Application of corrective feedback using emerging technologies among L2 university students

Abstract The majority of foreign language learning (FLL) has recently been conducted online due to unprecedented changes in society and this trend seems to be global and perpetual even if the situation changes. Unfortunately, there is still low awareness among language instructors of the impact it may have on the corrective feedback they provide to their students. Traditionally, corrective feedback was provided to the students during or after the class orally or in written form, but the personal aspect of this feedback was present throughout the evaluation process. With digital media, the situation is rather different caused by the remoteness of the learning process participants. The present systematic review attempts to provide a systematic summary of cutting-edge research into the topic of corrective feedback with the use of emerging technologies for FLL in the university context. The findings indicate that there is not much awareness of the specific context of digital corrective feedback among tutors. Moreover, from an objective perspective, it seems that digital corrective feedback can be very efficient if applied correctly. The review provides a summary of new trends but calls for a more systematic and analytical approach to the topic that is still rather neglected in the studies currently available.


Introduction
Corrective feedback seems to be one of the most important aspects of any second language (L2) education or foreign language learning (FLL) as it provides important data for the learners about their progress, mistakes, and other important language issues. When a significant proportion of FLL is conducted online, it is even more important to focus on this topic as corrective feedback can be a serious challenge in any FLL both for the instructor, as the person who sends the message, but also for the receiver of this coded information, i.e., the learner.
Feedback is endemic to students' learning and successful academic achievements, be it positive or negative. Generally, feedback is defined as a response that students receive about the performance they produce (Van Patten & Benati, 2015). Corrective feedback is considered to be negative since it identifies an error or errors in students´ performance (Ceman & Dubravac, 2019). Nevertheless, as several research studies (Ceman & Dubravac, 2019;Ellis, 2009;Hattie & Timperley, 2007;Lyster, 1998) reveal, corrective feedback, and especially the explicit one, when a teacher explains what mistakes were made and what the right form is, seems to be beneficial in achieving students´ learning outcomes.
Generally, there are eight types of corrective feedback (Fathimah, 2020), which can be divided into input-providing (the teacher explains the correct form) and output-prompting (the teacher asks the student about the correct form). The input-providing feedback includes the following categories: 1. recast-the teacher reformulates all or part of a student's utterance, minus the error, 2. explicit correction-the teacher provides the correct form and clearly indicates that what the student said was incorrect, and 3. translation-the teacher translates the wrong expression or utterances into student´s L1. The out-put providing feedback categories are then as follows: 4. metalinguistic feedback-the teacher provides additional information or questions related to the correct form of students' expression or utterances, without explicitly providing the correct form, 5. elicitation-the teacher attempts to induce the correct form by asking for completion of a sentence, or asking questions, or asking for a reformulation, 6. clarification-the teacher asks for reformulation of the expression or utterance since it is wrong and the information cannot be understood, 7. repetition-the teacher highlights student's wrong expressions or utterances by repeating them, 8. paralinguistic signal-the teacher by his/her gesture or facial expression signals that the student's expression or utterance is wrong.
At present, corrective feedback might be even more effective thanks to new emerging technologies, such as computer-assisted language learning (CALL) or tools of artificial intelligence (AI), which can provide automatic corrective feedback. Although there have been recently a few research studies on electronic written corrective feedback (cf., Altamimi & Masood, 2021;Mohsen, 2022), almost no research study has explored other types of computer-mediated, electronic, or digital corrective feedback. For instance, the findings of the study by Altamimi and Masood (2021) show that not much has been done on the exploration of different types of digital feedback in writing tasks. Most of the studies have dealt with recasts and metalinguistic types of feedback. This was also true for the study by Fathimah (2020) in which teachers most frequently used recasts to correct students´ mistakes and students themselves welcomed the direct and explicit feedback.
Furthermore, the meta-analysis on computer-mediated corrective feedback for L2 writing performed by Mohsen (2022) illustrated that non-automated corrective feedback had a larger effect size than the automated one. In addition, computer-mediated corrective feedback was far more beneficial for beginners and intermediate L2 learners and improved both their writing fluency and accuracy. Cornillie et al. (2012) examined the use of corrective feedback in digital games designed for language learning and they found out that explicit corrective feedback was more useful for students than implicit one. Moreover, Al-Olimat and Abuseileek (2015) in their study on computermediated corrective feedback in L2 writing performance also indicated that the most efficient corrective feedback was when it was provided by both teachers and peers.
Nevertheless, as recent research (Payne, 2021) demonstrates, and especially after the online teaching and learning during the COVID-19 pandemic, students prefer digital corrective feedback, especially the audio-video one, because it can be more effective at building rapport, it also gives them a sense of community online and involves them in the overall process of evaluation. In addition, such feedback appears to be more personalized and specific since students can react and asks verification and clarification questions. Thus, this kind of feedback creates some kind of social environment, rather than just an impersonal negative evaluation.
However, as further research points out (De Vries et al., 2010) corrective feedback, and as mentioned above the explicit corrective feedback in particular, has a significant impact on second language acquisition than the language input alone. Therefore, this review study will focus on the exploration of the application of different types of corrective feedback using emerging technologies among L2 university students.
Based on the available references and sources, the following research questions were formulated to be able to analyze the potential, but also limitations, of corrective feedback in digital FLL in the university context.
(1) Is digital corrective feedback in FLL effective in the university context?
(2) What types of digital corrective feedback are useful in FLL in the university context?
(3) What are the limitations of corrective feedback in digital FLL in the university context?

Methods
To obtain relevant data the PRISMA methodology for systematic reviews and meta-analyses was fully followed. The search was conducted in the Web of Science and Scopus databases to collect open-access peer review journal articles written in English. The focus of this review study was to bring together experimental studies with the exclusion of non-experimental ones, such as theoretical, conceptual, or review studies, in order to generate the most objective, reliable, and valid results. The key methodological concept necessary for this study was a delayed post-test to obtain statistically relevant, measurable, and replicable results.
The key concept relevant for this study was second (L2) or third language (L3) acquisition, or foreign language learning, therefore, all studies that dealt with other aspects of corrective feedback or in other, i.e., non-foreign language acquisition contexts were excluded. Also, all studies that dealt with first language acquisition (L1) were excluded. All grey literature and conference proceedings were not considered relevant for this study as it aims at bringing together all cuttingedge research but only the results that have been evaluated by a very rigorous peer review process, which is not guaranteed in conference proceedings that present rather preliminary findings that still need further verification. The same applies to pre-prints, that is why they were excluded as well.
The time span of the studies analyzed was not restricted, however, naturally, basically all studies were published in the few past years. However, it was important for this study to focus on university students as the corrective feedback would vary significantly in other age groups and it is not possible to provide a generally accepted approach for younger learners, high school, and university students. Therefore, this study deals with the context of university FLL.
The dataset for this systematic review was created in mid-July 2022 (15-20 July 2022) when the database search was conducted, all relevant articles were collected, manually checked for their relevance and, applying the following inclusion and exclusion criteria, selected all related and relevant studies. After their careful verification, their analysis was conducted, referential backtracking was conducted as well to obtain more related studies, then, the results were summarized and implications generated.

Search terms (both for web of science and scopus)
The following search string was created using Boolean operators to obtain relevant studies related to the topic: "corrective feedback" AND "university" AND "digital" OR "electronic" OR "computer-mediated".
This search string was applied in both Web of Science and Scopus databases. After that, all inclusion and exclusion criteria were applied to identify relevant articles. The analysis of the articles obtained was conducted manually by both authors of the article to ensure that all inclusion and exclusion criteria were met after the first set of articles was generated from the databases. Referential backtracking was also applied to identify further studies that were relevant and important for the current study but were missed by the standard database search.
The initial search generated 61 results from Web of Science and 67 from Scopus after the search string was applied. After the manual implementation of all inclusion and exclusion criteria, four studies from Web of Science and two studies from Scopus were obtained. A total of six studies were considered for further analysis and investigation. Due to the fact that Web of Science and Scopus did not provide enough journal articles, it was necessary to include Google Scholar and Science Direct in the search to obtain more results that could be processed. After the implementation of these two databases, three more articles were identified as relevant and they were also analyzed.

General information
Altogether nine relevant studies were identified. Three studies originated in the USA, two studies in China, followed by a study from Saudi Arabia, Kuwait, India, and Spain. Most of the studies have been published recently as they date back to 2020-2022, while the oldest study was published in 2009. The number of participants in each study also differs; it spans from three subjects to 311 subjects.

Methodology of the detected studies
Therefore, most of the studies also differ in their methodological design and procedure, as well as outcome measures. Only two studies in fact included pre-, post-tests and delayed post-tests-a tool that can provide very reliable data and their interpretation. Some of the studies not only focused on content analysis of the students´ performance but they also conducted questionnaire studies that seem relevant as they can provide a more subjective reflection of the given corrective feedback process by the students. The same is true for interviews conducted after the intervention was conducted, again, with the aim of receiving a subjective evaluation of the situation. The data yielded through qualitative means, such as questionnaires and interviews, could be compared with the statistical results, and thus, provide a broader insight into the area.

Type of digital corrective feedback
The main focus was on digital corrective feedback in writing tasks as they are relatively easy and manageable via various digital platforms, unlike speaking and listening, for instance. Altogether seven studies exclusively concentrated on written corrective feedback. One study (Tang et al., 2021) dealt with both written and oral corrective feedback and only one study (Calvo-Ferrer, 2021) explored the role of corrective feedback in the acquisition of L2 vocabulary. For example, Ene and Upton (2014) focused on the feedback provided by the teachers in writing and they concluded that it consisted of marginal comments related to minor issues, i.e., finetuning the manuscript of the participants, rather than having a significant impact on their language proficiency. Naturally, their feedback dealt mainly with grammar issues and vocabulary improvements. Therefore, their findings are not very counterintuitive and cannot provide much deeper insight. However, they claimed that electronic feedback was successful in 62.3%, which is, however, very difficult to replicate and verify. But it is clear that this kind of corrective feedback can be extremely helpful for grammar revision. Their subsequent study in2 2018 provided more insights and much more important findings. They claim that digitally-mediated feedback is an important affordance and the 21stcentury FLL must be ready to rely on that. It was obvious that their research identified several aspects that are crucial for enhancing writing skills but they still stick to their initial idea that annotated text can provide corrective feedback in the area of grammar development and vocabulary enhancement.
The majority of the researched studies focused on a crucial topic of the feedback provided by the teachers when assessing the performance of the students instructed online with the aim of analyzing the quality of this corrective feedback. The format of the feedback was usually in the form of an audio-visual mode, sometimes oral feedback and written feedback was the least prominent (Alharbi, 2022). The same research confirms that the students would welcome more specific and detailed feedback and the analysis of their language issues, which is, however, demanding and extremely complicated and time-consuming. Explicit corrective feedback is, naturally, an integral part of any digital L2 acquisition but due to its challenging nature, it is very often considered to be one of the most complicated parts of digital FLL (Calvo-Ferrer, 2021). Moreover, clear and proper corrective feedback can have a significantly positive impact on L2 acquisition as it provides a strong motivator for the learners. These findings were further expanded by Xu (2021) who claims that students who seek feedback also possess self-regulated learning writing strategies. In addition, his students admitted having a positive attitude towards online written corrective feedback because they could review it as many times as they wanted and they were engaged in their writing tasks more thanks to frequent teacher-student online interactions. Elsayed and Hassan (2020) confirmed that digital indirect feedback is more efficient than traditional indirect corrective feedback and it can also have long-term effects on various aspects of accuracy. In the context of traditional written feedback, digital corrective feedback can be a useful tool to improve the students´ learning outcomes as it provided them with an instant reaction to their mistakes, which seems crucial in any L2 acquisition, rather than a delayed written corrective feedback which will not have such an impact.
Sauro, as early as 2009, provided some interesting findings that implicit corrective feedback (recasting and metalinguistic comments) will support linguistic gains in L2. This was in fact confirmed in 2021 by Tang et al. in their study on telecollaboration (the so-called e-Tandem project) in which native speakers of corresponding L1 language provide corrective feedback to their L2 peers. They reported that especially in Skype discussions, the most commonly used feedback was a clarification request, which is again implicit corrective feedback.

Effectiveness of digital corrective feedback
Overall, the results of this review reveal that digital corrective feedback seems to be effective and quite stimulating for L2 learners since by applying it to their L2 learning, they could see progress in their learning achievements, which, eventually, leads to the improvement of their L2 proficiency. In addition, it might also have a very beneficial effect on their self-regulated learning. Moreover, the findings also indicate that audio-visual or online oral explicit corrective feedback appears to be the most effective and stimulating for students. In this respect, one could also claim that the future of the use of digital media in FLL will depend on the fact of how well they will be able to provide feedback to their learners in order to keep them motivated to study L2.

Benefits and limitations of the selected studies
One of the benefits of the reviewed studies might be seen in the development of handling the issue of corrective feedback, i.e., going from the handwritten one to the digital one, analyzing whether the digital one is efficient, exploring different digital media in using written corrective feedback, as well as identifying other language areas in which corrective feedback can be used, such as the acquisition of L2 vocabulary.
Not all studies deployed delayed post-tests and measured student language proficiency after completing the task, which must be seen as a major limitation The detailed results, based on the inclusion and exclusion criteria set in the part on Methodology, are summarized and presented in Table 1 below.

Discussion
Based on the results described above, the authors are able to answer the research questions set in the Introductory part.

Is digital corrective feedback in FLL effective in the university context?
The findings of this review study reveal that digital corrective feedback seems to be more effective than traditional, i.e., handwritten feedback (Alharbi, 2022); (Elsayed & Hassan, 2020); (Xu, 2021), although the combination of both might be even more effective in enhancing students´ writing performance (Ene & Upton, 2014, 2018. However, further studies are needed as the data presented do not seem sufficient to draw a clear conclusion about whether they provide a longterm effect (cf., (Al-Olimat & Abuseileek, 2015). More studies should also be connected to the digital nature of the current young generation (Gen Z, Millennials, or digital natives) as it is necessary to relate further studies to various drawbacks of digital media from a cognitive perspective and also from the perspective of wellbeing, such as increased screen time issues, a lack of social interactions, or possible AI addiction ; Klimova et al. (2021); Pikhart and Klímová (2020).

What types of digital corrective feedback are useful in FLL in the university context?
The results also confirmed that both explicit and implicit digital corrective feedbacks were undoubtedly beneficial and useful. The explicit corrective feedback is especially useful for online written corrective feedback, while implicit feedback appears to be more suitable for online oral feedback where more interactions between the teacher and the student or L1 student and L2 student are possible. In addition, even just the audio-visual and voice feedback modes seem to be more beneficial than just the online written feedback because the teacher might be more specific in his/her comments when providing the audio-visual and voice feedback (Alharbi, 2022). Nevertheless, when speaking about the types of feedback, one should also consider the student ´s learning style and the fact that the teacher´s feedback might be affected by the nature of the task itself (Ene & Upton, 2018;Sauro, 2009;Tang et al., 2021).

What are the limitations of corrective feedback in digital FLL in the university context?
Basically, all studies conclude that corrective feedback cannot be used intuitively or based on some previous experience with corrective feedback that was conducted personally when in The field of L2 writing should pay attention to digitally-mediated feedback as an important affordance of 21st century literacy and pedagogy.
(Continued) Klimova & Pikhart, Cogent Education (2022) Students had a positive attitude towards online WCF since they could review it as many times as they wanted and they were engaged in their writing tasks more thanks to frequent teacherstudent online interactions.
In addition, students' feedback-seeking orientation was positively associated with self-regulated learning writing strategies. The online WCF can be used as support to the face-toface writing courses. Teachers should provide some guidance on how to apply self-regulated learning writing strategies during the writing process by providing specific examples. When implementing the peer feedback or interactions, students´ proficiency level should be considered.
a classroom, which is a crucial aspect as the majority of the instructors have not received any systematic training on how to apply corrective feedback when using digital media for L2 acquisition (cf., Ene & Upton, 2014, 2018(Mahapatra, 2021); (Sauro, 2009); (Tang et al., 2021). If this condition is not met, the teachers will be just intuitive in their corrective feedback, which will lead to a systemic issue in digital learning, i.e., the tool used for FLL will not be used adequately and the learning aim will not be achieved or with a lot of effort.

Limitations of this review study
The major limitation of this study is a lack of a large number of sufficient data from various digital learning environments and the description and analysis of the implementation of corrective feedback with relevant statistical outcomes that could be further processed. However, despite this limitation, this study still provides reliable and systematic data that clearly show the potential and drawbacks of corrective feedback possibilities when using digital media for FLL.

Significance of this study and future research directions
The findings of this review study point to the significance of digital corrective feedback, especially oral one, which has a considerable impact on the improvement of students´ learning outcomes and their self-regulated learning. In addition, it might have a long-term effect on their language accuracy.
Overall, this review study has shown what is currently known regarding the opportunities of corrective feedback in FLL or L2 acquisition, i.e., from the perspective of applied linguistics, but it should also motivate further studies as it has been highlighted already that it is necessary to verify basically all previous findings with more empirical results so that more theoretical conclusions regarding FLL with the use of emerging technologies can be drawn.
This study also has the potential to stimulate further research activity so that more comparable data will be generated and a meta-analysis could be conducted. In addition, more pedagogydirected research investigating the effectiveness of digital corrective feedback should be performed.