Reverse Subtitling in the ESP Class to Improve Written Skills in English: A Case Study

The improvement of foreign language (L2) skills is of paramount importance in the university context, and a considerable number of students lack linguistic proficiency in their oral and written communication skills. This lack of proficiency needs to be properly addressed and methodologies employed in order to help students improve such skills. There are many cases in which international students, whose cultural background is either European or Asian, need to improve their oral and written production in English as an L2. This paper presents a study conducted at the Universidad Complutense de Madrid throughout the 2017–2018 academic year, in which a number of undergraduate students enrolled in English for Specific Purposes (ESP) used reverse subtitling, from Spanish or Chinese into English, as a learning tool. The main focus of this case study was to enable students to improve their written skills in English through the combined use of Business English and reverse subtitling in accordance with the course content and expectations. A mixed method was used in order to gather and evaluate quantitative and qualitative data. This study therefore aims to bring to the fore the potential of using reverse subtitling in the ESP class.


Introduction
The use of audiovisual translation (AVT) as a didactic tool in foreign language learning (FLL) is gaining more and more importance in the university context, considering the vast body of research as well as the numerous studies conducted in this field in the last decades (Díaz-Cintas, 1995;Talaván, 2006a;Gambier, Caimi, & Mariotti, 2015).
The improvement of skills in the foreign language (L2) class is a necessity which not only concerns students, but also teachers and researchers. In the context of the Spanish university system, some L2 students present a lack of proficiency in English which deserves attention (Dafouz & Camacho-Miñano, 2016). This fact may encourage lecturers to employ methodologies used in the FLL setting, and participants may be more inclined to use them to improve their skills. This study presents an experiment within the AVT scope as it makes use of reverse subtitling as a potential pedagogical instrument capable of assessing and enhancing undergraduate students' written production in English.
The validity of AVT as a tool for the L2 class can be said to be currently well-established and acknowledged. Numerous scholars have put this tool into practice and investigated its potential in the last few years (Díaz-Cintas, 2012;Incalcaterra-McLoughlin & Lertola, 2011;Lertola, 2012;Talaván, 2010Talaván, , 2012Talaván, , 2013Talaván & Ávila-Cabrera, 2015). As for the type of subtitles, Talaván and Rodríguez-Arancón (2014) refer to reverse subtitling as an active type of subtitling in which the students subtitle from their mother tongue into an L2. This type of subtitles is the focus of interest here.
In this paper, attention is paid to the manner in which students write in English by assessing a number of compositions written in this language while simultaneously practising subtitling in an active way (Sokoli, 2006;Talaván, 2006a). By making use of a multi-strategy design in the form of quantitative and qualitative data, this experiment presents a series of activities which account for the potential of reverse subtitling when several languages are used within a Spanish university context.

Research on Subtitling in the FLL Context
It is well known that translators usually transfer texts from a source language (SL) into a target language (TL) and in most cases, these are translated into their mother tongue. When putting this into practice in the FLL context, it can be surmised that if students subtitle from their mother tongue into an L2, they will be more fully exposed to the TL. This practice will help them improve their L2 written skills given that "translation facilitates foreign language encoding" (Danan, 1992, p. 497). However, there is a challenge in this type of practice as the TT may not always sound as idiomatic as the one produced by a native speaker, especially when this type of L2 pedagogical tool is used by students.
Díaz-Cintas presents the conventions of subtitling as part of a teaching technique (1995), discusses how the audiovisual media can be used in FLL (1997), and deals with the importance of subtitling in the L2 class (2012). Among some of the studies conducted on subtitling, Talaván (2006b) looks into the potential of subtitling presented as a technique to improve business communicative skills. Talaván (2010) also presents a theoretical-practical account of the pedagogical applications that emerge from subtitling in the teaching of an L2 (2013). She has also conducted experimental studies in collaboration with other authors on reverse subtitling in a collaborative and online context (Talaván & Rodríguez-Arancón, 2014), intralingual captioning for the improvement of writing and vocabulary skills (Talaván, Lertola, & Costal, 2016), the benefits of combining subtitling and dubbing for the improvement of both oral and written skills (Talaván & Ávila-Cabrera, 2015), and collaborative reverse subtitling activities in English as an L2 (Talaván, Ibáñez, & Bárcena, 2017). As can be inferred, the possible combinations and types of subtitles in the FLL setting are many.
The benefits of reverse subtitling for incidental vocabulary acquisition are explored by Lertola (2012) and also how subtitling as a task can be best used by teachers of languages (2015). Furthermore, Lertola and Mariotti (2017) analyse reverse dubbing and subtitling involving Italian learners of English as an L2, which proved that a combination of different AVT modes can be optimal for the improvement of different skills, both oral and written. Ávila-Cabrera (2018) presents a practice guide which enables secondary education teachers to use engaging subtitling activities to help students improve their written skills in different language combinations. The study conducted by Alonso-Perez and Sanchez Requena (2018) highlights the presence of AVT modes in FLL, which benefit different learning areas such as production and comprehension skills and intercultural awareness, among others.
It is important to mention that students can use subtitling in a passive or active way. The aforementioned studies concern the latter, that is, participants undertaking subtitling tasks, and it could therefore be highlighted as a practice that offers more benefits in writing skills to learners of foreign languages given the exposure of the participants to the L2 in its written form.
With regard to more recent terms used in the literature of subtitling, Díaz-Cintas (2018) presents a taxonomy of subtitles in accordance with the new forms which are emerging nowadays. These include "cybersubtitles" (divided into "fansubs," "guerrilla subtitles" and "altruist subtitles"). They can also be regarded as "genuine subtitles" or "fakesubs". Chaume (2018), on the other hand, gives an account of the new changes in the subtitling types by resorting to the term "creative subtitles" to refer to those which do not follow the traditional screen positions, size, orthotypographic conventions, etc., as it is the case with multilingual programmes for example. As can be observed, new terminologies are currently gaining an established position in the AVT panorama.
Finally, it is important to note the number of recently published manuals and books which delve into the techniques and strategies used in AVT, mainly with English-Spanish combinations such as those documented by Bartoll (2015) and Talaván, Ávila-Cabrera, and Costal (2016). These authors present a general approach to AVT, and AVT accessibility respectively. Rica Peromingo (2016) deals with technical and linguistic aspects in AVT. Roales Ruiz presents an AVT teaching proposal (2017) based on linguistic and technical conventions of subtitling, and also another on how to learn to subtitle thorough the subtitling editor LRN-2sub (2018), of his own creation. The aforementioned publications can provide lecturers and practitioners with useful linguistic and technical tools to use in the FLL setting.
The paper aims to contribute to the state-of-the-art research on subtitling in the FLL setting, as summarised in this section, through a detailed study of the benefits obtained by L2 students of ESP through the use of active reverse subtitling. By conducting a preliminary study that evaluates this AVT mode, we intend to present an analysis of this tool through a discussion of its results in order to promote its uses in the classroom.

The Experiment
This study 1 was conducted during the first semester of the 2017-2018 academic year by the author at the Universidad Complutense de Madrid. The participants were undergraduates from the second course of the Degree in Commerce who were enrolled in an English for Specific Purposes (ESP) course, and who worked exclusively from an international course book on Business English at the B1-B1+ level. With the aim of obtaining data from the students, two groups were chosen, each composed of 25 participants. The experimental group (EG) consisted of those who completed the following: two composition tasks (see Appendix 1 and 2), vocabulary and grammar tests, reading comprehension exercises, listening exercises, along with a reverse subtitling task (as the tool used for the purpose of this study), and a composition in the written exam (see Appendix 3); in addition, the control group (CG) comprised another 25 students, who completed the same exercises, with the exception of the subtitling task. In an attempt to obtain data that could be analysed and compared in more equal terms, each of the groups included 5 native speakers of Chinese, i.e. 20%.
The EG students were given a number of basic instructions on subtitling conventions to guide them in their task of subtitling a video of their choosing, the topic of which had to be related to business, trade or finance. The task of reverse subtitling included audio in Spanish or Chinese (the mother tongues spoken by the students under analysis) to be subtitled into English. In order to obtain quantitative data, the students were asked to submit two compositions in English to verify the improvement of their written production skills throughout the semester; the subtitling task was completed between the first and second composition. Having finished these tasks, the students completed a questionnaire (see Appendix 4) about their opinion of this AVT practice, which was then explored and evaluated to elicit qualitative information.
The students used a video subtitling platform called Amara 2 and were given the opportunity to contribute to the accessibility of online videos by uploading these subtitles to YouTube. This practice can be said to be related to a newfound respect defined as "social subtitling," which was referred to by Ávila-Cabrera (2016, 2017) as "the voluntary collaboration to promote media accessibility at the university and other settings where there is not sufficient funding to hire professional subtitlers to cover the vast needs created by the Information and Communication Society" (2017, p. 287). In this sense, the subtitles produced by the participants in the study could be considered altruistic when being shared on platforms and websites such as Amara and YouTube. Audiovisual materials are therefore made accessible from the university context to different types of viewers who may not be competent in the languages presented in the audio of the videos. When talking about media accessibility, we need to bring to the fore Greco's (2016Greco's ( , 2018 proposal in the field to extend the definition and scope of media accessibility beyond people with sensory disabilities in order to include those who for diverse reasons cannot access audiovisual content in its original form. Regarding this case study, media accessibility would then be understood as the latter account, that is, making videos comprehensible to users through subtitles in English with the aim of overcoming linguistic barriers.

Research Design
The mixed method (Robson & McCartan, 2016) was adopted as a means to gather quantitative data, which included compositions submitted by the students throughout the semester together with the subtitling practices and the composition of the final written exam. Further qualitative data was gathered in the form of a questionnaire containing questions on the practice itself. The author made use of the methodology implemented with the groups he was teaching and was able to conduct the study on the basis of the EG's and the CG's written outcomes, for which reason the research design cannot be said to be randomised. However, the selection of the participants in each of the groups was random in order to avoid biased results.
Triangulation allows researchers to observe a phenomenon from different perspectives. In this case, triangulation was to be achieved through the analysis of quantitative and qualitative data as explained above. As for the data analysis, the sequential explanatory design was implemented in the following manner: observations were made about the quantitative data, then the qualitative data was analysed in an attempt to verify or refute the findings (Creswell, 2003).
A further aim of this study was to assess the validity of subtitling from Spanish or Chinese into English, while addressing these specific nationalities of students in the classes under study. There were also students from other European countries, but they were not considered for the purposes of analysis as they did not have Spanish or Chinese as their mother tongue. Ultimately, the goal of this study was to provide further evidence in support of active subtitling (Talaván, 2012) as a tool for the improvement of written production in English in the ESP class.
The research question at the core of the methodology can be seen as follows: Is reverse subtitling a valid tool for the written skills improvement in an L2?
In order to address this question, observation was focused on the progress made by the students in the EG when writing in English, including subtitling activities, which were used as a means to achieve the goal. The CG's results were also compared with those of the EG with the aim of addressing this question and discussing the possible differences encountered. In addition, the EG's answers to the questionnaire were analysed in order to determine whether or not the quantitative results could be corroborated by their opinions following their participation in the experiment.

Timeline
The varying tasks were designed to ensure that the gathering of data was in accordance with the dates seen in Table 1. As for the evaluation criteria of the ESP subject, all tasks submitted by the students accounted for 10% of the final mark. Additionally, an oral presentation (10%) and active participation in class (10%) were also assessed. Last, the final written exam, which represented 70% of the final mark, included a composition as the object of study. This writing task was also evaluated to verify or refute any positive enhancement of the EG students' writing skills in English, and was also compared with the CG's results.

Procedures
Before the students started subtitling on Amara, an introduction to AVT and subtitling (Díaz-Cintas, 2001;Díaz-Cintas & Remael, 2007) was presented at the beginning of the course. While most students generally knew what subtitling is, they had to learn some basic concepts on the technical restrictions and linguistic considerations of this AVT mode. The following pieces of information were given to the students. Table 2.

AVT and General Subtitling Conventions
-Definition of AVT and subtitling -Subtitling conventions: ▪ Subtitles should appear on screen for a minimum of 1 second and a maximum of 6.
▪ One-line subtitles should have a maximum of some 35 characters and two-line subtitles some 70. ▪ Only the gist (the most important information both linguistic and paralinguistic) should appear in the subtitles. ▪ Segmentation or line breaks: subtitles must be semantically and syntactically properly segmented (1 or 2 lines), that is, the division of the subtitle should not split the phrases between the lines.
-For further information please read the BBC subtitle guidelines: http://bbc.github.io/subtitleguidelines/ Source: Díaz-Cintas & Remael (2007) Regarding segmentation, a sample with a correct and incorrect line break was provided so the students could have a clearer idea of this essential concept as shown in Table 3. Table 3. Once the students became familiar with the platform, a sample 3 of The Hateful Eight (Quentin Tarantino, 2015) trailer, with audio in Spanish and subtitles created in English by the author, was uploaded to Moodle, the virtual platform used by the university. In addition, some aspects of subtitling conventions, segmentation, etc. were again revised in class. Next, the students were provided with a series of steps on Moodle on how to proceed with the activity. It must be noted that this was the first time these students were required to subtitle videos, as it will be shown later on in the discussion of the questionnaire results. Moreover, they were not familiar with translation strategies or techniques (Talaván, 2017) given that their studies were not related to translation. Thus, very general instructions were given to them regarding how to subtitle. It was also made clear that professional quality results were not expected.

Data Analysis and Discussion
Quantitative and qualitative data were gathered and analysed for the purposes of this experiment as shown in the following sections. In addition, a discussion on the analysis of data and results is presented here.

Written Production Assessment
Throughout the course, the students submitted several compositions about business, trade and finance topics that were produced with the aid of online dictionaries so that their real written production could be assessed and feedback provided. This gave the students the opportunity to improve their skills before the final written test, in which no dictionary is allowed in accordance with the ESP final assessment guidelines of the Facultad de Comercio y Turismo (Faculty of Commerce and Tourism) of the Universidad Complutense de Madrid. A mock composition was corrected at the beginning of the study, although this did not count as a mark for the course. In this case, the mock writing served as a guide for the students so that they knew what was expected from their writing production tasks. Following this assignment, they submitted composition 1, completed the subtitling activity, submitted composition 2 and, finally, wrote the composition of the final written exam, which was also assessed as part of the data for this study. For the correction of the compositions, a rubric 4 based on the B1 assessment scale (which assesses content, communicative achievement, organisation, and language) by Cambridge English was used.
The reverse subtitling practice was done in different sessions and with the help of online dictionaries to be able to condense the TT as much as possible. In order to assess this task, a subtitling rubric was used (based on Talaván, Ibáñez, & Bárcena, 2017). Table 5 shows the items assessed.  (2017) With the aim of addressing the needs of a large number of students from abroad, Chinese native speakers subtitled from Chinese into English. Given the researcher's inability to understand Chinese, this entailed a limitation to this study given that the evaluation of Chinese students' subtitles had to be based on the TT rather than on the transfer from the ST to the TT. However, this aspect did not affect the Chinese students' composition assessments, which were the main focus on the study.
As can be seen in Table 5, the different aspects under analysis are the following: (1) Accuracy concerns the linguistic and punctuation marks conventions. The Spanish students subtitled a video from Spanish into English. Accordingly, Chinese native speakers subtitled from Chinese into English. In order to avoid a drawback when analysing the data (considering the author could not understand Chinese and their accuracy was centred on the TT linguistic features), the same number of Chinese students in the EG was also included in the CG, that is, 5 in each group in order to have more homogenous groups in terms of their mother tongue. All in all, both native speakers of Spanish and Chinese were assessed, although not in the same terms.
(2) As for condensation, the scale shows how condensed the subtitles in the TT were.
(3) A proper syntactic and semantic segmentation was also evaluated, as well as (4) synchrony between the sound and the subtitles. These last two aspects were evaluated, but their value was lower than that of accuracy and condensation, considering that these students are not enrolled in university degrees in which translation is learnt in detail and practised regularly.
Regarding the rubrics used for both the composition and the subtitling task, they have different features, considering that they are aimed at the assessment of two different tasks, although both of them deal with written production in English. Most importantly, it must be pointed out that while the improvement of written production in English in the form of compositions is the object of study, reverse subtitling as an aid in improving these skills. Thus, the subtitling task is the instrument employed to determine whether or not the goal can be accomplished.
Regarding the assessment of the EG's and CG's compositions, the results obtained can be seen in Figure 1.

EG's and CG's Composition Assessments
Paying close attention to the data shown above, an improvement was observed in the EG students' written production considering the rise of the average, which was 6.39 points (median 6 and mode 5.5) in the first composition, 6.72 points (median 6.5 and mode 5.5) in the second and, finally, 7.41 points (median 7.5 and mode 9) in the composition of the final written exam. Thus, an improvement of +1.02 points was observed from the beginning of the course to the end.
The subtitling activity could not be compared with the composition in equal terms, as it was an anecdotal datum. The results revealed that the average mark, median and mode (all of which accounted for 7.14 points, 7.5 and 8 respectively) were higher than the results obtained in the first and second compositions of the EG, but below the average mark of the exam composition. This could be explained as the students were given more sessions in which to complete the subtitling task and

Composition assessments
Composition 1 Composition 2 Exam composition might have found subtitling a more dynamic and entertaining activity. The final results obtained in the exam highlighted the quantitative improvement as shown in Figure 1.
As for the CG, in their first composition, the average was mark 5.52 points, the median 5 and the mode 4 respectively. In the case of their second composition, they obtained a mark of 5.81 points, a median of 5.5 and a mode of 6.5. Focusing on their exam composition, the improvement can be seen in an average mark of 6.18 points, a median of 6 and a mode of 6.5. In this case, the CG showed an improvement in their average mark of +0.66 points.
It can be observed that the EG exhibited an improvement in their written skills in English throughout the semester, as already discussed (see Figure 1). In addition, it was also revealed that the EG's final results (average 7.41 points) were considerably higher than those of the CG (average 6.18 points), who completed the same tasks with the exception of the subtitling practice. The difference in their improvements positions the EG +0.36 above the CG. It can be seen that the result showed more positive outcomes by the students who practised their written skills throughout the semester with reverse subtitling as a tool used to improve their writing skills. It can be argued that having the opportunity of practising one's skills in a more recurrent manner should benefit participants in their outcomes. In addition, we cannot ignore the fact that the EG's compositions were marked higher than the CG's and this could lead us to consider that the higher the level of the students is, the quicker their progress is.
With the aim of using a more thorough statistics formula, the independent sample T-test and the one-way analysis of variance (ANOVA) were also applied to analyse the quantitative results of the means of the three tasks. One the one hand, the results of the T-tests show that the t-value (2.11) is slightly higher than the critical value (2.01) and that p-value (0.04) is significant at p < 0.05, so it seems plausible to reject the null hypothesis and claim that the means of two independent groups are significantly different. On the other hand, ANOVA returns a statistically significant result given that the statistical value of the analysis (F= 4.46) is considerably distinct from 1 and the p-value (0.03) is lower than the significance level of 0.05. Given these data, it seems reasonable to accept the alternative hypothesis that the EG's and CG's means are statistically significantly different from each other.
In view of the results provided by the two tests, we could assume that there is a direct relationship between the two variables, namely the improvement of the EG's written production and the reverse subtitling task, used as a tool for the enhancement of written skills. In conclusion, it can be stated that, within the parameters and context of this study, students' learning did not develop independently of the use of a particular teaching methodology.

Questionnaire
Qualitative data was gathered in the form of a questionnaire in which the students of the EG were asked about certain aspects of their linguistic profile as well as their experience on this subtitling practice. Their answers are discussed below.
Although, up to 37 students practised with the subtitling task, we only analysed 25 random answers from the EG since we could only gather 25 participants of the CG who had completed all tasks of the course with the exception of the subtitling practice. However, in other experimental studies, different numbers of participants can be found in the EG and CG. Non-native speakers of Spanish or Chinese were excluded in the sample so that the results were not jeopardised. The participants were aged 20-22, and 64% were female students and 52.7% were male.
The students' mother tongues were two: 20 students were native speakers of Spanish (80%) and 5 were native speakers of Chinese (20%). Within the group of native speakers of Spanish, the nationalities involved were varied: 14 students were from Spain (56%), 2 from Colombia (8%), 1 from Equatorial Guinea (4%), 1 from Argentina (4%), 1 from Paraguay (4%), and 1 from Peru (4%). Hence, the two main languages spoken by all participants were the languages chosen for the transfer from Spanish or Chinese audio into English subtitles.
As for linguistic questions, the most significant data were that only 4% considered themselves to have an advanced level of English including writing and speaking skills, 20% stated that they could have an upper-intermediate level, 36% accounted for an intermediate level, and the remaining 40% situated their skills at a lower-intermediate and beginner level. It can be observed that the predominant group exhibited a lower-intermediate and beginner level in accordance with their own views. However, these data cannot be considered in scientific terms since no proof of a language certificate was required of the students. The purpose of this question was, rather, to observe how the participants felt about their linguistic proficiency in English.
For the question regarding the improvement of reading skills, a Likert scale was used on the basis of 4 responses in an attempt to avoid a neutral choice: 1 (not at all), 2 (slightly), 3 (very much), and 4 (extremely). Accordingly, the participants were asked if they believed their reading skills in English had improved: 4% indicated that they did not improve at all, 28% improved slightly, 60% improved very much and 8% extremely, answer 3 being the most common. If we add the percentages of the students who felt a positive enhancement of their skills, we obtain 68%, which can be said to be recurrent in the majority of participants. Thus, this positive opinion seems to be predominant.
The most important question, which sheds light on one the main goals of the study, can be seen in Figure 2.

Improvement in Writing Skills in English
Based on the above figure, those students who considered themselves to have improved in their written skills when subtitling from Spanish or Chinese into English account for 56%, 48% having chosen very much, and 8% extremely. By contrast, the percentage of students who indicated only a slight improvement was 40%, and those who considered no improvement at all was 4%. The results discussed here indicate a positive tendency of the improvement of writing skills. Accordingly, it could be claimed there is a direct relationship between writing and reading in the same language as both skills are involved in the production of the former. Thus, triangulation of data can be said to corroborate the quantitative (with positive outcomes in writing skills of the EG) and qualitative results (a positive view) discussed here.
Other questions concerned the improvement of speaking skills with the following results: 4% indicated that they did not improve at all, and 52% slightly; by contrast, 40% stated to have improved very much, and 4% extremely. These data indicate a more negative tendency regarding the fostering of this oral production skill, which can be explained by focusing on the conventions of reverse subtitling as all the participants of the EG subtitled from audio in their mother tongues into written text in English. As far as the listening skill is concerned, the data are not relevant here since the audio was in the students' mother tongues.
To illustrate the subtitled videos submitted by the EG, the following example is an excerpt with audio in Chinese subtitled into English, which presents news segments on the Asian company Alibaba broadcast by TVBS news:

Chinese to English Subtitling Sample
Source: https://goo.gl/NHx2dD The following video has audio in Spanish and is subtitled into English. Here Victor Küppers talks about motivation in the clip XXIII Convención Grupo Expofincas (13 th Grupo Expofincas Convention).

Spanish to English Subtitling Sample
Source: https://goo.gl/jdHv25 Last but not least, the students were asked if they would like to continue subtitling with Amara. It is true that this was the first time participants subtitled videos, but we aimed to observe the way the students felt about using this platform. 96% indicated that they would like to continue subtitling with it and among the responses obtained were: "Yes, because it is easy to work with this platform", "This webpage has helped me a lot, it's very easy to use and a very innovative way to learn English", "Yes, because it is fun", and "Subtitling helped me read more fluently in English, everyday expressions can be easily learnt." In general terms, they stated to have enjoyed editing the subtitles and learning how to synchronise the text with the audio, that Amara was a user-friendly platform, as well as the fact that reverse subtitling was a good way to improve English written skills. Some responses concerned accessibility, as can be seen in the following: "Yes, I think it's a very good idea to make videos accessible to everybody", "Yes because you can learn new vocabulary and you can help other people to understand other languages that they don't know". It would be interesting to conduct further research on the promotion of linguistic accessibility to audiovisual content in these types of studies.
To summarise, positive outcomes were reported by the participants. It could therefore be inferred that this was a rewarding activity for the students.

Discussion
The data gathered in this study is directly related to the students' performance in written English by using reverse subtitling, in an active way, as an L2 learning tool. The research question established is shown and answered as follows.
• Is reverse subtitling a valid tool for the written skills improvement in an L2?
In order to address this question, observation was focused on the progress made by the students when writing in English, including the subtitling tasks and the composition of the final exam as a means to reach the goal.
If we observe the quantitative results from the EG, the data show an improvement of +1.02 points.
As for the subtitling practice, the average mark was higher than the first and second compositions under analysis. It could therefore be inferred that although the subtitling activity cannot be directly compared with the numerical data obtained from the composition assessments, the subtitling task was acknowledged by the students to have been a positive practice as indicated in their answers in the questionnaire. In addition, reverse subtitling was an innovative tool for them, hence the positive results that could be related to this fact. On the other hand, the average marks obtained by the CG were lower than any of the previous quantitative marks obtained by the EG. This group, nonetheless, showed an overall improvement of +0.66 points. Accordingly, the difference between the EG's and the CG's improvement was +0.36 points in favour of the former.
After applying the statistical formulas of the T-test and ANOVA, it could be said that the improvement in the participants' written skills in English with the use of this reverse subtitling tool is not accidental. Triangulation of data can be said to have been achieved by observing this quantitative improvement in the EG's results, along with the participants' positive views regarding the way their written skills could have improved by making use of reverse subtitling.
Based on the responses obtained from the questionnaire, the students mentioned that they believed their reading skills had improved: 68% indicated that they felt that they had improved very much and extremely. This observation could concern the necessary operations to transfer an audio in an L1 into subtitles in an L2, that is, in English in this case. Speaking and listening skills were not considered by the students given the features of this subtitling task.
Regarding the participants' opinions on the benefits of reverse subtitling, it must be mentioned that they indicated that they found the platform user-friendly, learnt new expressions, and found this activity to be an innovative manner in which to practise their skills in English. The benefits of making videos accessible to more users was also mentioned, but not explored in depth. However, this first step is essential given that the majority of participants expressed their willingness to use this platform in the future, which might lead to further research in which the importance of accessibility of audiovisual content can be one of the goals of studies of AVT in the FLL context.

Conclusions
This paper has attempted to support claims by scholars who have investigated the potential of AVT modes in the FLL context and has also shown how important it is for undergraduate students to use reverse subtitling or active interlingual subtitling as a means to improve written skills in foreign languages. Both lecturers and researches need to consider methodologies that can be beneficial to students as learning aids, but also attractive to them as 21 st -century learners. It has been shown that the participants in this particular study tended to welcome activities that contained audiovisual content and AVT. The potential of subtitling as a tool for the ESP class has been described and assessed. Thus, the goals established in the research design have been addressed.
This study investigated whether or not reverse subtitling can be employed as a tool for the improvement of written skills in English, and what other benefits its active use can entail. Given the results obtained and the small number of students under analysis, it can be said to provide evidence of a case study.
Focusing on the average marks that the students in the EG obtained through the evaluation of their written production tasks, the improvement of their writing skills in English materialised in +1.02 points. The students from the CG obtained a higher average mark in their exam composition, but their improvement (+0.66 points) was lower than the EG's. Based on the statistical formulas, there could be a high probability that a correlation between the subtitling tasks and the final outcomes of the EG exists, making reverse subtitling a pedagogically valid tool in the FLL context.
As for the qualitative data explored, it can be noted that reverse subtitling was positively received by the EG, as an L2 tool in the ESP class, as the majority of participants indicated a willingness to use Amara in the future. The research question was triangulated when students stated to have improved their written skills and when comparing the quantitative results. As for the questionnaire responses, the students mentioned they had also improved reading skills and the importance of helping other viewers consume audiovisual programmes whose audio tracks are not in their mother tongue. This idea therefore supports the fact that although the participants' cultural backgrounds were different, with Spanish and Chinese being the native languages involved, subtitling in the university context and amateur subtitling can help in overcoming the linguistic and cultural barriers present in communication.
As for the limitations in this study, the reduced number of participants might have, in one way or another, had a direct influence on the quantitative results. In addition, more assessments of written production and subtitling activities can offer researchers more thorough results when assessing the potential of a pedagogical tool. The researcher's lack of proficiency in Chinese also affected the manner in which the subtitling tasks of the native speakers of Chinese were evaluated. However, this fact did not directly affect the quantitative analysis of the results of the composition tasks submitted by these students. Thus, the main goal of this study has aimed to prove how in this case study the use of reverse subtitling may have benefitted the participants' written skills in English in the ESP class.
Aiming for the promotion of using reverse subtitling in the FLL setting, a future research avenue might be to observe whether or not the students reflect on the importance of making videos linguistically accessible to other users, as supported by "social subtitling." All in all, this preliminary case study helped shape an innovative teaching programme conducted in the same university for the 2018-2019 academic year within the Innova-Docentia programme, under the name SubESPSKills (Subtitling tasks in the English for Specific Purposes class to improve written production SKills) (Ávila-Cabrera & Rodríguez-Arancón, 2021). A group of lecturers from different Spanish universities (as well as two undergraduate students to help with the Chinese transcripts) designed a number of subtitling activities for the improvement in written English in the ESP class. It is hoped that the participants' enthusiasm and completion of tasks can provide the researchers with fruitful results that can contribute to the literature on AVT as a teaching tool in the FLL setting.