The Effects of Self-Access Web-Based Pragmatic Instruction and L2 Proficiency on EFL Students’ Email Request Production and Confidence

Abstract

The present study pursued three objectives: (i) to examine whether self-access web-based instruction could significantly improve EFL students’ ability to formulate pragmatically appropriate email requests; (ii) to determine whether L2 proficiency influenced students’ pragmatic performance and their gains following instruction; and (iii) to explore changes in learners’ confidence when evaluating the appropriateness of their own email requests. Sixty-eight first-year English Studies students at a Spanish university completed a five-week intervention integrated into their curriculum. Their L2 proficiency was assessed using the Oxford Placement Test, which categorized them into B1 (n = 22), B2 (n = 23), and C1 (n = 23) levels. Using a pre–post-test design, learners’ performance was assessed through email tasks varying in imposition, and their confidence was measured via Likert-scale ratings. Results showed statistically significant improvements across all dimensions of an analytic rubric—particularly in request appropriateness and organization—indicating that self-access instruction effectively enhanced learners’ pragmatic competence. Gains were similar across the three proficiency groups, with B2 students showing slightly higher though not statistically significant improvements. Participants also reported significantly increased confidence in evaluating their own email appropriateness post-intervention. Findings support the integration of self-access pragmatic resources into EFL curricula to develop academic communication skills in higher education contexts.

1. Introduction

In academic contexts, email has become a primary means of communication between students and university faculty. Among the range of purposes academic emails serve, requests are both frequent and pragmatically demanding. Writing appropriate email requests in a second language (L2) requires the ability to balance politeness, clarity, and contextual sensitivity. This is particularly challenging for learners of English as a Foreign Language (EFL), who often lack exposure to authentic models and opportunities to receive instruction on how to manage social distance and power relations appropriately in written communication (Biesenbach-Lucas, 2007; Economidou-Kogetsidis, 2011; Halenko & Winder, 2021). Previous research has consistently shown that pragmatic competence and, in particular, the ability to produce appropriately mitigated requests, do not develop automatically through increased exposure or language use alone. Longitudinal studies involving authentic email communication have revealed that frequent practice, in the absence of explicit instruction, does not lead to measurable gains in students’ pragmatic performance (Halenko & Winder, 2021; Rau & Rau, 2016). These findings underscore the need for instructional interventions that explicitly target the pragmatic dimensions of L2 email writing.

A growing body of work has examined the effects of explicit instruction on learners’ ability to formulate pragmatically appropriate email requests in both English as a Second Language (ESL) and EFL settings (e.g., Alcón-Soler, 2015a, 2015b; Y. Chen, 2015; Nguyen, 2018; Usó-Juan, 2021, 2022). While this research has generally reported positive effects, interventions have been delivered through teacher-led instruction within the classroom. However, such approaches may not always be feasible in university contexts, where time constraints and curricular demands limit the integration of pragmatics-focused instruction.

To address this challenge, recent pragmatics research has begun exploring the potential of digital modes of instruction (see González-Lloret, 2022). In the present study, we focus on self-access web-based learning, since it is a flexible and scalable alternative for fostering pragmatic development. This mode of instruction allows learners to engage with input and tasks at their own pace and has shown promising results in the field of L2 pragmatics (e.g., Kerber et al., 2023; Muhammad, 2022). Findings also suggest that learners’ L2 proficiency may play a critical role in mediating the effectiveness of self-access instruction, with higher-proficiency students better benefiting from this type of instruction, particularly with respect to pragmatic production (e.g., Yang, 2017).

Despite growing interest in digital instruction, few studies have examined the effects of self-access web-based materials on EFL learners’ ability to produce pragmatically appropriate email requests in contexts involving high social distance—such as initial requests to unknown faculty members. Moreover, little is known about how such instruction affects learners at different proficiency levels. The present study addresses these gaps by investigating the effects of a pragmatics-focused self-access web-based instructional module on Spanish university students’ ability to write email requests to faculty, and by exploring the mediating effect of L2 proficiency.

Finally, the present study also explores the impact of instruction and learners’ L2 proficiency level on their confidence when assessing the appropriateness of their requestive emails. This third objective was set since, despite the positive role that confidence can play in language learning (MacIntyre et al., 1998), limited research has examined how explicit instruction may influence learners’ confidence in their pragmatic abilities (e.g., Martínez-Flor, 2006), particularly in relation to complex communicative tasks such as writing email requests (Y. Chen, 2015; Usó-Juan, 2022).

2. Literature Review

2.1. Email Requests to Faculty

Research on email communication in academic settings has shown that students use emails to achieve various communicative goals, with requesting being the most frequent function (e.g., Biesenbach-Lucas, 2007; Y. Chen, 2015; Economidou-Kogetsidis, 2011). Specifically, requests for information (e.g., “Which is the date of the exam?”) and requests for action (e.g., “Could you please send me the feedback on my project?”) have been documented to be the most common types (Codina-Espurz & Salazar-Campillo, 2019; Economidou-Kogetsidis, 2011).

Requests typically consist of a head act—where the request is made—and peripheral modifiers that either soften (mitigators) or intensify (aggravators) the imposition (Codina-Espurz, 2022). Internal modifiers include both lexical (e.g., polite markers like ‘please’, mitigators like ‘a bit’, or downtoners such as ‘if possible’) and syntactic strategies (e.g., use of past tense, conditionals, subjunctives, embedded clauses). External modifiers, or supportive moves, function as preparatory or follow-up elements to the head act and may include grounders (reasons for the request), preparators (statements that precede the request), precommitments (attempts to gain the recipient’s agreement beforehand), and offers of reward. In emails, the request head act together with its external modifiers are usually referred to as the “content moves”, while the subject line and the opening and closing lines make up the “framing moves”. Very importantly, both types of moves play a key role in managing the interpersonal aspects of the request.

Email writing poses significant challenges for L2 learners due to the need to convey intended meaning in a language they may not fully master. As Halenko et al. (2021) note, learners often face difficulties in selecting appropriate linguistic resources and in adapting these to specific contextual demands. The private nature of emails further complicates learning, as students have limited access to appropriate models of this genre (Halenko & Winder, 2021). In addition, research indicates that learners often prioritize their own needs in email requests while using insufficient mitigation strategies, which may unintentionally result in impoliteness (e.g., Biesenbach-Lucas, 2007; Codina-Espurz & Salazar-Campillo, 2019; Economidou-Kogetsidis, 2011; Hendriks, 2010). Furthermore, longitudinal studies of real email interaction have shown that frequent email writing alone is insufficient for developing pragmatic competence. For example, Rau and Rau (2016) observed that Taiwanese EFL graduate students, despite writing emails regularly over 12 weeks, did not adjust their language use to reflect growing rapport with their English-speaking instructor. Similar patterns were found by Halenko and Winder (2021), who studied Chinese ESL students during a 10-month academic stay in the UK and found minimal gains in email pragmatics despite increased exposure and writing practice. This has led to calls for explicit instruction in pragmatic strategies, especially the use of modifiers to soften requests (e.g., Usó-Juan, 2021).

2.2. Effects of Pragmatic Instruction on the Production of Email Requests to Faculty in English

Only a limited number of studies have investigated the effects of instructional interventions on the production of email requests in ESL (Alcón-Soler, 2015a, 2015b; Ford, 2006) and EFL contexts (Y. Chen, 2015; Nguyen, 2018; Nguyen et al., 2015, 2017; Usó-Juan, 2021, 2022). Overall, this body of research suggests that explicit instruction has a generally positive effect on students’ ability to formulate appropriate email requests, both in elicited and real-life conditions.

Focusing on an ESL setting, Ford (2006) was one of the first researchers to examine the impact of instruction on email requests. In his study, study-abroad students in the US received a single 50 minute lesson on email etiquette. Subsequent analysis of their discourse completion tests (DCT) revealed improvements in the structural organization and variety of content moves used in their requests. However, these improvements were not sustained in the delayed posttest, suggesting that a longer instructional period might have been necessary for lasting gains. Also exploring development in an ESL context, Alcón-Soler (2015a, 2015b) investigated the effect of explicit instruction on Spanish university students’ use of request mitigation strategies during a 36-week study abroad program in England. While all participants were exposed to authentic requests in context, only the group receiving four 20 minute lessons on pragmatics outperformed the control group, demonstrating the added value of explicit instruction.

While the previous studies focused on ESL contexts, the remaining body of research has focused on pragmatic instruction directed to EFL learners, a context that is particularly relevant for the present study. Y. Chen (2015) examined the impact of five hours of pragmatic instruction which followed a genre-based approach on Chinese learners’ email request performance. The analysis of 224 emails revealed that instruction significantly improved learners’ use of framing moves, likely due to their formulaic nature. However, content moves remained challenging, with students continuing to struggle with appropriate request strategies and the inclusion of supportive information. In addition, Y. Chen’s (2015) data showed that email requests that involved a lower distance relationship with the lecturer benefited more from instruction than higher-distance ones, as indicated by the greater gains observed in the first type of request scenario.

Working in the Vietnamese context, Nguyen (2018) also observed that instruction resulted in better use of request strategies. However, students made only modest progress in producing supporting moves. Notably, the positive effects were retained in a delayed posttest conducted eight months after the treatment, which consisted of six hours of metapragmatic instruction delivered over four weeks. Further studies by Nguyen et al. (2015, 2017) shed additional light on the role of feedback in combination with instruction. In Nguyen et al. (2015) two experimental groups received the same explicit instruction but were given different types of feedback: direct feedback or metapragmatic feedback. While both groups outperformed a control group in a production task, the group receiving metapragmatic feedback achieved slightly higher scores, suggesting the value of raising learners’ metalinguistic awareness. In a follow-up study, Nguyen et al. (2017) examined whether the inclusion of required revisions following feedback would affect students’ use of syntactic downgraders. Three experimental groups received feedback after metapragmatic instruction, with variation in the number of required revisions (zero, one, or two). While all experimental groups significantly outperformed the control group (which received instruction but no feedback), no significant differences were found among the three experimental groups, indicating that the act of revising may not be the key variable once instruction and feedback are in place.

The most recent instructional studies have been conducted in Spain by Usó-Juan (2021, 2022). In her 2021 study, Usó-Juan examined both short- and long-term effects of metapragmatic instruction on students’ use of request mitigation strategies. The instructional treatment consisted of three two-hour sessions and targeted scenarios where students were familiar with the professor and had regularly attended class, thus involving low social distance. Analyses showed a significant increase in the frequency and range of mitigation strategies used in both immediate and delayed posttests. Participants initially relied heavily on grounders and conditional forms; however, after instruction, their use of a broader variety of mitigators increased. Finally, in her 2022 study, Usó-Juan uniquely analyzed naturally occurring emails rather than controlled tasks. Her data set comprised 110 real student emails sent to instructors, and the results showed that three two-hour sessions of instruction significantly improved the overall appropriateness of students’ email requests. The analysis also confirmed earlier findings by Y. Chen (2015) and Nguyen (2018) that framing moves were more amenable to instruction than content-related email components.

In addition to analyzing the effects of instruction on email production, Y. Chen (2015) and Usó-Juan (2021) explored an underexplored dimension of L2 pragmatic development: EFL learners’ confidence in their pragmatic competence. Specifically, both studies measured participants’ confidence through Likert-scale self-assessments of their perceived appropriateness of their email requests to faculty. Y. Chen (2015) found that instruction positively affected learners’ confidence, which significantly increased in the posttest significant. Furthermore, data showed that participants felt more confident when writing email requests that involved a lower distance relationship with the lecturer. Usó-Juan’s (2021) data showed, first, that learners’ confidence level was already quite high before instruction, and second, that it significantly increased immediately after instruction and remained higher than pre-instruction after two months. These findings are highly relevant when considered through the lens of positive psychology in language instruction, which argues that language teaching should not only promote learners’ proficiency, but also support their well-being, positive emotions and traits (e.g., MacIntyre et al., 2016). First, in a broad sense, feelings of confidence are intimately connected to the extent to which individuals perceive themselves as competent. Given that feeling competent is considered a fundamental human motivation (Keller, 2010), fostering learners’ confidence contributes to their general well-being. Second, from a situated perspective, state self-confidence (the feeling that one has the capacity to communicate effectively at a particular moment) has been shown to be positively related to dimensions such as willingness to communicate and L2 motivation (e.g., MacIntyre et al., 1998). Given these connections, it seems highly relevant to expand existing research on the effects of pragmatic instruction on EFL learners’ confidence in their pragmatic production.

Considering the overall findings from the abovementioned studies, it can be argued that research to date supports the effectiveness of metapragmatic instruction, particularly when it integrates instructional elements such as explicit and implicit awareness-raising, scaffolded practice, and meaningful feedback. However, the extent of improvement appears to be influenced by the pragmatic features targeted and by learner and contextual factors. For instance, instructional gains tend to be more evident in framing moves (e.g., greetings and closings) than in content moves (e.g., request strategies and supportive reasoning). In addition, it seems that the length of the instructional period positively influences the pragmatic gains students obtain, as six-hour treatments led to longer term gains than brief interventions such as the one implemented by Ford (2006). Nevertheless, the limited classroom time typically available in university-level English courses may constrain the implementation of such extended instruction. As a result, there is a growing interest in exploring alternative modes of delivery—such as web-based or self-access formats—that can support pragmatic development without demanding additional classroom time. Given that previous research has shown that students benefit from teacher-led classroom instruction that explicitly targets the different moves in request emails, the present project intends to contribute to this strand of research by shedding light on a type of instruction that has not been explored for teaching L2 email requests to EFL students: self-access web-based materials.

2.3. Self-Access Web-Based Pragmatic Instruction

In view of the time constraints of traditional classroom instruction, recent research has begun exploring the potential of alternative modes of instruction such as self-access websites for developing learners’ L2 pragmatic competence (e.g., Kerber et al., 2023; Muhammad, 2022; Yang, 2017, 2024). This mode of instruction offers learners the flexibility to engage with content at their own pace and can be integrated into formal coursework or used independently. Preliminary findings suggest that self-access instruction can be beneficial for the development of L2 pragmatic skills, though outcomes depend on task and learner-related factors, such as the amount of practice provided or students’ proficiency level. Muhammad (2022), for example, developed a digital module combining instructional videos and guided practice tasks aimed at improving the oral and written communication skills of 21 Indonesian students preparing to study in the US. The approximate completion time for each of the two sections (focusing on oral and written skills) was only three hours. While the study reported some improvements in participants’ pragmatic performance, the author attributed the limited gains to a lack of extended, sustained practice.

Considering learner-related factors, L2 proficiency has been found to mediate learners’ ability to notice and process pragmatic input in self-access web-based instruction. Kerber et al. (2023) explored the pragmatic development of 28 beginner-level Spanish learners as a result of their engagement with the web-based platform Dancing with Words, which focuses on Spanish speech acts and follows a strategy–instruction approach (Sykes & Cohen, 2008). Specifically, they assigned participants to one of two conditions: the self-access group engaged with the content independently, while the instructor-led group received instructional support and feedback, awareness and spoken production of Spanish apologies. The study assessed learners’ awareness and oral production of apologies before and after a five-week instructional period. Data showed that while the instructor-led group showed improvement in both awareness and production, web-based instruction did not significantly affect pragmatic development of the self-access group. From these findings Kerber et al. (2023) concluded that beginner learners would benefit from scaffolding and individualized feedback that complemented web-based instruction to make meaningful progress in pragmatic development.

Yang (2016, 2017, 2024, 2025) designed a website to aid the development of L2 Chinese expressions of gratitude that followed an explicit consciousness-raising approach. In her 2017 study, Yang examined the influence of L2 proficiency on the effectiveness of web-based pragmatic instruction. Thirty-six English-speaking university students studying Chinese as a foreign language were divided into two groups based on their scores on a Chinese Proficiency Test: a lower-level group with approximately one year of study, and a higher-level group with over two years of study. Over five weeks, both groups received the same instruction through the self-access website. Data analyses of DCTs showed that while all participants improved in the appropriateness and range of the thanking strategies used, higher-level learners achieved significantly greater gains than lower-level peers. Specifically, more advanced participants produced more varied and accurate expressions, made fewer errors, and demonstrated better adaptation to social distance. Based on these findings, Yang (2017) concluded that proficiency played a key role in learners’ capacity to process and internalize pragmatic information delivered through a self-access website.

In fact, Yang’s findings mirror general results found in recent meta-analyses that have explored L2 proficiency as a moderating factor of the effects of pragmatic instruction (Ren et al., 2022; Taguchi et al., 2022; Yousefi & Nassaji, 2019). These analyses have revealed that L2 proficiency level is a particularly strong predictor for the learning of productive pragmatic skills and suggest that intermediate groups seem to benefit more from instruction than novice and advanced learners. It is thus relevant to explore whether these results also apply to instruction that is delivered through a task-supported self-access web-based module such as the one employed in the present study.

3. The Present Study

Despite growing interest in digital instruction, little research has examined the effects of self-access web-based materials on the development of pragmatic competence in email requests to faculty, particularly within EFL settings. Moreover, most prior work has explored low-distance scenarios, where students are supposed to have an established relationship with the recipient, and thus more research is needed on how instruction affects students’ ability to produce appropriate email requests in scenarios involving higher social distance. Building on previous findings that highlight the efficacy of metapragmatic explicit instruction on production (e.g., Y. Chen, 2015; Nguyen, 2018; Usó-Juan, 2021, 2022) and the relevance of learner proficiency (e.g., Kerber et al., 2023; Taguchi et al., 2022; Yang, 2017), the present study investigates the impact of a self-access web-based instructional module on EFL students’ ability to write email requests to faculty members in first-contact scenarios. These situations are characterized by high levels of social distance, requiring learners to navigate delicate face-threatening acts while lacking prior rapport with the recipient. This context mirrors real-life academic communication in which students often need to make initial requests for assistance, clarification, or accommodation. In addition, this study investigates whether a self-access intervention—without individualized instruction or feedback—can enhance students’ confidence regarding the appropriateness of their email requests. Finally, it explores the effects of L2 proficiency on participants’ confidence, as learners’ linguistic resources are considered to affect their feelings of confidence (MacIntyre et al., 1998).

The study was guided by the following three research questions:

• RQ 1: Does self-access web-based pragmatic instruction work as a means to improve EFL learners’ ability to write appropriate email requests to faculty in English? If so, which aspects of email requests are affected by instruction?
• RQ 2: Do effects of self-access web-based pragmatic instruction vary according to learners’ L2 proficiency level? If so, which aspects of email requests are affected by proficiency?
• RQ 3: Do self-access web-based pragmatic instruction and L2 proficiency level have an effect on EFL learners’ confidence in the appropriateness of their email requests to faculty?

4. Methodology

4.1. Participants

Participants were 68 first-year English Studies majors, studying at a Spanish public university. They belonged to an intact group of 75 students, of which seven students did not participate in all the tasks of the teaching intervention. Sixteen students were male and fifty-two were female, reflecting the gender composition of their class. Their general English proficiency was measured using the Oxford Placement Test (Allan, 2004) and their scores reflected three levels of the Common European Framework: 22 students had a B1 level, 23 had a B2, and 23 had a C1 level. At the beginning of the semester, participants completed a background information questionnaire that showed that sixty of them had been taught how to write “formal emails” either during their secondary education or in extra-curricular lessons (such as in courses preparing them for official B2 level certifications). However, they reported that the content of such instruction had not focused on pragmatics, but on general email structure, formulaic language to be used in openings and closings and on avoidance of contractions and informal vocabulary. Furthermore, participants acknowledged that they were able to recall very little knowledge from that previous instruction. They could be described as novice users of academic email, since it was their first year at university and they expressed a general lack of real-life practice, noting that they had rarely written emails, whether in their first language (L1) or in English. Nevertheless, the majority were planning to spend their third year abroad and, as a result, they expected to use English to communicate with their lecturers in the near future. They were enrolled in a compulsory general English course that followed a coursework with no specific emphasis on pragmatics. The instructional treatment was included in the course as a project that students had to complete in addition to their coursebook work.

4.2. Self-Access Instruction

Pragmatic instruction was provided via a self-access web-based module created by the first author of the present study. The module adopted a task-supported, awareness-raising approach and was structured to incorporate elements believed to foster L2 (pragmatic) development. These included exposure to authentic language input, opportunities to produce output focused on communication, explicit instruction in metapragmatic knowledge, as well as opportunities for reflection through immediate metapragmatic and direct feedback after interactive activities and sample responses (e.g., Wain et al., 2019).1 In terms of structure, it was divided into four main sections. The first three sections focused on (i) key aspects of politeness, (ii) email structure and framing moves, and (iii) content moves. Each one included metapragmatic explanations, examples, reflection questions, and interactive activities that provided immediate feedback. Each section ended with a reflective questionnaire that invited students to reflect on their learning process through guided questions. The fourth main section provided practice tasks (detailed below).

As suggested by Fryer et al. (2014), the materials were integrated into the curriculum, creating a blended learning format “that ensures that students are engaged in studying the prescribed content for, at a minimum, a certain amount of time every week” (p. 27). As shown in Figure 1, participants performed in-class tasks in weeks 1 and 5, while engaged in autonomous out-of-class work from week 2 to week 4. During weeks 2 and 3 students could navigate the first three sections at their own pace and read the materials and complete the interactive activities as many times as needed. To help them organize their autonomous work, they were given suggested deadlines for working with each section and for submitting the corresponding reflective questionnaires. Access to Section 4 (Practice) was granted during the second half of week 3, and participants were given compulsory deadlines for submitting the tasks in this section. These tasks involved (i) a guided analysis of three email requests, (ii) the production of three emails, (iii) the guided comparison of those emails to three sample answers and, optionally, (iv) the creation of request scenarios followed by the writing of their corresponding emails.

4.3. Data Collection

The data were collected through two email writing tasks which included scenarios adapted from previous research (e.g., Y. Chen, 2015). The main characteristics are summarized in

Table 1. Email writing tasks characteristics.

Email	Distance	Imposition	Communicative Purpose
Email 1	+	−	Request for help accessing the class online learning platform.
Email 2	+	+	Request for permission to enroll in the lecturer’s class.

below:

Both situations involved first email encounters, and thus the social distance between the student and the lecturer would be considered high. The requests had different degrees of imposition: the first was a request for action that would be included within the lecturer’s duties (helping the student access the course online learning platform), while the second one would be more of a favor, as it involved giving the student permission to attend the lecturer’s course, which was already full. The scenarios were written in the students’ L1 for two main reasons: to ensure the participants’ understanding of the situation and to prevent them from copying verbatim from the instructions. The instructions and layout of the first email writing task are displayed in Figure 2 (an English version of both tasks can be found in Appendix A).

Following Y. Chen (2015) and Usó-Juan (2021), each scenario also included self-report measures of confidence, attention and difficulty. Students reported their confidence levels concerning the general appropriateness of the email written through a 5-point Likert scale ranging from 1 (not confident at all) to 5 (completely confident). Two open-ended questions were aimed at exploring the features they had paid special attention to and the difficulties they had encountered. While the analysis of the open-ended answers is beyond the scope of the present study, examples are presented to illustrate participants’ justifications of their confidence levels before and after instruction.

Data were collected in two writing sessions taking place in Weeks 1 and 5, during class time. Participants used their own electronic devices (i.e., laptops and tablets) and had 30 min to write the emails. A total of 272 emails were collected (68 participants × 2 emails × 2 times). The present study was approved by the first author’s institutions ethics committees. Student consent was requested at the end of the semester to minimize the potential feeling that students’ refusal to participate would negatively affect their grades. A passive consent approach was followed to avoid self-selection bias (Merrison et al., 2012).

Participants’ emails were analyzed by means of an analytic scoring rubric developed taking as a basis recommendations for rubric development (e.g., Vercellotti & McCormick, 2021) and rubrics employed in previous L2 (email) writing research (e.g., Y. Chen, 2015; Ishihara, 2010; Kuiken & Vedder, 2022; Nguyen et al., 2015). This type of rubric was chosen as opposed to a holistic one, because the former can provide detailed information on specific areas of improvement and capture differences such as the ones derived from varying levels of proficiency more sensitively. In addition, raters have been claimed to provide more consistent ratings when they rely on analytic criteria (e.g., East, 2009; Weigle, 2002).

Initial dimensions and descriptors were established based on commonalities found in previous rubrics. Twenty emails (belonging to both the pretest and the posttest, the two scenarios, and written by students at the three proficiency levels) were analyzed by three experienced raters (the authors of the present study) in order to refine this first version of the rubric, so that it could capture most phenomena across levels and tasks. An iterative process of collaborative refinement and resolution of discrepancies led to additions and refinements to the rubric, until a final version was agreed on (see Appendix B). The rubric includes five levels of performance—from 0 (Absence) to 4 (Completely appropriate)—and five dimensions:

• Readability: Effort required on the part of the reader to understand the email purpose and ideas.
• Organization: Presence or absence of framing moves and of a general structure that facilitates identification of key information.
• Request appropriateness: Directness of the request and use of internal mitigators (both syntactic and lexical–phrasal) in relation to the given communicative situation.
• Use of content moves: Relevance and adequacy of the content moves included in the email according to the specific communicative situation.
• Overall sociopragmatic awareness: Appropriateness of register in the communicative situation (student–professor relationship, rank of imposition), evidenced in the level of formality, politeness, and word choices.

Thirty-two additional emails were individually rated and collaboratively discussed by the three raters. Inter-rater agreement scores were calculated using adjacent agreement rates for each of the dimensions in the rubric. This method, commonly used in language testing (e.g., Huang et al., 2018), measures rater agreement within an acceptable discrepancy threshold k. For the present study, we followed previous L2 writing research (e.g., Liu & Kunnan, 2016), and set k = 1, meaning raters’ scores were treated as matching as long as their difference was no greater than one point. To calculate this measure, the number of instances where the raters’ scores differ by one point or less is totaled and then divided by the overall number of cases. Inter-rater adjacent agreement percentages were 82% for Raters 1 and 2, 80% for Raters 2 and 3 and 96% for Raters 1 and 3. Given these results, the remaining emails (n = 220) were analyzed by Raters 1 and 3. In order to avoid rating bias, the emails were anonymized and coded so that raters could not identify whether they belonged to the pre- or the post-test.

The maximum score to be obtained by a participant in each test was 40 points, that is, 20 points per email, given a maximum of four points in each of the five dimensions in the rubric. Scores for global performance, for individual emails, and for dimensions in the rubric were calculated. As the data were normally distributed, paired sample t-tests were conducted to measure the effect of self-access web-based pragmatic instruction (RQ1). ANOVA tests were conducted to explore the possible effects of students’ proficiency level on their performance and on their potential gains between the pre and the posttest (RQ2). Participants’ confidence level in the appropriateness of their emails in the pretest and the posttest, as measured through the Likert-scale data, were compared using the non-parametric Wilcoxon signed-rank test (RQ3).

5. Results and Discussion

5.1. Effects of Self-Access Instruction

Research Question 1 asked whether self-access web-based instruction improved EFL learners’ ability to write appropriate email requests to faculty in English.

Table 2. Mean scores and gains from pretest to posttest.

	N	Pretest M (SD)	Posttest M (SD)	Gains M	t	p
Total score	68	18.81 (5.33)	32.82 (4.88)	14.01	−23.086	<.001
Email 1	68	9.51 (2.85)	16.71 (2.42)	7.19	−22.521	<.001
Email 2	68	9.29 (2.80)	16.12 (2.87)	6.82	−18.265	<.001

Note: The maximum possible total score is 40 points (20 per email).

illustrates mean scores and gains from the pretest to the posttest.

The data indicated that students significantly improved their production of emails after being engaged in the instructional period. Specifically, paired sample t-tests showed statistically significant improvements in both the total score [t(67) = −23.086, p < 0.001] and in the scores for each individual email [Email 1: t(67) = −22.521, p < 0.001; Email 2: t(67) = −18.265, p < .001]. However, the effect sizes were small, as measured by Cohen’s d, indicating that while improvements were statistically significant, the magnitude of the change was relatively modest. As shown in Table 2, scores and gains were slightly higher for Email 1 (the lower imposition request). However, significant differences were not found between the two emails [pretest t(67) = .966, p = .338; posttest t(67) = 2.298, p = .055; gains t(67) = −23.086, p = .281]. Therefore, the effect of instruction did not vary depending on the level of imposition of the email, and production in both scenarios benefitted to a similar extent.

To obtain a better understanding of the effects of self-access web-based instruction on email request production, the different dimensions in the rubric were explored. The paired-samples t-tests showed statistically significant improvements in the five dimensions, as shown in

Table 3. Mean scores and gains from pretest to posttest in each dimension.

Total Score	Pretest M (SD)	Posttest M (SD)	Gains M	t	p
Readability	4.75 (1.16)	6.45 (1.19)	1.70	−11.624	<.001
Organization	3.87 (1.50)	7.25 (1.02)	3.38	−18.891	<.001
Request appropriateness	2.60 (1.37)	6.09 (1.49)	3.49	−16.045	<.001
Content moves	3.57 (1.28)	6.23 (1.17)	2.66	−17.515	<.001
Sociopragmatic awareness	4.01 (1.35)	6.79 (1.08)	2.78	−17.442	<.001

Note: The maximum possible score for each dimension is 8 points (4 per email).

.

Before instruction, participants scored higher in general email Readability (M = 4.75) than in the other four dimensions of the rubric. In contrast, the lowest score was obtained in Request appropriateness (M = 2.60). This dimension was in fact the one that experienced the largest gains after self-access instruction (M = 6.09), followed by Organization, which became the aspect with the highest score (M = 7.25). Remarkably, all dimensions in the rubric reached mean scores above 6 (out of 8 possible points) in the posttest, which indicate high levels of attainment in the task. Figure 3 and Figure 4 below illustrate qualitative changes in students’ performance before and after instruction.

Quantitative analyses showed that students’ performance improved in all the rubric dimensions, but that instruction was particularly effective in increasing students’ ability to produce more appropriate requests (in agreement with previous research, e.g., Alcón-Soler, 2015a, 2015b; Y. Chen, 2015; Nguyen, 2018) and better organized emails. As illustrated in Figure 3 and Figure 4, many emails did not include a clear request in the pretest; indeed, Student 12 writes down “For me is very important if you could help me” (see Figure 3) but does not specify the nature of their request for help. Contrarily, in the posttest both students employed syntactic downgraders targeted in the web-based module, specifically “I was wondering if + it would be possible”. This complex bi-clausal structure was particularly frequent in Email 2 (see Figure 4), which involved a higher level of imposition. This choice of such a sophisticated conventionally indirect strategy mirrors that of expert English users in previous research (Halenko & Winder, 2021). This improvement is thus considered key, as requests strategies convey the main communicative intention in the email (Y. Chen, 2015). Concerning organization, which focused on the inclusion of framing moves and on general email structure, both posttest emails show how after engaging with the module, students better structured their emails into paragraphs and included effective optional framing moves such as pre-closing lines (i.e., Thank you very much for your attention), as can be seen in the two examples from the posttest.

Self-access instruction also had a positive effect on sociopragmatic awareness, and qualitative data showed that this improvement was quite evident in how such awareness was expressed through framing moves. To begin with, students’ ability to write more concrete subject lines improved in the posttest, as has been the case in previous studies (e.g., Y. Chen, 2015; Usó-Juan, 2022). In Figure 3, we can see how Student 12 wrote “problems with the virtual campus” in the pretest while she tried to specify the type of problem in the posttest with the word “access”. Concerning the greeting, even though it was present in all the emails in the pretest, the sociopragmatic awareness of this move varied greatly among students. Emails included a wide range of expressions, from formulas typically found in formal letters (e.g., ‘To the attention of X’, ‘Dear Madam’) to informal salutations with inaccurate terms of address such as ‘Hi Mr Chris Smith’. However, the most frequent expressions (which can be found in Figure 3 and Figure 4) were combinations of greetings that are typically used in Spanish, the participants’ L1, such as ‘good morning’ or ‘good afternoon’ and a term of address (e.g., Nguyen, 2018; Salazar-Campillo, 2018). Finally, in the pretest, the sender’s self-introduction was either missing (Figure 3) or incomplete (Figure 4), as was the case in previous studies (e.g., Y. Chen, 2015; Nguyen, 2018; Usó-Juan, 2021). This framing move is particularly important in first email encounters such as the ones in the present study, as it helps the lecturer identify the student and the course they are enrolled in.

Taken together, opening and closing moves used in the pretest showed that participants had a general awareness of the different statuses lecturers and students had, and believed that the academic context required them to write “formal” emails. However, their pretest emails showed a lack of situational appropriateness, that is, they were not adapted to specific social distance nor to the level of imposition of the email request. Figure 5 exemplifies this shift, as we can see how Student 30 (B2) conveyed formality in Email 1 through highly elaborated (albeit incorrect) formulas in the opening and closing lines, and the repetition of the word “professor”. In contrast, in the posttest the student finetuned the framing moves for the particular situation and included a clear self-introduction, necessary in a first-contact email.

Finally, participants also became more effective in using content moves after instruction. Figure 4 above illustrates, for instance, how the grounder in Email 2 changed from the pretest to the posttest. In the pretest, Student 73 merged a very vague self-introduction with the reason for writing the email, which is not contextualized in a way the recipient can easily understand: “I am a foreigneer [sic] student that has found the course Advanced Grammar too difficult”. In contrast, in the posttest the student provides a clear identification followed by a more contextualized and appropriate grounder: “I’m writing because I’m interested in joining the World Englishes course”. Furthermore, in general terms there was a shift from the students’ main reason provided in the grounder in the pretest (difficulty of Advanced Grammar) to more instances of sweeteners such as “I find World Englishes very interesting” in the posttest.

In addition to this change, before instruction students tended to emphasize their individual needs in their emails, a feature that has been documented in previous research (e.g., Nguyen, 2018). Participants expressed the urgency or importance of their requests with intensifiers such as “as soon as possible” in Figure 3, or the performative verb “I beg” in Figure 4. In contrast, the use of these expressions after self-access instruction was infrequent, and students tried to find ways to lower the sense of pressure put on the teacher. For instance, in their Email 1 written in the posttest several students included a clarifying sentence explaining that they had already asked their classmates for the activities they had to prepare for the following class. Therefore, data show that students started paying attention to key sociopragmatic issues such as lowering the imposition of their request.

Taken together, our data support previous empirical research that shows that explicit pragmatic instruction helps EFL students develop their ability to write appropriate email requests to faculty (e.g., Y. Chen, 2015; Nguyen, 2018; Nguyen et al., 2015; Usó-Juan, 2021, 2022). The major novelty of the present study is that it shows that students’ performance in such emails can be positively impacted by a self-access web-based module that follows a task-supported approach and incorporates a combination of input, metapragmatic information, practice, and feedback through sample answers.

5.2. Influence of L2 Proficiency Level on Self-Access Instruction Effects

Research Question 2 explored the influence of L2 proficiency on the effects of self-access web-based instruction.

Table 4. Pre- and post-test mean total scores and gains across proficiency groups.

Group	N	Pretest M (SD)	Posttest M (SD)	Gains M	t	p
B1	22	16.04 (4.61)	30.04 (4.96)	14.00	−10.915	<.001
B2	23	18.91 (5.38)	33.65 (3.54)	14.74	−15.907	<.001
C1	23	21.35 (4.78)	34.65 (4.98)	13.30	−13.941	<.001

illustrates pre- and post-test total scores and gains across the three proficiency groups.

Paired sample t-tests showed that the three proficiency groups experienced significant gains after instruction, both in their total scores and in individual emails. Descriptive statistics indicated that B2 level students achieved slightly higher gains (M = 14.74) compared to B1 level students (M = 14.00) and C1 level students (M = 13.30). However, when a one-way ANOVA test was conducted, no statistically significant differences were found between groups in terms of amount of gains (F(2, 65) = 0.46, p = .630).

The mean scores and gains obtained by the three groups across the five dimensions in the pre and posttests are presented in

Table 5. Mean scores and gains across proficiency levels and rubric dimensions.

	B1		Gain	B2		Gain	C1		Gain
	Pre M (SD)	Post M (SD)		Pre M (SD)	Post M (SD)		Pre M (SD)	Post M (SD)
Readability	4.00 (0.87)	5.59 (1.05)	1.59	4.61 (1.03)	6.61 (0.99)	2.00	5.61 (0.99)	7.13 (1.01)	1.52
Organization	3.32 (1.42)	7.04 (1.13)	3.72	3.96 (1.55)	7.43 (0.97)	3.47	4.30 (1.40)	7.26 (1.05)	2.96
Request appropriateness	2.31 (1.39)	5.41 (1.38)	3.10	2.52 (1.20)	6.13 (1.22)	3.61	2.96 (1.50)	6.70 (1.33)	3.74
Content moves	3.00 (1.02)	5.59 (1.10)	2.59	3.65 (1.33)	6.39 (1.03)	2.74	4.04 (1.30)	6.70 (1.15)	2.66
Sociopragmatic awareness	3.41 (1.26)	6.41 (1.14)	3.00	4.17 (1.40)	7.09 (0.79)	2.92	4.43 (1.24)	6.87 (1.21)	2.44

. As can be seen, the three groups experienced significant gains in all five dimensions of the rubric. B1 students experienced larger gains in organization, followed by request appropriateness, while B2 and C1 groups also had the largest gains in these two dimensions but in reverse order. Readability, which was the highest scoring dimension in the pretest for the three groups, was the area that registered the lowest gains. The ANOVA tests showed no significant differences in terms of amount of gains for any dimension across proficiency levels.

Data indicated that scores were generally aligned with proficiency levels—with the C1 group outperforming the B2 group, and the B2 group outperforming the B1 group in total scores and individual dimensions. The only exception was found in the B2 group’s score for Sociopragmatic awareness in the posttest, which was higher than the C1 group’s one. Only one significant difference emerged from the one-way ANOVA tests conducted, and it was in Readability. C1 level students achieved significantly higher Readability scores than B1 level students, both in the pretest [C1 M = 5.61; SD = 0.99; B1 M = 4.00; SD = 0.87; t = 5.57; p < .001] and the posttest [C1 M = 7.13; SD = 1.01; B1 M = 5.59; SD = 1.05; t = 5.07; p < .001].

Therefore, quantitative data showed that self-access web-based pragmatic instruction was effective for improving learners’ ability to write appropriate email requests to faculty, irrespective of their proficiency level. In fact, students’ posttest emails evidenced that the self-access module provided learners at diverse proficiency levels with knowledge (such as formulas and pragmatic strategies) they could apply to fulfill their communicative intention in their email requests. Figure 6 and Figure 7 illustrate how the pre-closing formula “Thank you for your time” was used by both a C1 and a B1 student to solve issues which, at pretest time, were of a different nature.

Figure 6 illustrates how proficiency and pragmatic performance do not necessarily go hand in hand if no pragmatic instruction is received. Student 36 (C1 level) closed her email with the pre-closing “Please, get back to me as soon as you can” and used “Thank you” as a closing to express gratitude. Despite her high level of proficiency, she resorted to transferring a typical use of her L1 to perform a directive (imperative + please) (e.g., Félix-Brasdefer, 2012) to a context where it may not be the most pragmatically appropriate option (Ishihara, 2018). In fact, in her open answer she expressed her uncertainty about her closing lines: “I’m not sure if the way I did it was the most appropriate. What I was aiming for was to end with a formal sentence that conveyed respect towards the teacher, but I didn’t know how to phrase it” (original in Spanish). After instruction, she used the pre-closing line “thank you for your time” to express gratitude and closed the email with the formula “Kind regards” and a complete signature.

In the case of Student 29 (B1), illustrated in Figure 7, the participant wrote the incorrect pre-closing line “Thank you previously and I hope you have a nice day”, in the pretest which reflected her need for linguistic resources to better fulfill her intentions, which were to express the idea of “in advance”. In fact, in her open answer she acknowledged that she was not confident about her pre-closing line as she felt that she had “Spanishised it a little bit” (la última frase no me transmite mucha seguridad ya que a lo mejor la españolicé un poco). In the posttest she wrote “Thank you for your time. I hope you are having a good day”, keeping her original intention to use a sweetener, but also correctly using one of the preclosing formulas suggested in the module (and adding an appropriate closing and an informative signature).

Our data are not in line with previous research focusing on self-access web-based pragmatic instruction (e.g., Kerber et al., 2023; Yang, 2017), which suggests that this type of instruction may not be so beneficial for lower-level students. This difference in results may have been caused by the nature of the target items explored and the data collection methods used: while other studies have focused on oral speech acts (such as expressing gratitude or apologizing) and measured learning through written discourse completion tests, the present study explored a written genre (requestive email) and knowledge was tested through communicative writing tasks. The match in modality between target, practice and assessment may have better captured progress for all groups (e.g., Bardovi-Harlig, 2018).

A further possible explanation for the positive effect of self-access instruction even for B1 students is that the self-access module included a high amount of metapragmatic feedback embedded both in controlled interactive activities and in the tasks that involved email analysis, production and model comparison. This would be in line with results in previous research on instruction and feedback on request development. For instance, Y. S. Chen et al. (2023) explored the effects of different types of feedback for the development of request downgraders and found that, while high-proficiency learners benefited similarly from direct and metalinguistic feedback, metapragmatic feedback proved more effective for low-proficiency learners.

5.3. Effects of Instruction and L2 Proficiency Level on Confidence Level

Research Question 3 asked whether self-access web-based pragmatic instruction had an effect on EFL learners’ confidence level when judging the appropriateness of their email requests to faculty and whether confidence level was affected by L2 proficiency. The Wilcoxon signed-rank test showed that students reported confidence experienced a significant increase, both in terms of global ratings and in each individual email, as shown in

Table 6. Pre- and post-test mean confidence scores and gains.

	N	Pretest M (SD)	Posttest M (SD)	Gains	Z	p
Global rating	68	3.55 (0.63)	4.03 (0.46)	0.48	−5.133	<.001
Email 1	68	3.51 (0.72)	4.10 (0.55)	0.59	−4.835	<.001
Email 2	68	3.58 (0.73)	3.97 (0.54)	0.39	−3.281	<.001

Note: The maximum possible confidence rating was 5.

.

Confidence ratings rose from an average of ‘somewhat confident’ in the pretest (M = 3.55; SD = 0.63) to ‘quite confident’ (M = 4.03; SD = 0.46) in the posttest, and gains were observed across all proficiency levels.

As illustrated in

Table 7. Mean confidence ratings and gains across proficiency levels.

Group	N	Pretest M (SD)	Posttest M (SD)	Gains	Z	p
B1	22	3.36 (0.51)	3.89 (0.51)	0.53	−3.152	.001
B2	23	3.54 (0.62)	4.11 (0.45)	0.57	−3.245	<.001
C1	23	3.74 (0.72)	4.11 (0.42)	0.37	−2.499	.011

, B2 students exhibited slightly greater improvements in confidence, followed by B1 students; however, differences among groups were not statistically significant. Therefore, while instruction had an effect on confidence level, L2 proficiency did not.

The fact that instruction positively affected participants’ confidence levels is in agreement with previous research (Y. Chen, 2015; Usó-Juan, 2021) that supports the effectiveness of explicit pragmatic instruction in raising students’ level of confidence in the appropriateness of their email requests. It should be highlighted that the novelty of the present study relies on the fact that it is the first one to explore the effects of self-access instruction on students’ state self-confidence. To better understand how confidence developed under these conditions, students’ open answers were analyzed to shed light on their sources of insecurities in the pretest, as well as their feelings of confidence after instruction.

Students’ insecurities in the pretest were mostly related to their perceived lack of (i) linguistic resources and (ii) “real-world” practice in these communicative scenarios. Students’ sense of insecurity due to insufficient L2 resources was common across B1 level students, as illustrated in Example 1 below:

(1)

“With these activities, I feel quite insecure because I think I don’t express myself as well as I should, or that I don’t have the vocabulary I should have. I feel like I’m constantly making spelling mistakes or that I’m expressing myself poorly.” (Original in Spanish) (Student 51; B1; Email 2; pretest)

In fact, even though no statistically significant differences were found between groups, descriptive data showed that no student in the B1 group chose the rating 5 (completely confident) in the pretest, while students in the other two proficiency groups did. This would be in agreement with the role that linguistic competence plays in students’ L2 confidence levels (MacIntyre et al., 1998).

More specifically, participants also expressed their concerns about their lack of ‘formal’ vocabulary and formulas they felt were needed to construct emails that were appropriate to the context. As shown in Example 2, B1 students resorted to using the terms in their linguistic repertoire even when they felt they may not be the most appropriate ones, or to adapting words or coining expressions from their L1 (such as “thank you previously”):

(2)

“It really has been a bit difficult for me to write it [the email] overall, because I think some of the words I used may not be appropriate in this context, but I couldn’t find another way to say them in English. Also, I think some of the sentences sound a bit ‘Spanish-like’.” (Original in Spanish) (Student 62; B1; Email 2; Pretest)

This issue was not restricted to lower-level students, as even C1 participants were worried about not being able to find appropriate English expressions. Even though more advanced participants knew they had grammatically correct options available, frequently they were not sure whether these were an adequate alternative, as illustrated in Example 3 below:

(3)

“I’m not sure whether it [the email] is entirely appropriate. [...] I’ve realized that I can’t think of any other way to begin the email besides saying ‘I am getting in touch with you,’ and I’m not even sure whether that phrase is the most appropriate for a formal context.” (Original in Spanish) (Student 3; C1; Email 2; Pretest).

The second concern that appeared across all proficiency levels was the lack of practice and of real-world application of the knowledge acquired in previous instruction. Examples 4 and 5 show how two students from B1 and B2 groups explain that they learned how to write formal emails just to pass an official exam:

(4)

“I find it difficult to write emails in English when I have to use formal language because it’s something I’m not used to and don’t use in my daily life. [...] I’ve only used this register in assignments or official exams, but never as part of my routine.” (Original in Spanish) (Student 71; B1; Email 2; pretest)

(5)

“The biggest difficulty is using formal language when writing an email in another language, especially since when I studied this, it wasn’t taught with the goal of being used in the future, but just to pass an exam.” (Original in Spanish) (Student 22; B2; Email 1; pretest)

Students’ answers illustrate that, in addition to language competence, prior L2 contact and experiences also play an important role in state communicative confidence (MacIntyre et al., 1998). In the case of the present study, even though students had different proficiency levels, they were all novice academic email writers. Therefore, their shared limited L2 experiences with this communicative task may have interacted with their differing proficiency in the L2, making their confidence levels very similar to each other.

The answers included in the posttest shed light on the most frequent sources for their increased confidence. As reflected in Examples 6 and 7, the two most repeated elements that appeared in them were the amount of practice the module had provided, and the reflection performed during the module, and particularly in the third task. This task involved comparing their own emails to sample answers accompanied by metapragmatic explanations.

(6)

“After writing so many emails I don’t find it difficult anymore; I just try to find the best way to write them so they are specific, clear and appropriate” (Student 37; C1; Email 1; posttest)

(7)

“I haven’t really found any difficulties, I only paid more attention to the things that I saw that my emails lacked of [sic] when I compared them to the models in Task 3.” (Student 2; B2; Email 1; posttest)

As suggested by Keller (2010), for materials to foster confidence they should provide students with feedback and means for action with such feedback, opportunities for successful practice and control over their actions. In this sense, the module provided students with guided and clear feedback, as well as control over the amount of time and practice they wanted to invest, together with activities that were considered optional and that gave them freedom to choose the topics that were most relevant to them. In addition to actionable feedback and practice, the module included six reflection activities (one after each module and one after each task) such as the ones proposed in positive psychology interventions to promote language learner confidence. According to Shelton-Strong and Mynard (2020) “these tasks provide opportunities for learners to not only document and reflect on their learning and feel a sense of achievement, but to think deeply about the [learning] process” (p. 5). As a result, students were able to reflect on their learning through the module and many of them became aware of how their conceptualizations of what an appropriate email was had changed, as illustrated in Example 8 below:

(8)

“I think that thanks to the skills and abilities I’ve developed during this project, I feel more confident about what I’ve written. I used to think that the more text I wrote, the better the task would be. But now I’ve realized that sometimes brevity and clarity are better than writing a lot.” (Original in Spanish) (Student 51; B1; Email 2; posttest)

Taken as a whole, our results show that a carefully designed self-access web-based module (which includes explicit metapragmatic explanations and feedback, and opportunities for input analysis, output production and reflection on the learning process) can have a positive impact on EFL learners’ ability to produce more appropriate email requests as well as on their state communicative confidence, irrespective of their proficiency level and even when with no individualized feedback or interactions with a lecturer are provided.

6. Conclusions

The present study pursued three objectives: (i) to examine whether self-access web-based instruction could significantly improve EFL students’ ability to formulate pragmatically appropriate email requests; (ii) to determine whether L2 proficiency influenced the effects of instruction; and (iii) to explore changes in learners’ confidence when evaluating the appropriateness of their own email requests. The results showed statistically significant improvements across all five rubric dimensions following instruction. The most substantial gains were observed in request appropriateness and organization, suggesting that self-access instruction was particularly effective in enhancing learners’ ability to mitigate requests and structure emails appropriately. Participants achieved slightly higher posttest scores for the lower-imposition request than for the higher-imposition one. Analysis by proficiency level revealed that all groups benefited similarly from instruction, with no statistically significant differences in gains. This may be regarded as an important finding, as our data suggest that carefully designed self-access modules can be used successfully in contexts with a disparity of proficiency levels, as is often the case in university settings. In addition to improvements in performance, students reported a significant increase in confidence when evaluating the appropriateness of their own emails. Confidence ratings rose from an average of ‘somewhat confident’ to ‘quite confident’ post-intervention, with gains observed across all proficiency levels. While B2 students exhibited slightly greater improvements in confidence, these differences were not statistically significant.

This study makes several contributions to the field of L2 pragmatics and digitally mediated language instruction. First, it offers empirical evidence that a self-access, web-based instructional module can significantly enhance EFL learners’ ability to produce pragmatically appropriate email requests to faculty, supporting the viability of autonomous learning approaches in university settings. Second, by focusing on high social distance scenarios—namely, first-contact emails to unfamiliar instructors—it addresses a previously underexplored dimension of email pragmatics. Third, the study contributes to ongoing discussions about the role of L2 proficiency in pragmatic development. Fourth, the research expands existing frameworks by incorporating learners’ self-reported confidence as a meaningful outcome, demonstrating that self-access instruction not only improves performance but also fosters greater self-assurance in email communication. Finally, the findings have direct pedagogical relevance, highlighting the potential of blended learning environments and web-based resources to promote pragmatic competence in resource-constrained EFL classrooms.

While the study yielded positive outcomes, the absence of a control group (due to ecological reasons) introduces the possibility that task repetition contributed to the observed improvements. Although most participants indicated they did not recall the pretest scenarios, this potential influence cannot be entirely dismissed. Therefore, future research should test the possible effects of task repetition. In addition, in the present study authentic emails were not collected due to the limited availability of such data. Participants were novice writers who reported infrequent use of English for emailing instructors, as most of their coursework during their first semester was conducted in their L1. Future research should therefore investigate how self-access instruction influences students’ email communication in real-life academic contexts. Even though the module offered feedback through model emails and metalinguistic explanations, this type of feedback is not customized to individual student performance. Given the importance of feedback in language development, future research should examine whether integrating individualized feedback into the web-based module enhances learners’ pragmatic gains to a greater extent. A possible avenue for this research would be training artificial intelligence to offer feedback through analytic rubrics such as the one used in the present study (e.g., Bertoldi & Gesuato, 2025). Finally, exploring other individual variables such as learner engagement with the web-based module would deepen our understanding of what drives successful autonomous pragmatic learning.