' beliefs regarding intercultural communicative competence were after the course was completed. Once the course was over we expected to have changed their initial beliefs regarding this issue and, thus, to find some interesting contributions in the self-reflective task related to the need to develop this competence for both students and lecturers.
The rest of the paper is divided into four further sections. The second section briefly reviews the key theoretical concepts dealt with in the paper-i.e., Intercultural Competence and Intercultural Communicative Competencefocusing on definitions of the two concepts and the main issues raised by the canonical definitions. The third section describes the methodology, with special attention to the description of the participants' and the data-gathering process and tools. Section four presents and discusses the results, both quantitative and qualitatively, before our conclusions are drawn in section five.

Theoretical Background
In one of the canonical publications on Intercultural Competence (IC), Byram (1997, 70) defines IC as a person's "ability to interact in their own language with people from another country and culture, drawing upon their knowledge about intercultural communication, their attitudes of interest in otherness and their skills in interpreting, relating and discovering, i.e. of overcoming cultural difference and enjoying intercultural contact." In their definition, Chen and Starosta (1996) point out that interculturally competent people are those who can interact "effectively and appropriately" with people with different cultural identities. The same authors consider that intercultural competence consists of an affective perspective (attitudes), a cognitive perspective (knowledge) and a behavioral perspective (skills). Indeed, knowledge of one's culture and that of others is not enough to constitute IC. Thus, the very notion of IC primarily rests on the psychological and attitudinal aspects of the individual themselves. As Deardorff (2006) has pointed out, all the seminal definitions of the concept of IC involve the development of the "individual's skills and attitudes in successfully interacting with persons of diverse backgrounds" (9). However, as argued by Dervin and Gross (2016b): "an approach to intercultural competence that fails to point coherently, cohesively and consistently to the complexity of self and the other fails to accomplish what it should do: Helping people to see beyond appearances and simplifying discourses -and thus lead to 'realistic' encounters" (3).
Knowledge of one's own society could thus be considered the departure point for developing IC, as the most influential theories have put forward (Byram 1997(Byram , 2020Chen 1989Chen , 1992Deardorff 2011). Taking this into account, our study focuses first of all on the participants' perception and knowledge of the self, not only regarding strengths and weaknesses but also as a member of one's own society. This intrapersonal dimension, which constitutes our first category of analysis, is in line with King and Baxter Magolda's (2005) Intercultural Maturity Model, which is based on the constructive developmental tradition of human development (Kegan 1980;Piaget 1950). In fact, acquiring this knowledge is a lifelong learning process that needs continuous learning on the side of the individual. In their refinement of the model, Pérez et al. (2015, 760) explain that the term constructive refers to "individuals' abilities to interpret and make meaning of their experiences [...]; developmental refers to the growth of increased capacities to do so in more adaptive and complex ways over time". Thus, it is possible for an individual to find themselves at an initial, intermediate, or high level of intercultural maturity along three dimensions 1 : cognitive, intrapersonal, and interpersonal. According to King and Baxter Magolda (2005: 574), the cognitive dimension includes "complex understanding of cultural differences", the intrapersonal dimension can be defined as the "capacity to accept and not feel threatened by cultural differences", and, finally, the interpersonal dimension refers to the individual's "capacity to function interdependently with diverse others". From our perspective, this model is interesting because it includes knowledge of the self (intrapersonal dimension) besides knowledge of one's culture, which are often subsumed within the same dimension in other models. Consequently we took intrapersonal knowledge into account when deciding on the dimensions to be studied for this work, considering continuous learning as our first category.
Together with the individual's knowledge of the self, another major aspect stressed by the seminal definitions of the concept of IC is the notion of "successful interaction" (Byram 1997;Chen, 1989Chen, , 1992. More recently, IC has been defined as "the skill in facilitating successful intercultural communication outcomes in terms of satisfaction and other positive assessments of the interaction and the interaction partner" (Jandt 2013, 35, emphasis added). This is the main reason why we consider interpersonal engagement, closely related to the interpersonal dimension of the Intercultural Maturity Model (King and Baxter Magolda 2005), as our second category.
IC definitions often stress differences rather than commonalities (see Maffesoli and Strohl 2015;Egekvist 2016). As argued by Phillips (2007), a stress on similarities might help develop a deeper sense of global citizenship and foster the social construction of culture, diminishing ethnocentric perspectives and intercultural anxiety (Gudykunst 2005a) or culture shock (Roskell 2013). In fact, as Liu et al. (2011, 273) point out: "To overcome the barrier of differences, we need to build a mutual understanding with the other interactant. One way to achieve this is to focus on similarities rather than differences. As we gain more knowledge about each other, we may find that, despite the visible differences we share similarities in a number of ways." Furthermore, the role psychological and emotional factors play in intercultural encounters has also been frequently addressed. These factors may interfere with interaction in two ways: first, affecting our attitude when we approach the interlocutor and, secondly, affecting our motivation to engage in the communicative act itself. As Liu et al. (2011, 272)  situation, such as fear, like or dislike, anger, or stress. Emotions affect our motivation to interact with others from different cultures." As a result, we have included a third dimension, on psychological attitude, in our research tool.
The terms IC and ICC often seem to be used interchangeably. Some authors (Chen 1989;Byram 2020) consider there is one main difference between both terms. For example, Guilherme (2000, 297) points out that the main difference is that ICC involves "the ability to interact effectively with people from cultures that we recognise as being different from our own". In the same vein, Wilkinson (2020: np), argues that ICC "acknowledges the fundamental role of language difference, and therefore the need for linguistic as well as cultural competence". In other words, intercultural competence needs to include language-more specifically, a lingua franca-as a key aspect. As such, the perspective adopted in this paper falls within the field of ICC.
Taking all the above into consideration, we have focused on the following four macro dimensions: continuous learning (CL), interpersonal engagement (IE), psychological attitudes (PSY) and language skills, both verbal (VB) and nonverbal (NVB).
(i) Continuous learning can be defined as the lifelong process through which an individual develops their intercultural communicative competence (in line with Deardorff's (2006) Process Model; see also Sercu (2010)). As pointed by Bok (2009): "Intercultural competence is a lifelong process; there is no pinnacle at which someone becomes "interculturally competent". Given the lifelong learning inherent in intercultural competence development, therefore, it is imperative that learners regularly engage in reflective practice in regard to their own development in this area." (xiii, emphasis added) More specifically, continuous learning encompasses the individual's awareness of the self ("know thyself"), and their curiosity and openness to continuing to learn: 2 (1) my room for improvement is still very large but my desire to improve as well.
(ii) Interpersonal engagement can be defined as the willingness and ability not only to interact but also engage with others, especially with individuals from different cultures. This is in line with King and Baxter Magolda's (2005, 576) Intercultural Maturity Model, where the interpersonal element is also a key component: (2) I think that preparing and studying hard could bring me new deals in this area of EMI as a manner of interchange culture with lectures and international students.
(iii) The psychological attitude dimension can be defined as the individual's ability to identify, understand, and manage their emotional reactions -anxiety, fear or stress, for example-when receiving intercultural stimuli in communication.
As already mentioned, positive and negative emotions affect the individual's motivation to approach and interact with others (Liu et al. 2011, 271). (3) and (4) illustrate this dimension, (3) shows a positive reaction and (4) a negative one: (3) I find it [teaching in English] very appealing and I am truly motivated.
(4) Getting to be intercultural competent is a hard job.
(iii) Language skills refers to the capacity an individual has to use a lingua franca effectively and appropriately in a particular intercultural communicative situation, and to be able to negotiate a shared understanding and avoid communicative dissonance 3 : (5) English is our "lingua franca" in this process, but it is really important to improve (at least in my case) my level of English.
Our concept of language skills comprises two subdimensions: verbal and nonverbal. The former refers to the use of a particular linguistic code, written or spoken, of a language and all its components. The latter to the conscious or unconscious use of kinesics and proxemics to convey meaning in a communicative act. In (6), the only participant that reflects on the importance of kinesics or proxemics points out: (6) In relation to the [sic] behaviour into [sic] an international classroom, paying special attention to issues that often go unnoticed such as kinesics or proxemics.
It is important to point out that these four dimensions are not rigidly separate, and there may be a slight degree of overlap between them (see Section 4). Figure  It is important to point out that these four dimensions are not rigidly separate, and there may be a slight degree of overlap between them (see Section 4). Figure 1 shows the interrelationship of the dimensions. For the sake of clarity, the rest of the section has been divided into three parts. The first part describes the course. The second part deals with the participants involved in the study and the third section describes the data gathering tools and procedure.

Methodology
As mentioned in the Introduction, we aim to compare the participants' beliefs about their ICC before and after the training. More specifically, we intend to answer these research questions: (i) what were the participants' initial beliefs about their own ICC prior to the training? And (ii) to what extent has the training changed the participants' initial perceptions of their ICC? For the sake of clarity, the rest of the section has been divided into three parts. The first part describes the course. The second part deals with the participants involved in the study and the third section describes the data gathering tools and procedure.

Course description
This course is an in-service training EMI course consisting of four modules-i.e. linguistic skills, interactive practices, academic literacies and interculturality. Each module involves four hours of face-to-face training which is preceded by participants completing a set of preparatory tasks. The full course lasts for one week and follows an interactive learning, student-centred methodology that uses pair work and small group discussions during the first four days. In the final session participants carry out individual micro-teachings on any topic they choose, trying to put into practice some of the different strategies they have been provided with during the course.
The module on Interculturality is on the fourth day of the course. Following mostly an inductive approach, participants are requested to reflect on the notion of culture, its components and how they might reflect on working in a multicultural class. The module also introduces the concept of ICC.
Prior to the start of the module, participants were asked to fill in a questionnaire (see Appendix 1) with the objective of them starting to reflect on their own ICC. The post-course activity consisted in a self-reflective report (see Appendix 2) where participants were asked to reflect on their own learning, their impressions of EMI and their own personal impressions as future EMI lecturers.

Participants
This study involved 21 participants (12 female and 9 male) from two Spanish Universities who took part in three different editions of the course (see 3.1), rendering three different groups. Initially, there were 25 participants but 4 could not be included as they failed to hand in the post-course task.
All participants are university lecturers, of different professional grades and from a wide range of disciplinary cultures (Biology, Chemistry, Economics, Law, Medicine, Political Sciences and Psychology). Each had recently gained a B2 qualification in English in the Common European Framework of Reference (Council of Europe, 2018). None had previously taught in EMI, which was the main rationale behind their taking this course. It is important to mention that all participants were recruited on a voluntary basis.

Data Gathering Tools and Procedure
The pre-and post-course tasks, which participants had to complete to be certified for their attendance (see 3.1), were also the principal instruments used to gather the data for the study, combining, therefore, a quantitative and qualitative analysis in a mixed-method approach. The questionnaire consists of 20 Likert-scale (1 to 5) questions, divided into four blocks: continuous learning (CL), interpersonal engagement (IE), psychological attitude (PSY) and language skills, subdivided into verbal (VB) and non-verbal (NVB). Each block of questions was followed by a blank space where participants could add their own comments. Unfortunately, only four participants made additional comments and these were quite brief.
The questionnaire, adopted from that designed by the Kozai Group (2009), is specifically developed to evaluate the skills critical to interacting effectively in intercultural contexts. Both researchers decided to use this survey as they agreed that it encompassed the main aspects of Deardorff's process (2006). This particular questionnaire was chosen as the basis for the current work as it includes questions specifically targeted at the individual's knowledge of the self-emotions and feelings when interacting interculturally, curiosity, openmindedness, for example-as well as covering the other aspects of interest here in terms of interaction-interpersonal engagement and linguistic skills, both verbal and non-verbal. The survey was accompanied by an adapted version of the rubric developed by the Association of American Colleges and Universities (2019) which also focuses on knowledge (of the self and awareness of their own culture), psychological skills, linguistic skills (both verbal and non-verbal) and attitudes. Following the rubric, participants were thus able to 'measure' their own intercultural competence prior to participating in the module and in this way identify their own strengths and weaknesses.
The second data-gathering method was the self-reflective report that the participants had to write after completion of the training (see Appendix 2), which comprised two sections. The first required trainees to write a 350-400 word reflection summarising the most salient points in their learning throughout the training. This process was guided by the use of questions, which the participants did not need to answer in full, and could simply use to sum up their ideas. The second block required participants to write a brief description (350-400 words) of their role as a prospective EMI lecturer in the context of their own institution. They were prompted by four concepts: needs, resources, enablers and barriers. This rendered a corpus of circa 17,905 words. For the analysis of the self-reports, we opted to carry out a thematic analysis (Guest et al. 2011). The process was in three phases: (i) Initialisation phase: this involved two stages: first, both researchers read all the reports individually and several times, in line with the recommendation for the "careful reading of transcripts, and listing meaningful, recurrent ideas and key issues in data" (Vaismoradi et al. 2016, 103). Secondly, each researcher made notes and observations on their respective initial reading of the data.
(ii) Construction phase: each researcher labelled the participants' comments, classifying them according to the four dimensions explained. Thus, when participants described themselves in terms that identified them as curious, willing to learn, or aware of strong and weak points, these comments were coded as belonging to the continuous learning (<CL>) dimension. Likewise, if comments were made about their relationship with others (e.g., other lecturers, students), these comments were coded as belonging to the interpersonal engagement (<IE>) dimension. Participants' comments on their fears, anxiety, challenges, etc. were coded as belonging to the psychological attitudes (<PSY>) dimension. Finally, comments on the use of language, etc. were coded as either verbal <VB> or non-verbal <NVB> language skills. For <PSY>, <VB> and <NVB>, a further distinction was made between positive and negative reflections. For example, when participants described interculturality in the EMI classroom as a challenge, it was tagged as <PSY, NEG> as opposed to participants who regarded it as an opportunity (<PSY, POS>). Likewise, comments by participants who felt confident with their use of English as a lingua franca in intercultural contexts were tagged as <VB, POS> while those comments stressing a participant's need for help from native speakers or their lack of confidence when using English were analyzed as <VB, NEG>.
(iii) Rectification phase: The two researchers compared and contrasted their individual analyses using Cohen's kappa test for agreement, which produced a classification of substantial agreement (0.76) with respect to inter-rater reliability. After this first round, the researchers discussed those cases which they had not initially coded in the same way (31 comments out of a total of 130). Reaching consensus through discussion on these remaining cases led to perfect agreement of inter-rater reliability (0.100) being achieved in the second round.

Results and Discussion
For the sake of clarity, this section has been divided into two subsections relating to each of the two data-gathering methods: section 4.1 deals with the questionnaires and answers the first research question, section 4.2 on the analysis of the self-reflective reports written by the participants with special emphasis placed on the second research question.

Questionnaires
This section looks at the responses to the pre-course questionnaires, with each of the dimensions defined in Section 2 considered separately.

Continuous learning
The first block focuses on the participants' continuous learning; more specifically, their degree of curiosity, their knowledge of the self and their open-mindedness. As this section comprised three items, the maximum rate was 15. Scores in the range of 13 to 15 were considered highly skilled, from 10 to 12 as quite skilled, 7 to 9 averagely skilled, 3 to 6 poorly skilled, and finally 0 to 2 very poorly skilled. Of the 21 participants 10 had scores of 13 or above (2 at 15; 3 at 14; and 5 at 13), another 9 rated between 12 and 10 (5 at 12; 2 at 11; and 2 at 10), while the remaining 2 participants rated 9. By and large, this suggests that, before the training, most participants understand themselves to be highly skilled on CL (47.61%) or as quite skilled (42.85%), and just 9.5% think they were averagely skilled.
The three items in this block were as follows: Item 1 (Do you continually seek to understand and learn about the activities, behavior, and events that occur around you?), Item 2 (Would you say that you know yourself (e.g., your values, strengths, weaknesses, interpersonal style, behaviour?) and Item 3 (Are you interested in other cultures?). Table 1 summarises the results for each item (i.e., the extent to which the participant responded positively or negatively to the item). As can be observed, more than two thirds of the participants consider themselves very or quite curious and willing to learn. This might explain why they are also willing to try a new kind of teaching they have never experienced before. The same applies to their self-awareness (item 2), with 76% of participants in the first two categories, and 14% in the third one. There are two participants (10%) that consider their knowledge of the self to be poor. Finally, all participants declare themselves to be open-minded, as was to be expected due to the voluntary nature of the course.

Interpersonal Engagement
The second block aims to gauge the extent of participants' actual interaction with other cultures (items 5 to 10), as well as their potential interest in doing so 4 In all the tables, the first figure refers to the number of participants, being n= 21. In brackets, we have included the ratio. As n=21, ratios are approximate and do not include decimals. This explains why some of the items do not add up to 100% (e.g. item 7, item 12, item 14, item 17), but add up either to 99% or 101%. Exceptionally item 16 adds up to 102%.

Psychological Attitudes
The third block includes three questions targeted at the way participants perceived their own emotional intelligence, ethnocentrism and their degree of anxiety in intercultural contexts (Gudykunst 2005a). Since the maximum rate is 20, the range of 20 to 17 is considered highly skilled, the range from 16 to 13 quite skilled, 12 to 9 averagely skilled, 8 to 5 is poorly skilled, and finally 4 to 0 very poorly skilled. In this block, 47.61% of participants believe themselves to be highly skilled, i.e., they rated 17 or above (2 people rated 20; 1, 19; 4, 18; and 3, 17) and 52.38% to be quite skilled (4 rated 16; 2, 15; 3, 14; and 2, 13). Again, no participant considered themselves to be averagely skilled or below. In summary, before the training, the participants regard themselves as ready to deal with intercultural exchanges without any anxiety. Table 3 presents the results for each item:  (2) 1 (5%) 0% 0% 1 (5%) Very poorly(1) 0% 0% 0% 0% As can be observed, the most common tendency is for participants to respond positively to these four items. In the same vein, none of the participants choose the last option (very poorly) for any of the items. However, it is interesting to see that almost 91% of the participants report they know themselves as highly skilled (24%) or quite skilled (67%) (Item 12). Most of them also report feeling highly skilled or quite skilled (28% and 57% respectively) in intercultural encounters, with just a minority (5%) showing a slight degree of intercultural anxiety (Item 14).

Language Skills
In the last block, participants were asked about their verbal and non-verbal communication skills. This block includes a specific question to find out the number of foreign languages participants spoke. This is the only item in the questionnaire which is not based on a Likert scale as participants were asked to choose 1 if they only spoke a foreign language (often English) and increase this number according to the number of additional languages they spoke. The remaining six items were 5-Likert items regarding their perceived proficiency in English (Item 15), and their willingness to learn more languages (Item 16). Finally, Items 17 and 18, respectively, asked participants whether they were aware of kinesics in their own culture and in other cultures. Items 19 and 20 gauged the participants' awareness of the proxemics in their own or other cultures, respectively. This block, excluding the answer for the number of foreign languages participants spoke, added up to 30 points. The ranges considered are: 30 to 24, highly proficient, from 23 to 18, quite proficient, 17 to 12 averagely proficient, 11 to 6 poorly proficient, and finally 5 to 0 very poorly proficient. It is in this block where participants perceive themselves as less skilled. Only 19.04% ranked highly (2 at 26, 2 at 25), 54.14% quite (1 at 23; 2 at 22; 5 at 21; 1 at 20; 3 at 19), 19.04% poorly (2 at 16; 1 at 15; 1 at 12) and 4.76% very poorly (1 at 11). No participant considers themselves to be averagely skilled. Table 4 sums up the results for items 15 to 20: Highly (5) 1 (5%) 14 (67%) 6 (28%) 1 (5%) 4 (19%) 0% Quite (4) 11 (52%) 2 (10%) 8 (38%) 1 (5%) 6 (28%) 3 (14%) Averagely (3) 7 (33%) 2 (10%) 7 (33%) 13 (62%) 7 (33%) 11 (52%) Poorly (2) 2 (10%) 1 (5%) 0% 4 (19%) 0% 3 (14%) Very poorly (1) 0% 2 (10%) 0% 2 (10%) 4 (19%) 4 (19%) When asked how many languages they spoke apart from their L1 (Spanish), all the participants report speaking at least one foreign language. Clearly, they refer to English, as a B2 qualification in English was compulsory to join the course. It is, though, quite surprising that roughly 25% (5 people) speak from three to five foreign languages. This may mean that this group enjoys a high degree of multilingualism, although does not necessarily mean that they are interculturally competent (see Methodology for more details on this specific item). Regarding participants' level of proficiency in their L2s (Item 15), more than half of the group perceive themselves as highly proficient (5%) or quite proficient (52%), while about one third identified themselves as having an average level (33%) and only 10% consider themselves poorly proficient. When they were asked about their willingness to learn more languages (Item 16), 67% answered highly and 10% quite, meaning they would rather to do so. Only 15% acknowledge that they are not very interested in learning other languages, answering poorly (5%) and very poorly interested (10%). It is also worth mentioning that in this section participants wrote more qualitative comments than in other blocks. In fact, one of the participants, who speaks five foreign languages and expresses her willingness to learn even more, states: (8) the learning of different languages helps me to enrich my knowledge and personal growth.
As for kinesics, 28% and 38% of the informants respectively seem to be highly aware and quite aware of how body language works in their culture, while almost one third (33%) feel they have average knowledge. In contrast, more than half (62%) state they have an average knowledge and more than a quarter (29%) say they know poorly or very poorly how kinesics works in other cultures.
The participants' responses about the proxemics in their own culture (Item 19) and in other cultures (Item 20), were quite contrary to our expectations. In fact, the sum of participants that rated 5 or 4 adds up to only 47% who considered that they knew how distance and space work in their own culture. Another third (33%) feel they have average knowledge. 19% of the participants admit having a poor or very poor knowledge of proxemics in their own culture. When asked about other cultures, roughly half of the participants (52%) think they have average knowledge. We expected participants to reflect a higher level of knowledge of the proxemics and kinesics of their own culture and much lower knowledge of these in other cultures. A plausible explanation might be that they did not really know what these terms meant, as we purposefully refrained from offering any further explanation or definition of the term, specifically in order to raise their curiosity regarding what the concepts might involve and to be willing to find out when they studied the module on interculturality. In fact, three participants left the question on proxemics unanswered, and one participant commented: (9) I don't understand the meaning of proxemics.
Another participant included an extensive marginal note to herself with the definition of this term in her L1, which she later admitted to having looked up online.
In summary, before the training took place, the participants perceived themselves as being highly intercultural, with more than average knowledge of themselves and their own culture. Likewise, they regard themselves as highly tolerant, open-minded and relaxed when taking part in intercultural encounters.
However, in their own perception, their weakness seems to be the use of English as a lingua franca, even though they all had a B2 English qualification.

Self-reflective reports
This section focuses on the results and analysis of the self-reflective reports the participants were required to write after the intervention took place. In this section we will compare the results with those obtained before the intervention in the questionnaire stage. Figure 2 shows an overview of the percentages of comments per dimension: In summary, before the training took place, the participants perceived themselves as being highly intercultural, with more than average knowledge of themselves and their own culture.
Likewise, they regard themselves as highly tolerant, open-minded and relaxed when taking part in intercultural encounters. However, in their own perception, their weakness seems to be the use of English as a lingua franca, even though they all had a B2 English qualification.

Self-reflective reports
This section focuses on the results and analysis of the self-reflective reports the participants were required to write after the intervention took place. In this section we will compare the results with those obtained before the intervention in the questionnaire stage. Figure 2 shows an overview of the percentages of comments per dimension: Regarding Continuous Learning, it is the dimension most frequently mentioned in self-reflective reports. In fact, all the participants (n=21) mention this dimension. From the total of 130 comments gathered for analysis, 46 refer to this dimension (i.e., 35%). This is in line with the results obtained in the questionnaire (where about 89% of participants rated themselves as quite or highly skilled or above), and shows that individuals are still curious and willing to learn: (10) my room for improvement is still very large but my desire to improve as well.
Initially, this might suggest that the intervention did not really have a strong effect on the participants as they were highly curious and they have remained as such because every participant addresses this issue in their self-reflective tasks. However, there is a contrast between their initial self-confidence (as reported in the questionnaires) and their realisation (after the training took place) that the intercultural classroom might entail a different context, requiring a further set of intercultural skills. This might explain why in the self-reflective report trainees mostly focus on their increased awareness of how internationalisation might affect their teaching. As some of the participants reflect: (11) We have raised our awareness of the factors characterizing teaching and learning in the international classroom. (12) This course has not only make [sic] me gain more awareness about the international classroom and its special characteristics, (but also has provided me with specific 'ready-to-use' strategies and resources, which I personally find extremely helpful).
With reference to Interpersonal Engagement, it is less frequently commented upon than Continuous Learning. Of the total of 130 comments, 19.2% refer to this dimension and, surprisingly, 8 participants (out of 21) did not mention it at all. This contrasts with their answers to the questionnaire, where everyone considered themselves either highly or quite highly skilled in their ability to engage with others from different cultural backgrounds. After the training, the recurrent idea in the self-reflective reports is around the importance of fostering students' interaction: (13) Moreover, we can foster integration among international and local students through our teaching methodology. Mutual exchange and cooperation surely enrich the teaching and learning experience. (14) I hope to improve the learning outcomes and promote a great interaction between domestic and international students.
Even though both before and after the training all participants said they found it easy to engage with others, it is after the training that they have become aware this was not going to be an easy task with students, who might be reluctant to interact with international students and remain grouped according to cultural homogeneity: (15) Initially, I may find some barrier on the part of the students when it comes to creating mixed groups between native and international students. (16) some students are narrow-minded and do not share their experiences with the international ones or people who do not have a fluent level in English or another foreign language; therefore the communication could be difficult for them.
The Psychological Dimension is the least mentioned. Only 18 of the 130 comments refer to it (i.e., 14%), and the comments come from only a minority (8) of participants. This contrasts with their answers to the questionnaire, where everyone considered themselves either highly or quite highly skilled. In other words, before the training, most of the participants regard themselves as ready to face intercultural exchanges, with just one of the participants (5%) acknowledging to be poorly skilled when facing intercultural encounters (Item 14). After the training, the tables seem to have turned, with 72% of the comments in this section made by 18 of the participants conveying negative ideas as they now talk about fears, anxieties, barriers and challenges when facing an intercultural classroom: (17) Teaching in a foreign language is a great challenge for me, so speaking into [sic] English (in this case) will be intimidating and also much more complicated than in Spanish. (18) Teaching in English in a multicultural class to mostly non-native English speakers, present [sic] challenges for both students and instructors.
However, 28% of the comments were tagged as positive. In these cases, the word opportunity tends to appear: (19) It has given me the opportunity to open my mind and to discover so many things that had been hidden for me, talking in international classroom environment [sic].
One explanation we consider could account for this contradiction between the questionnaire and the self-reflective reports is that, in their regular intercultural encounters, participants might feel comfortable, but they regard the classroom as quite a different context: Remarkably, only three of the comments were positive, such as that in (23): (23) Thanks to this course, I do feel I have gained confidence to teach in English.
In their responses to the questionnaire before the training, participants indicated that they felt quite confident regarding their language skills (i.e., using English as a lingua franca). In fact, one participant (5%) stated to be highly proficient and 11 of the participants (52%) described themselves as quite proficient. A group of 7 participants (33%) regarded themselves as averagely proficient, with only two participants (10%) rating themselves as poorly proficient.
It could be argued that the intervention did not have the expected results, as one of the trainers' main objectives was to "empower" these lecturers and help them gain self-confidence in EMI contexts. As one of the participants writes: (24) The first thing I need to do is to improve my level of general English in order to gain confidence and ability to face a multicultural group in this language.
If we have a look at the verbal and non-verbal distinction, just one comment (0.40%) refers to non-verbal communication: (25) in relation to the behaviour into a [sic] international classroom, paying special attention to issues that often go unnoticed such [sic] kinesics or proxemics.
We could interpret this as a lack of awareness of the role of the non-verbal in communicative acts even after completion of the course, so maybe more emphasis should be placed on this issue on future in-service training courses.

Conclusions
The aim of this study was to contrast the participants' beliefs about their intercultural competence before and after taking a course related to the topic as part of an EMI in-service training course. With regard to continuous learning, there is a difference between initial self-confidence and participants' realisation after the course that the intercultural classroom is a much more challenging context than they had perhaps realised. Moreover, they also changed their perception with respect to interpersonal engagement, with participants realising that students' reluctance might pose a considerable barrier. All of this may have taken its toll on the psychological dimension, where the intervention seems to have raised the participants' anxiety. At the language level, the training did not fully succeed in empowering these lecturers in EMI contexts. Some participants still consider a high level of language competence to be a key issue for them to become effective EMI lecturers. This paper is not without its limitations. The number of participants is relatively low and does not allow for generalisability. In addition, the use of the original questionnaire in a question format posed problems for the interpretation of the Likert-scale answers. A possible solution might be to fully rephrase the items as statements rather than questions. Finally, other interesting variables such as differences between disciplines, the specific context of each institution, and the fact that the instructors were different in each case, have not been considered for the sake of space. Future research is planned to tackle these issues.