Technology-Enhanced Language Learning: Subtitling as a Technique to Foster Proficiency and Intercultural Awareness

Abstract

Computer-Assisted Language Learning (CALL) is an umbrella term that encompasses diverse technologies with the purpose of enhancing language learning. In the existing literature on CALL, intercultural awareness and the pedagogical use of multimedia products have received less attention. This study explores how the process of creating subtitles for short clips may enhance language skills and intercultural awareness when implemented through lesson plans designed following the framework proposed by the TRADILEX project. A pre-experimental longitudinal design was implemented. The sample consisted of 43 participants who were enrolled in a B2 English course at the University of Córdoba (Spain). During the course, participants consistently attended theoretical sessions. The intervention took place during the practical sessions from February to April 2024, involving four subtitling-based lesson plans on literature and gender. After the intervention, the practical sessions shifted to a traditional, textbook-based format from April to June 2024. The instruments employed to assess the effectiveness of the intervention consisted of a commercial test by MacMillan and the ERI scale on interculturality. The results showed that after the intervention, there was a significant improvement in language proficiency, which increased at a slower rate during the traditional sessions. However, when it comes to intercultural awareness, there was a peak of attainment after the intervention, but attrition rapidly took place. Regarding the pedagogical implications of this study, subtitling could be an appropriate technique that allows contact with the L2 culture and shows positive effects in terms of proficiency.

1. Introduction

Technological advancements have had a great impact on a multiplicity of spheres, and education is no exception. Computer-Assisted Language Learning (CALL; Gillespie, 2020), Mobile-Assisted Language Learning (MALL; Bárcena, 2020) or Artificial Intelligence-Assisted Language Learning (AIALL; Huertas-Abril & Palacios-Hidalgo, 2023) are great examples of applications of technology to language learning. From the use of computers to mobile phones or AI, researchers have explored the benefits of the implementation of these technologies for language education. In this regard, Gillespie (2020) offered a comprehensive taxonomy of the most researched topics in CALL, listing them in the following order: writing, computer-mediated communication, vocabulary, speaking, corpora, NLP (Natural Language Processing), design, teacher education, reading, listening, web 2.0, grammar and feedback. This order is important because it provides a general outlook of the most researched topics in CALL. In contrast, Gillespie (2020) highlighted that intercultural awareness and multimedia were topics that have received less attention when it comes to CALL and added that translation, digital literacies and literature are topics that have been scarcely published in this field. Given this gap, this study aims to contribute to the field by focusing on the intersection between multimedia and intercultural awareness through pedagogic translation. It is important to acknowledge that CALL is difficult to define due to the wideness of its scope. Gallego Espinosa (2018) affirmed that this term may allude to any educational process linked to language learning in which computers play a significant role. Therefore, the use of multimedia products can be considered as part of CALL, and there is a wide range of opportunities and options available for including these products in language learning contexts. For instance, Wang and Lee (2021) found out that when it comes to gloss presentation, students showed a higher preference for multimedia glosses as they enriched the text and facilitated the understanding of new words. Other studies addressed the way in which the inclusion of short video lessons in online courses enhances engagement (Castrillo de Larreta-Azelain & Martín Monje, 2016). There are also studies focusing on short videos such as those on TikTok. These studies focus on the cognitive burden, its relationship with the number of likes, and the possible impact of their use on language learning (Zhang et al., 2023).

One of the topics that has been traditionally addressed is the use of subtitled materials in language learning contexts. In this respect, subtitles or captions are elements that appear on screen that translate or reproduce the dialogues (Díaz-Cintas & Remael, 2021). It is important to clarify that subtitles are not mere transcriptions, as they need to adhere to specific regulations related to time and spatial constraints. They can either be in the same language as the video or function as a translation (Ogea-Pozo, 2020a). There is a specific field in Translation Studies (TS), which is Audiovisual Translation (AVT). It explores different modes like subtitling, voice-over, dubbing, Audio Description (AD), or Subtitling for the Deaf and hard of Hearing (SDH). This type of specialized translation is closely related to technology due to the very nature of the texts involved in the process. With reference to this, Bolaños-García-Escribano (2025, p. 155) explained that translation is narrowly linked to technology due to the fact it “progressively evolved from prints and typewriters to computer text editors and started to make the most of new specialist software”. In this case, there is a wide range of software devoted to specific AVT modes (Bolaños-García-Escribano, 2025). As Ogea-Pozo (2018) clarified, when it comes to subtitling, translators use editors like AEGISUB (v. 3.4.2.) or Subtitle Edit (v. 4.0.11). However, it is important to take into consideration the distinction made by Talaván (2010), who argues that subtitles serve as a support while subtitling functions as a task that enriches language learning.

Research on the benefits of using subtitled materials in language learning is extensive. Different studies have examined the benefits of implementing subtitled materials (subtitles as a support) into the language classroom, with Vanderplank (1988) being one of the trailblazers in this field. Subsequent studies like the one by Frumuselu et al. (2015) empirically proved that English language learners preferred intralingual subtitled materials as they perceived that consuming a video in English with English subtitles was more effective than having the captions translated into their L1. Alotaibi et al. (2023) examined the impact of the inclusion of subtitles into the L2 classroom and concluded that the results obtained in the meta-analysis suggested a positive and significant effect, but the authors warned that there were factors such as language skills, learners’ L1 or educational level affecting the results, and that is why they should be addressed with caution.

Whereas subtitles as a support could be of great interest in language learning, subtitling as a task involves an active and task-based approach to teaching languages in which the learners are not only consumers but also producers of audiovisual products, which are known as prosumers (Bolaños-García-Escribano & Navarrete, 2022). It is important to emphasize that Talaván et al. (2024) coined the term Didactic Audiovisual Translation (DAT) to refer to the process of using audiovisual translation as a means to teach languages, and it comprises different modes like didactic subtitling, voice-over, dubbing, AD, SDH or free commentary among others.

It was Talaván (2013) who designed a task-based method based on the principles of active methodologies to include subtitling in foreign language lesson planning coherently. After decades of testing, Talaván and Lertola (2022) proposed an instructional structure for the design of DAT-based lesson plans by conceiving them in four stages that align with the PPP (presentation, practice, production) approach (Harmer, 2015). According to the authors, these lesson plans consist of the following stages:

• Warm-up: This is the first stage of the lesson plan. It should include oral and written reception tasks related to the video, which has been specifically chosen for the lesson plan, as well as exercises about grammar or vocabulary to reinforce these aspects.
• Video viewing: In this second stage, learners are shown the video, and they should answer a set of questions to ensure they have fully understood the contents of the video they would later be asked to translate. It is important to note that the video should not last more than 2 min.
• DAT task: This is the most challenging stage of the lesson plan, as students are expected to undertake their own translation of a 1 min excerpt retrieved from the video they viewed in the second stage, according to the AVT mode (subtitling, voice-over, dubbing, AD, SDH or free commentary) and the language combination established by the teacher (keyword captioning, interlingual, intralingual or creative). Stages 1 (warm-up) and 2 (video viewing) are essential for completing this task, as they provide the specific scaffolding needed for students to produce their translation.
• Post-DAT task: This final stage aims to consolidate the new knowledge through a mediation task. If the DAT task involves captioning, the task should include mediation and oral production. For DAT tasks based on revoicing, this final stage should involve mediation and written production.

The LP(s) could be delivered in PDF format or in Google Forms, together with a link to free captioning software (e.g., AEGISUB v. 3.4.2) or free revoicing software (e.g., LOOM v. 5.5.82 or Audacity v. 3.6.). They can also be presented on platforms like TRADILEX (https://www.tradilex.es/ (accessed on 2 September 2024)), which were specifically designed for DAT environments. LPs are usually part of a sequence, and the effectiveness of the sequence is usually tested in the field of DAT studies (Plaza-Lara et al., 2024). However, there are other studies, like the one by Ogea-Pozo and Talaván (2024), that assessed the effectiveness of a particular LP.

As previously mentioned, CALL is an umbrella term for the inclusion of computers in language learning in different ways. In the context of CALL, DAT could be defined as a teaching technique that involves producing new audiovisual materials through adapted and scaffolded audiovisual translation tasks and the use of specialized software and platforms. Before proceeding to examine the state of the art of subtitling, it is important to discuss the theoretical framework behind this technique, delving into its educational background.

1.1. Didactic Subtitling: Educational Background

The efficacy of consuming subtitled materials for language learning has been widely supported in the literature. Research has shown benefits across various domains, including materials for vocabulary (Hestiana & Anita, 2022), content comprehension (Negi & Mitra, 2022), language acquisition (Frumuselu et al., 2015) or interculturality (Black, 2022). Vanderplank and Feng Teng (2024) provide a solid theoretical framework based on the Cognitive Load Theory (CLT) and the Dual Coding Theory (DCT). According to CLT, when applied to multimedia products, captions can improve comprehension and reduce the cognitive load (Paas & Sweller, 2014). The DCT has a cognitive nature and refers to verbal and non-verbal stimuli, suggesting that they are stored in different parts of the brain, which are interconnected. According to this theory, exposure to multimodal stimuli facilitates the ability to recall and retrieve information. Classic studies in language learning and acquisition (Krashen, 1981) focus on three concepts: input, output and intake. The input hypothesis by Krashen (1989) indicates that if an individual is exposed to a comprehensible input, language acquisition takes place. Therefore, it may be inferred that being exposed to audiovisual input whose comprehensibility is enhanced by captions might foster acquisition. In this regard, Danan (2004) identified potential advantages of using subtitles for learning languages, including the development of production skills, the improvement of word recognition and vocabulary building and the possible positive effects on language anxiety.

Notwithstanding, there are several important differences between subtitling as a task and subtitles as a support, as Talaván (2010) highlighted. Subtitling goes beyond the mere consumption of captioned materials as it involves interaction with the multiplicity of the channels and interaction with its semiotic elements. Talaván (2013) establishes subtitling as grounded in TBL (Task-Based Learning), as the former conception of the four-stage lesson plan model by Talaván and Lertola (2022) proves. When it comes to captioned viewing, there are models that address cognition and emotions, like the one by Vanderplank and Feng Teng (2024). This model, together with the proposal by Talaván and Lertola (2022) and the educational bases for didactic subtitling in Didactic Audiovisual Translation and Foreign Language Education (Talaván et al., 2024), have grounded the proposal of a cognitive-affective model of language learning through this DAT mode (Figure 1).

The third stage of Talaván and Lertola’s (2022) proposal is of special interest when it comes to developing the cognitive-affective model since it is one that is innovative and requires the use of technology to mediate the process (Talaván, 2006).

To plan an intervention based on didactic subtitling, the instructor should carefully consider the purpose of viewing certain excerpts, the interest of students in viewing and subtitling the videos (henceforth, the audiovisual corpus) and the length of the intervention. Each lesson plan should contain a 1 min excerpt that should be subtitled. Choosing the videos that comprise the audiovisual corpus is a challenge as they should be coherent with the proficiency level of the participants and with the degree of familiarity with AVT tasks (Lertola, 2019). Besides, choosing a specific AVT mode depends on the purpose of the activity. Revoicing modes enhance oral production skills, whilst captioning tends to reinforce listening and writing skills.

If the audiovisual corpus is not appropriate because it does not meet the input+1 criteria (Krashen, 1982), then it should be replaced by suitable material as it may cause frustration and will probably have negative effects on motivation. Once the appropriateness of the level of the video has been ensured, the subtitling stage takes place. During this stage, three events occur simultaneously because learners are drafting their subtitles, watching the video, and paying attention to the product since they can see their own subtitles on the screen.

Cognitively, it is a demanding task. Additionally, drafting involves the affective filter (Krashen, 1982; Talaván et al., 2024) as it is supposed to trigger motivation because of the immediateness of the results as well as engagement because of the very nature of the task. Anxiety levels are expected to decrease, provided the language requirements for the activity align with the language competence of the learner. In addition, according to the principles of Emotional Psychology, if an individual is exposed to the feared stimuli in an appropriate way, through the exposure, the levels of anxiety naturally tend to decrease (Abascal et al., 2010). Resilience is another emotional factor that should be considered, as learners should face a multiplicity of challenges linked to language, culture and technology. When it comes to the linguistic dimension, adhering to the spatial and time constraints is a challenge that activates their repertoire as they should try to preserve the meaning, bearing in mind the maximum number of characters and the time that the subtitle should appear on the screen (Bolaños-García-Escribano, 2025).

Following Krashen’s (1989) theory of input, output and intake, and applied to the concrete case of subtitling, learners receive input that should be adapted through translation to produce output. Metalinguistic and intercultural skills are necessary for the drafting phase as learners engage with a multimodal text in which the multiplicity of semiotic elements is interacting. By repeatedly engaging in this process, an intake is expected to occur. Then, after completing the task, learners will probably select language for their own purposes, which is known as the adaptation phase. Eventually, the adoption phase occurs when learners incorporate these elements into their repertoire and can produce them correctly. Technology, thus, plays a crucial role since it mediates the whole process. Having discussed the pedagogical fundaments of subtitling as a task, it is now necessary to explore its effectiveness by examining the state of the art.

1.2. Didactic Subtitling: State of the Art

Scholars like Díaz-Cintas (1995, 1997) predicted the potential of subtitling as a language teaching technique through the design of several proposals of a pedagogical nature. It was after the turn of the millennium, when the democratization of PCs took place in Western societies, that projects and experiments on didactic subtitling emerged. This is the very case of Learning via Subtitling (LeViS), a European project led by Sokoli (2006), which can be considered the first project in the field of DAT Studies. During the first decade of the new millennium, PCs allowed users to subtitle, but they could barely be used for revoicing modes, as they required a great number of computational resources. That is why there was a growing interest in research on subtitling.

Thanks to the advancements in technology and institutional funding, a more feature-rich platform was developed, enabling captioning tasks and revoicing tasks, launched as part of the ClipFlair project (Sokoli, 2015). McLoughlin and Lertola (2015) designed a proposal for teaching Italian as a foreign language using ClipFlair, which combined captioning and revoicing. Another experiment conducted by Soler Pardo (2020) proved that English learners using ClipFlair experienced positive results in terms of language gains by triangulating the empirical data with the observations. The author added that subtitling also enhanced ICT skills and triggered motivation. Even though it may seem that AVT is linked to translating from an L2 into an L1 or vice versa, Talaván et al. (2016) designed an experiment with a sample of 41 students who performed an intralingual subtitling of 10 excerpts from a sitcom. The results confirmed the effectiveness of intralingual subtitling for the enhancement of written production skills and vocabulary. Another experiment by Talaván and Rodríguez-Arancón (2024b) addressing reverse subtitling (L1→L2) with a sample of 24 participants evidenced the way in which this technique fostered translation skills, written production, vocabulary and grammar. The study by Talaván et al. (2017) also explored reverse subtitling, but it added a factor of uttermost importance in virtual environments: collaboration. The results are consistent with the former ones, with written production being one of the most developed skills, whilst there seems not to be a significant improvement in terms of grammar.

Besides general English, there were projects that specifically addressed ESP (English for Specific Purposes). Ávila-Cabrera and Corral Esteban (2021), within the framework of the SubESPSkills project, found similar results to those analyzed, with written production being the skill that showed the greatest development. In addition, González-Vera (2021), also in the field of ESP, conducted an interesting experiment that filled a gap in the literature as listening had not been studied in such depth until she found out that learners who engaged in subtitling tasks showed better oral reception skills than those following a traditional language course.

In contrast, Ogea-Pozo (2020a, 2020b) approached the same task from the perspective of the translator trainer and found that this approach also fostered translation skills in a significant way. As Larrea-Espinar and Raigón-Rodríguez (2019a) affirm, culture and audiovisual texts are interwoven, and the intercultural axe is essential when it comes to language learning. In this regard, PluriTAV (Martínez-Sierra, 2018) was a project that focused on plurilingualism and interculturality, developing a series of didactic sequences that addressed the formerly mentioned aspects (Reverter Oliver et al., 2021).

The main drawback of the formerly mentioned studies consisted of the sample size, and that is why the TRADILEX project offered new insights. One of the peculiarities of this project was that it did not focus on a specific AVT mode, as it aimed at measuring the effectiveness of a sequence that comprised the five main AVT modes (subtitling, voice-over, dubbing, AD and SDH). The language tests were validated and designed specifically for the project by Couto-Cantero et al. (2021), and a pilot study was conducted in order to explore the validity of the design (Talaván & Rodríguez-Arancón, 2024a). Fernández-Costales et al. (2023) empirically proved that the proposed sequences were effective for B1 and B2 and that there were statistically significant differences between the pre-test and post-test with a sample of 566 participants who completed the 15 lesson plans in the intervention.

Linked to the TRADILEX project, different experiments were conducted to explore the validity of the methodology. Veroz-Gónzalez (2024) was a pioneer in exploring the effectiveness of subtitling in French through an experiment that reported significant language gains, particularly in production skills, and examined the intercultural dimension, which showed improvement. In the same vein, Rodríguez-Arancón (2023) explored the way in which DAT-based tasks fostered intercultural awareness. Similarly, Rodríguez-Arancón et al. (2024) affirmed that DAT enhances intercultural learning through CALL. These two studies consisted of a questionnaire that was implemented after the intervention of the TRADILEX sequences. Nonetheless, an analysis of the literature reveals two gaps: one related to proficiency and the other to interculturality.

There seem to be no studies addressing these specific issues in a design with both pre- and post-tests. Given the importance of intercultural awareness in education (Arnaiz-Castro et al., 2022; Gómez-Parra, 2018; Gómez-Parra et al., 2021), it is necessary to delve into this topic. These two questions are of great importance from the point of view of language teaching, as they enable the exploration of the extent to which subtitling can enhance both proficiency and intercultural awareness. Another gap that was found in the literature consisted of the lack of studies approaching the longitudinal effects of the intervention, as attrition is a variable that should be approached after instruction, according to Tracy-Ventura et al. (2024). Therefore, the study addresses the following questions:

RQ1. To what extent does the intervention have effects on language gains?

RQ2. How does the intervention influence the development of intercultural competence?

This study aims to fill a knowledge gap in the field of DAT studies because the literature on the long-term and medium-term effects of DAT-based interventions is still scarce, and it could be considered pioneer research because, to date, there are no longitudinal studies on the effects of DAT-based interventions in the mid-term on interculturality and proficiency.

2. Materials and Methods

The nature of this study is quantitative, and it consists of a pre-experimental design with a pre-test, a post-test, and a delayed post-test aimed at exploring the impact of a subtitling-based intervention and its effects on both language skills and the development of intercultural competence in the medium term.

2.1. Sample and Important Considerations on the Implementation of the Intervention

The sample consisted of 43 participants, 11 of whom identified themselves as male, and 32 identified themselves as female. They were university students taking a B2 English course at the University of Córdoba (Spain). The average age was 19, with only a few exceptions (min = 18, max = 22). They voluntarily agreed to take part in this experience. The experiment was developed between mid-February 2024 and April 2024, and it took 6 weeks. The participants were asked to complete the tasks during practical sessions of the course. Due to ethical considerations and the exploratory nature of the project, all participants were offered the opportunity to participate in the experiment. In that way, equality of opportunities was ensured.

Methodologically, it is important to acknowledge that the lack of a control group affects the generalizability of the results. However, this decision was justified to preserve the fairness of the educational setting and is coherent with the nature of the project. Another limitation that should be mentioned is the fact that participants were taking a B2 English course—the intervention was, in fact, part of that course—and it could be hypothesized that it was the syllabus instead of the intervention that made the effect. Nonetheless, the delayed post-test was crucial to address this issue as it helped mitigate this concern because it provided insight into what happened during the course when there was no intervention.

It is important to mention that the intervention took place during the practical sessions—which shifted from subtitling-based lesson plans during the intervention to traditional textbook-based practices afterwards—whilst the theoretical lessons remained consistent throughout the course. In that way, the design, despite its limitations, aims at isolating the impact of the subtitling tasks on proficiency and intercultural awareness over time.

2.2. Instruments

To measure their language gains, a commercial multilevel test by Macmillan (London, United Kingdom), designed by Day and Skerrit (2020) and included in their method, English Hub, was used, but it has not been reproduced in this article due to copyright reasons. This instrument consisted of a test with 70 graded multiple-choice questions that assessed vocabulary, grammar and language use. The questions gradually increase in terms of difficulty, offering a snapshot of participants’ proficiency. Test–retest reliability was assessed by correlating the scores of the pre-test with those of the post-test. The results showed a strong and significant Pearson correlation (r = 0.817; p-value < 0.001), which indicated stability over time. It is important to clarify that this instrument measures the use of English, and it does not focus on reception or production skills. Therefore, language gains in this context should be interpreted in terms of vocabulary, grammar and language use.

Regarding interculturality, the ERI scale by González Plasencia (2020) was used to assess the intercultural skills of participants. In the same vein, it has not been reproduced for copyright reasons, but it is available in the cited study. This scale consisted of a five-point Likert scale with 97 items that approached the three dimensions of competence: cognitive, attitudinal and affective. This scale was validated by the author with a large sample, and the sample of this study reported strong reliability with a Cronbach α of 0.941 (95% CI: 0.911–0.963).

2.3. Materials

The cognitive–affective model was essential to preparing the materials, as they were expected to be motivating and aligned with participants’ interests and language skills. The topic was literature and gender, and it was linked to the English for Social Purposes and Cooperation approach by Huertas-Abril and Gómez-Parra (2018), which states that social topics should be part of the language classroom, as they accomplish a double aim: teaching language through topics that are close to the realities of students and including social values as an essential element of the syllabus.

The lesson plans of the intervention were designed according to the model by Talaván and Lertola (2022). The excerpts that participants had to subtitle belonged to films about women writers from the anglophone worlds, which involved contact with the L2 culture and, therefore, were expected to improve their intercultural competence. Each lesson plan contained a warm-up, a video viewing, a subtitling task and a production task. Participants were encouraged to use AEGISUB because it is a free, user-friendly editor. Each lesson plan was built upon the figure of a woman writer:

• Mary Shelley (https://www.tradilex.es/secuencias/27/lessons/75, accessed on 2 September 2024);
• Emily Dickinson (https://www.tradilex.es/secuencias/27/lessons/76, accessed on 2 September 2024);
• Jane Austen (https://www.tradilex.es/secuencias/27/lessons/77, accessed on 2 September 2024);
• Virginia Woolf (https://www.tradilex.es/secuencias/27/lessons/78, accessed on 2 September 2024).

Through this sequence, participants were not only expected to have contact with the language but also with the culture of the L2 in a deeper way, focusing on the invisible culture (values, norms, behaviors, etc.). It is important to bear in mind that audiovisual texts portray the core values of a society, as Larrea-Espinar and Raigón-Rodríguez (2022) affirmed.

2.4. Schedule

The experiment was developed between mid-February 2024 and mid-April 2024, and it took 6 weeks according to the following schedule:

• Week 1: Participants were asked to complete the language skills test, which was administered in a paper-based format in a controlled environment; they also were asked to fulfill the ERI scale.
• Week 2: The participants completed a lesson plan on Mary Shelley. The objectives included developing audiovisual mediation and reception, improving written production, deepening lexical and grammatical aspects such as conditional sentences and reported speech and fostering sensitivity towards motherhood from a feminist perspective. The structure consisted of a warm-up with five exercises, a critical viewing of a film scene followed by a short-written reflection, creating their own English subtitles and a consolidation task requiring a 250–300-word essay imagining the life of a hypothetical sister of Shakespeare.
• Week 3: The participants engaged in a lesson plan centered on Emily Dickinson, aimed at enhancing audiovisual mediation and reception skills. The lesson plan also sought to improve written expression, consolidate grammatical elements like conditional sentences and foster an understanding of the underrepresentation of certain groups in the literary canon, including non-heteronormative identities. The session began with a warm-up featuring reading comprehension, grammar practice and mediation tasks. Students then critically analyzed the trailer for the Dickinson series before creating their own English subtitles for a selected scene. The lesson concluded with a reflective writing task, drawing from biographical notes, poems, and the trailer to explore Dickinson’s work from a feminist perspective, with particular emphasis on the role of sexuality in her poetry.
• Week 4: The participants worked on a lesson plan centered on Jane Austen. The objectives were to enhance audiovisual mediation and reception, improve written production and strengthen specific grammatical structures such as inversions and modal verbs. Additionally, the lesson encouraged a deeper understanding of human relationships through a feminist and class-conscious perspective while exploring Austen’s work from a feminist approach. The session began with a warm-up that provided historical and biographical context, followed by multiple-choice questions on key facts and a grammar-focused task. During the critical viewing phase, students took notes while watching the Pride and Prejudice trailer and then reflected on the topic of marriage as a trade. For the subtitling task, they were asked to create subtitles from scratch for a well-known scene from the film adaptation. The lesson concluded with a consolidation activity, where students wrote an essay analyzing a quote from Pride and Prejudice in relation to the topic of marriage as a trade.
• Week 5: The participants engaged in a lesson plan focused on Virginia Woolf. The objectives were to develop audiovisual mediation and reception skills, enhance written production and reinforce grammatical aspects such as conditional inversions, question tags and the use of infinitives and gerunds. Additionally, the lesson encouraged reflection on the importance of mental health and explored Woolf’s work from a feminist perspective, with particular attention to mental health issues. The session began with a warm-up that included reading a passage from Mrs. Dalloway, followed by comprehension questions and grammar exercises contextualized within the text. In the critical viewing phase, students watched the trailer for The Hours, a film in which Virginia Woolf is portrayed, and were asked to write brief predictions about the film’s content, focusing on mental health. For the subtitling task, they created subtitles for a scene in which Woolf discusses her mental health with her husband. The lesson concluded with a consolidation task, in which students were asked to write an essay on the role of cinema in raising awareness of mental health.
• Week 6: An additional week was granted to allow participants who had not yet finished the lesson plans to complete them.
• Week 7: During the final week, they repeated the assessments from Week 1, which included the language skills test, administered in a paper-based format under controlled conditions, as well as completing the ERI scale.
• June: two months after the end of the experiment, students were provided with the same language skill test and the ERI scale in order to gather the data for the post-delayed test. From April to June, students had the same regular theory lessons together with practical sessions based on traditional textbook practices.

Figure 2 provides a general perspective of the experimental design, whose aim is to have a preliminary approach to the effects of the inclusion of intralingual subtitles into a regular English course focusing on interculturality and proficiency. Even though the limitations of the design were already acknowledged, it is important to consider that regular theory lessons remain constant throughout the study. It ensures that any observed effects are probably attributable to the transition from innovative subtitling-based practices to traditional textbook-based tasks in the practical sessions. Including a delayed post-test aims at addressing the absence of a control group. At the same time, it provides additional insights into language gains and intercultural awareness attrition after the intervention.

3. Results

This section was divided into two sections that addressed the RQs below:

RQ1. To what extent does the intervention have effects on language gains?

RQ2. How does the intervention influence the development of intercultural competence?

The first section is devoted to the study of language gains (grammar, vocabulary and use of English) and the second section to the development of intercultural competence.

3.1. Language Gains

One of the most important aspects of this research is exploring the effectiveness of the intervention in terms of language gains in the mid-term (RQ1). Previous research on second language attrition (Tracy-Ventura et al., 2024) suggests that proficiency may either decline or improve over time, depending on the level of exposure to the L2 after the intervention. Peak attainment is expected to appear according to former studies on DAT-based interventions in which significant differences were observed between the pre-test and the post-test. In order to explore the results, a repeated measures ANOVA test was conducted with JASP 0.19.0.2.

Table 1. Descriptive statistics for language gains.

Language Skills	N	Mean	SD
Pre-test	32	50.563	8.496
Post-test	32	53.656	8.635
Delayed post-test	32	54.781	8.932

shows the descriptive data retrieved from the language skills test in the three different moments in which measures were taken. Even though there were initially 43 participants, only 32 completed the language skills test in the three different moments, as they could leave the experiment whenever they wanted. The expected peak of attainment is observed in the post-test measure (M = 53.656), with a growing tendency reflected in the mean of the delayed post-test (M = 54.781). This slight increase suggests that continued exposure to the L2 after the intervention may have contributed to this result. This exposure was probably linked to the theory lessons and the traditional practical sessions.

The results of the within-subjects ANOVA test for the variable under study are shown in

Table 2. Within-subjects ANOVA test for language gains.

Cases	Sum of Squares	df	Mean Square	F	p	η2
Language gains	305.438	2	152.719	7.148	0.002	0.187
Residuals	1324.562	62	21.634	-	-	-

. From these results, it can be inferred that the intervention had a significant effect on language gains, with an F-value of 7.148 and a p-value of 0.002. In other words, the level of language skills varied significantly with time. It is important to focus on the mean square for the effect of the intervention on language gains (152.719) because it was greater than the unexplained variability linked to the residuals (21.634). The effect size reported by an η² of 0.187 showed that 18.7% of the variance in language gains could be explained by the differences among the measurement time points, meaning that it was linked to the effects of the intervention.

Figure 3 consists of a raincloud diagram, showing participants’ marks in language gains across time. Language gains varied significantly over time, with a moderate effect. The level does not decay after the instructional attainment peak; it rises at a slower pace, which is linked to the period in which traditional practical sessions were delivered.

Calculating the percentage growth could be useful in this regard. From measure 1 (February) to measure 2 (April), the percentage was 6.12%, but taking into consideration that this period spans 2 months, the monthly growth rate was 3.06%. When it comes to measures 2 (April) and 3 (June), the percentage growth rate was 2.10% and bearing that there is a span of 2 months, the monthly growth rate after the intervention was 1.05%. Therefore, during the intervention, language gains increased 2.91 times faster.

Post-hoc analyses were conducted to explore which means differ significantly from each other, and these results are shown in

Table 3. Post-hoc comparisons for language skills.

Pre-Test	Post-Test	Mean Difference	SE	t	Cohen’s d	pbonf
Pre-test	Post-test	3.094	1.001	3.090	0.356	0.013
	Delayed-post-test	4.219	1.320	3.197	0.485	0.010
Post-test	Delayed-post-test	1.125	1.123	1.002	0.129	0.973

. Comparing the pre-test and post-test results, we obtained a medium effect size (Cohen’s d = 0.356) and significant differences (p_bonf = 0.013). Regarding the pre-test and the delayed post-test, the effect size (Cohen’s d = 0.485) was moderate, and the differences between the two measures were statistically significant (p_bonf = 0.010). As can be inferred from the descriptive data and the growth rate, the differences between the delayed post-test and the post-test were not statistically significant (p_bonf = 0.973), and the effect size was small (Cohen’s d = 0.129). It is important to note that the differences between the pre-test and the post-test were significant, which coincided with the subtitling-based practical sessions. Whereas when assessing the differences between the post-test and the delayed post-test, which was the span of time devoted to traditional practical sessions, the effects were low, and no significant differences were reported.

3.2. Interculturality

A parallel analysis of the development of intercultural competence was conducted by implementing the ERI scale. As mentioned in the theoretical framework, studies on DAT and interculturality are still scarce, and they were based on self-reported questionnaires implemented after an intervention. In this case, the study differs from previous research in that it uses a pre-test and post-test pre-experimental design. Instead of employing a custom-designed instrument for DAT studies, this study used an existing tool that measures participants’ intercultural resources. This approach provides a different perspective on the effectiveness of DAT interventions.

In

Table 4. Descriptive statistics for interculturality.

Intercultural Competence	N	Mean	SD
Pre-test	30	325.533	36.066
Post-test	30	347.700	45.845
Delayed post-test	30	345.700	59.699

, the reader may find the means and the SD of the ERI scale at different moments of measurement. In this case, only 31 participants out of the 43 completed the ERI scale in the three different moments because they were free to withdraw from the project at any moment, and they could decide whether to submit the data or not. These data suggest that intercultural competence reached the peak of attainment but then gradually decreased. It might suggest that audiovisual materials provided participants with rich cultural materials and that a peak of attainment occurred. However, the post-test revealed intercultural competence attrition after the end of the intervention, even though participants went on with the course-taking of their theory lessons and traditional lessons. It also suggests that continuous exposure to the L2 in culturally relevant contexts could be important for sustaining and further developing intercultural competence.

Table 5. Within-subject ANOVA test for intercultural competence.

Cases	Sum of Squares	df	Mean Square	F	p	η2
Intercultural competence	9020.556	2	4510.278	3.084	0.056	0.096
Residuals	84810.778	58	1462.255

shows the results of the within-subjects ANOVA test in order to evaluate the intervention in the three main measurements. It is worth noting that the analysis showed a marginally significant effect with an F-value of 3.084 and a p-value of 0.056. It can thus be inferred that intercultural competence varied significantly over time. Furthermore, the mean square for this variable was 4510.278, whilst the mean square for these residuals was 1462.255. It suggests that the variability due to the changes in intercultural competence was greater because of the intervention than due to other unexplained factors. The effect size (η² = 0.096) indicates that 9.6% of the variance in intercultural competence was linked to the intervention.

As shown in Figure 4, the development of intercultural competence increased during the intervention, and the data also suggest a decrease after the intervention, with the measure taken in the post-test being the attainment peak. It is important to note that attrition seems to take place after the intervention.

The post-hoc results in

Table 6. Post-hoc comparisons for intercultural competence.

Pre-Test	Post-Test	Mean Difference	SE	t	Cohen’s d	pbonf
Pre-test	Post-test	22.167	8.841	2.507	0.462	0.054
	Delayed-post-test	20.167	9.919	29	0.421	0.154
Post-test	Delayed-post-test	2.000	10.765	0.186	0.042	1.000

show that the largest effect size (Cohen’s d = 0.462) occurred between the pre-test and the post-test, with this difference being marginally significant according to the adjusted p-value (p_bonf = 0.054). When it comes to the comparison between the pre-test and the delayed post-test, the difference was not statistically significant (p_bonf = 0.154) and had a moderate size effect (Cohen’s d = 0.421). Additionally, the size effect (Cohen’s d = 0.042) was small, and the differences were not statistically significant (p_bonf = 1.000) for the case of the post-test and the delayed post-test.

4. Discussion and Conclusions

This study aimed to explore aspects of CALL that have received less attention. Subtitling and subtitles have been largely studied, but there was a gap in the literature, as no studies in the mid-term have been implemented to the author’s knowledge. Despite Talaván et al. (2016) reporting positive results in terms of vocabulary and writing enhancement in a pre-test and post-test design, this study measured the effects on proficiency through a commercial test by Day and Skerrit (2020) at three different moments. Interculturality was addressed by Veroz-Gónzalez (2024), Rodríguez-Arancón et al. (2024) and Rodríguez-Arancón (2023). Nonetheless, these experiments consisted of a post-test without a pre-test. In this study, a validated instrument by González Plasencia (2020) has been used to measure this competence in three different moments.

As was formerly acknowledged, there is room for improvement in this study in terms of methodology since there was not a control group that was already justified. It is also important to explain that during the intervention, there were theoretical sessions on the English B2 course. Following the intervention, traditional textbook-based lessons were delivered instead of subtitling-based lesson plans. During the period of time devoted to the experiment and the period of time in which they attended traditional practical sessions, they engaged in theoretical lessons. Therefore, it was a constant over the experiment. Measuring outcomes after these traditional lessons provides greater control over the effects of the initial intervention.

The results suggest that there was a peak of attainment in language gains and in interculturality immediately after the intervention. However, there are differences between the behavior of these two variables. When it comes to proficiency, results showed that there was a peak of attainment after the subtitling intervention, with significant effects. After the intervention, when participants followed traditional practical sessions, there was a light but nonsignificant increase in their proficiency. Interculturality, however, different behavior is reported, as there is a peak of attainment after the intervention followed by quick attrition during the period of time devoted to traditional practical sessions.

Larrea-Espinar and Raigón-Rodríguez (2019b) explored the way in which textbooks approached culture in a diachronic study, and their results seem to align with the findings because there was a decrease in intercultural awareness after the intervention during the period in which they were still attending language lessons (theory and practice) using traditional textbooks. The authors argued that textbooks tend to focus on “big C” culture (the major, formal aspects of culture, such as history, art and literature), but they do not give sufficient attention to “little c” culture (the everyday, lived experiences and practices of people).

The pedagogical implications that derive from the results are that textbooks may provide a solid foundation for language proficiency, but the intercultural axe requires additional resources, as exposure to other cultures, specifically to the L2 culture, is essential for a better understanding and command of the language. Proficiency seems to improve faster when learners are exposed to authentic materials with which they can interact. In this regard, Didactic Audiovisual Translation facilitates this interaction through the consumption of authentic materials and the production of materials in the L2. The role of students as prosumers should be considered, particularly in the current context. Additionally, intralingual translation has proven to be effective for learners in advanced English courses.

CALL, through the technique of subtitling, seems a promising tool to foster intercultural awareness and language learning. Therefore, including a DAT-based lesson plan in regular language courses could be an interesting option to enrich the syllabus. This type of task could be included every fortnight to enhance motivation and provide learners with exposure to the language and culture. Platforms like TRADILEX are free and allow teachers to have a specific profile to follow their learners’ progress. This platform includes more than 60 lesson plans for B1, B2 and C1 with a guide for teachers, so it is relatively easy to include this technique in any syllabus. Replication studies in different contexts are encouraged, and the cognitive–affective model should be taken into account when deciding to implement this experiment in another context, as the selecting factors are context-dependent or when designing the materials for a different experiment. Future research should also address the affective filter by delving into engagement, anxiety and other psychological and pedagogical variables related to learning processes.