The Benefits of Social Media Comments to L2 Listening Comprehension

The purpose of this study is to determine whether a rich repertoire of meaningful, relevant comments posted on social media sites can facilitate or increase L2 listeners’ comprehension of videos. The study utilized a quasi-experimental design with an experimental group (n = 29) that watched two YouTube videos and read comments, as well as a control group (n = 31) that watched the same videos without comments. After watching each video, students in both groups wrote summaries of the video content, and only the experimental students responded to brief questionnaires assessing their perceptions of the video’s comments. T-tests revealed that summaries written by the experimental group had significantly more, and also more informed and accurate, unit ideas. Results of the questionnaires revealed that there is a generally positive attitude toward comments as an L2 listening help option. These findings suggest that social media comments may be a useful tool to assist or enrich listening comprehension of videos. Pedagogically, when social media videos are selected for teaching or learning L2 listening, the richness, relevance, and meaningfulness of comments should be among the video’s selection criteria.


Introduction
In the last decade, we began to see revolutionary developments in second language (L2) listening materials and delivery methods (Cross, 2018;Vandergrift & Goh, 2012), including technological advances in commercial media websites, social media sites, and streaming services.Such websites and services caught the attention of practitioners and researchers in computer assisted language learning (CALL) field, who in turn always recommend them as excellent resources as engaging, appealing, and authentic videos for L2 listening, teaching, and learning (e.g., Blake, 2016;Hubbard, 2017;Terantino, 2011).It is also obvious that L2 learners today, with the Internet, the availability of free online videos, and the affordability of smartphones, have plenty of options for L2 listening practice, such as YouTube, that go beyond the classroom materials (e.g., Blake, 2016;C xakmak & Erc xetin, 2018;Godwin-Jones, 2017, 2019;Hubbard, 2017;Wang & Chen, 2020).
However, when it comes to the help options that could assist L2 listening in such self-learning environments, L2 listening CALL research has not fully kept up with the new technology.More specifically, L2 listening CALL research has been and still is, even in the most recent research, hovering around a limited set of traditional help options that were developed long before the internet, not to mention web 2.0 and social media.For example, even the most recent L2 listening CALL studies (e.g., Ca´rdenas-Claros, 2020;Hsieh, 2020;Kam et al., 2020;Montero Perez et al., 2018;Pujadas & Mun˜oz, 2019;Teng, 2020;Vanderplank, 2019;Yang, 2020) of those who utilized YouTube videos have considered traditional help options such as captions, subtitles, transcripts, audio/video control buttons, transcript, dictionary, and glossary.
We do not doubt the benefit of the abovementioned help options; nonetheless, they have existed for many years-when only offline materials were available (e.g., on CDs) and when social media videos did not exist.Indeed, L2 listening research seems to have refrained from exploring the unique help options that the internet, and particularly the Web 2.0 and social media, can afford to L2 listening comprehension, such as social media comments.To explain: YouTube videos and similar social media sites are typically made available to the public with three embedded, ready-to-use traditional help options (namely captions, subtitles, and audio/video control buttons) and one yet to be built (comments).The former set of help options need to be fully designed before publishing or must be made fully available by platform once the video is published online (further improvement or modification is likely not possible).On the other hand, comments only begin to generate after publishing of the video is complete; however, they continue to grow and endlessly increase.As long as the comments section expands and improves, effectiveness presumably also improves.
The point of this research is to draw the attention of CALL practitioners and researchers to the potential benefit of social media comments to listening comprehension.Also, we aim to inform L2 learners that their comprehension can be helped by online comments right beneath the videos they regularly watch on YouTube.Indeed, content analysis of YouTube (e.g., Dubovi & Tabak, 2020;Khan, 2017;Madden et al., 2013;Schultes et al., 2013) found that comments perform various rhetorical functions (e.g., quoting, restatement, and elaboration) that could afford some sort of linguistic assistance to L2 listeners.
Hence, this study contributes to the body of L2 listening CALL research by investigating the value of social media comments to listening comprehension and introducing them to L2 learners, as well as CALL practitioners and researchers, as a new listening help option.

L2 Listening Help Options
The use of videos in preference to audio-only materials in L2 listening learning and teaching was founded on the belief that visual information such as body language, facial expressions, or the associated visuals would facilitate comprehension (Cross, 2018).Empirical CALL listening-related research was keen to invent and design the best tools that could provide additional support for the listening comprehension of L2 input.As a result, various help options (or modified input, as it is called by Ca´rdenas-Claros & Gruba, 2010), mostly text-based and including captions, subtitles, transcripts, scripts, glossed words, textual annotations, dictionary, cultural notes, grammar notes, visual annotations, and audio control, were exploited in L2 listening research and teaching (e.g., Ca´rdenas-Claros, 2020;Ca´rdenas-Claros & Gruba, 2014;Danan, 2016;Hegelheimer & Tower, 2004;Jones, 2008;Jones & Plass, 2002;Pujadas & Mun˜oz, 2019;Pujola`, 2002;Sydorenko, 2010;Vanderplank, 2019;Winke et al., 2010).
According to Cross (2017), help option studies essentially have centered on three themes: learner perceptions and experiences; comparisons of different conditions; and learner variables.Learners, in general, expressed positive attitudes toward textual help options.When compared to non-text help options, captions and transcripts, for example, were perceived as being the most useful tool for text comprehension (Ca´rdenas-Claros & Gruba, 2014;Jones, 2003).Regarding comparisons of different conditions, experimental studies (e.g.,Jones, 2003Jones, , 2007;;Jones & Plass, 2002) found a relatively positive impact from various textual help options on listening comprehension and vocabulary learning.Concerning learner variables, Cross confirms that proficiency level has been the only factor that has been investigated with regard to its influence on the use/non-use of textual help options, yet related studies (e.g., Grgurovic & Hegelheimer, 2007;Hegelheimer & Tower, 2004) have reported contradictory findings, making it difficult to draw any useful conclusions.

Non-Use of Help Options Issue
Researchers have noticed that L2 learners overlooked or ignored some potentially helpful options, leading them to identify the possible causes for this (e.g., Ca´rdenas-Claros & Gruba, 2014;Ca´rdenas-Claros, 2020;Pujola`, 2002;Jones, 2003;Grgurovic & Hegelheimer, 2007).Ca´rdenas-Claros and Gruba (2014), for instance, identified five variables that prompted or inhibited L2 listeners from using help options: relevance, challenge, familiarity, recovery, and compatibility.In brief, they determined when learners tend to use help options that are relevant to task management and completion, and they learned that the use of help options was broadly influenced by learners' views about language learning, comprehension failures, familiarity with help-option use, and the ease of help option interaction.
In a more recent study, Ca´rdenas-Claros (2020) examined how certain challenging textual features influenced learners' spontaneous selection and use of help options.They observed five problematic textual features, including four types of linguistic complexity (lexical, phonological, syntactic, and discourse) and speed of delivery.Four help options (audio/video control buttons, transcript, dictionary, and glossary) were found to be associated with specific textual challenges.Audio/video control was obviously linked to high speeds of delivery, transcripts were connected to fast speech problems as well as lexical and phonological complexity, and dictionary and glossary were related to lexical complexity.

Current Listening Help Option Research
The aforementioned help options have been and continue to be useful for offline and online listening materials alike.For instance, captions, subtitles, and audio/video control are very useful built-in features of YouTube.However, L2 listening CALL research seems to have been, and still is, confined to a limited set of traditional help options that were developed a long time ago, not only before Web 2.0 and social media but also before the internet, as can be seen in some recent studies (e.g., Ca´rdenas-Claros, 2020;Hsieh, 2020;Kam et al., 2020;Montero Perez et al., 2018;Pujadas & Mun˜oz, 2019;Teng, 2020;Vanderplank, 2019;Yang, 2020) in which some utilized YouTube videos.
There is no doubt about the benefit of those help options; nonetheless, such help options existed many years ago in offline listening materials that were provided, for example, on CDs.The point of this research is to send a wake-up call to CALL practitioners and researchers that a significant change in listening practice and extensive viewing has occurred.L2 learners today, with the Internet, the availability of free online videos and the affordability of smartphones, practice much of their L2 listening through YouTube and other streaming services (e.g., Blake, 2016;C xakmak & Erc xetin, 2018;Godwin-Jones, 2017, 2019;Hubbard, 2017;Wang & Chen, 2020).Thus, it is essential to expand CALL listening research and explore the unique help options that Web 2.0 and social media sites might provide, as well as what the social media comments can offer in terms of L2 listening comprehension.

YouTube Comments as a Help Option
Research on YouTube suggests that the commentary facility is probably the most prominent feature of social media sites such as YouTube (Benson, 2017;Dynel, 2014;Schultes et al., 2013;Shapiro & Park, 2018).YouTube has become more than a website for users to watch videos (Chen, 2020), since ''commenting is the key interactive writing space on the site'' (Barton & Lee, 2013, p. 39).YouTube comments have been described as ''vernacular literacies,'' implying that ''they are voluntary and self-generated, rather than being framed and valued by the needs of social institutions'' (Barton & Lee, 2012, p. 283).Nevertheless, an examination of the findings of discourse analysis studies (e.g., Cru, 2018;Dubovi & Tabak, 2020;Kulgemeyer & Peters, 2016;Madden et al., 2013;Shapiro & Park, 2018;Schultes et al., 2013) suggests that YouTube comments have the potential to become a listening support resource to all viewers, including L2 listeners.In particular, the results of two studies, that is, Madden et al. (2013) and Dubovi and Tabak (2020), indicate that comments are not only merely texts for users to express opinions or reactions to the videos they watch, rather useful texts to L2 listeners who can read comments to assist comprehension.
For instance, Madden et al. (2013) created a large corpus of 66,637 comments posted on 60 YouTube videos in order to provide a classification schema of YouTube comments.The authors found that YouTube comments provide factual information or an explanation about the video content or context, summarize or paraphrase something in the video content, directly quote something in the video, and sometimes may include blue-colored hyperlinks and a timestamp that direct the reader to a specific point in the video.Similarly, Dubovi and Tabak's (2020) analyzed a collection of 1,019 comments and found that comments made by viewers might negotiate meaning of the video contents (e.g., by providing restatements, by identifying inconsistencies, or by asking for clarification).Other users might choose to post comments to share or add opinions (including statements of observation, definition, or description).Incorporating the results of these studies, YouTube comments can serve a variety of rhetorical functions.For example, comments can summarize, restate, emphasize, paraphrase, elaborate on, or explain the video content by captioning the same words, phrases, or even whole sentences mentioned in the video.Such functions make comments extend or modify input of the video.By reading such comments, YouTube viewers can have their comprehension and predictions about the language consolidated and their affected understanding recovered.Given that CALL listening environments help ''expand the degree to which learners can self-regulate their own learning by providing them with more options for processing authentic material and scaffolding listening input'' (Danan, 2016, p. 5), an examination of Madden et al. and Dubovi and Tabak's findings reveals that YouTube comments, by being potentially helpful text, belong to the repertoire of the textual help options introduced in the literature.
It should not go without mention that YouTube comments have often been misconceived as haphazard, incoherent, and irrelevant input.However, that claim has been refuted by the conclusions of many empirical studies, such as Kulgemeyer and Peters (2016) and Schultes et al. (2013).The former study analyzed 1,365 comments in an attempt to find out which quality measures (numbers of likes, views, and comments) should be the indicator of the quality of 51 YouTube videos.They were able to label 392 comments as relevant comments, a proportion considered acceptable given that YouTube commentary facility is a ''largely unregulated space of discourse'' that accumulates ''massive amounts of comments'' (Androutsopoulos & Tereick, 2016, p. 361).Schultes et al. similarly found ''a substantial amount of comments that do not contain offensive statements and can be perceived as content carriers'' (p.667).

Advantages of YouTube Comments as a Help Option
Unlike some help options that learners tend to ignore for various reasons, YouTube comments have credibility as a favored help option.For example, the content of comments was found to be interesting, relaxing, and entertaining to L2 learners (Chen, 2020;Khan, 2017;Schultes et al., 2013).Also, YouTube comments have a userfriendly, simple, and intuitive design (Ca´rdenas-Claros & Gruba, 2013) that allows users to scroll down the comments section while keeping the video playing at the top of the screen, causing minimal or no distraction or interruption.For all their potential value for listening comprehension, YouTube comments have never been studied as a prospective help option.
As suggested by Madden et al. (2013)-''the benefit of a classification schema is that it allows researchers and practitioners to identify particular types or uses of comments and separate these out for analysis'' (p.711)-this experimental study aims to find out if YouTube comments can assist or increase L2 listening comprehension of videos.In addition, we will explore participants' perceptions toward three constructs of comments as a listening help option (relevance, value, and feasibility).Thus, this study seeks to provide answers to the following research questions: RQ1: Does reading YouTube comments result in better listening comprehension of L2 videos?RQ2: What are L2 learners' perceptions of YouTube comments as a listening help option?

Participants
The participants were male, Arabic-speaking Saudi EFL learners who were studying in a mandatory one-year English program at a Saudi public university where the researcher is a faculty member.The English program is designed for students seeking admission to health/medical or engineering colleges where English is the medium of teaching.The mission of the program is to improve students' skills in reading, writing, speaking, and listening in English.Using convenience sampling, 60 participants were drawn from two intermediate listening/ speaking classes (taught by the researcher) that met twice a week for 1 hour and 40 minutes each session.Students (M age = 18.45 years, SD = 0.59) had previously received formal English instruction for about 7 years in public schools.They are considered digital natives and have extended experience with computer-based listening.The two classes were randomly assigned to either the control (n = 31) or experimental group (n = 29).
The recruitment of intermediate-level learners for this experiment was done to ensure that summaries produced by experimental students were not solely based on listening competence if advanced-level learners were recruited.Thus, it was determined that less proficient learners were more likely to appreciate additional listening support (Yeldham, 2017) and would benefit from comments and, at the same time, have reasonable reading competence.

Materials
In this study, YouTube was chosen to represent social media firstly because it has many advantages over other social media platforms, namely, the main feature is posting videos and which it dominates social media platforms in this regard.Secondly, YouTube videos often have richer, more informative, and more educational content which means it should elicit richer and more informative comments.Thirdly, YouTube also allows the uploading of long videos (128 GB or 12 hours, whichever comes first), which can generate a large number of meaningful comments.Forth, YouTube allows public comments without any limitation in length meanwhile other platforms, such as Twitter, allow posting limited-length videos and short, 140-character comments/replies, or Snapchat, which does not allow public comments of any kind.
In order to achieve a more robust body of evidence for the effectiveness of YouTube comments on L2 listening comprehension, two YouTube videos were selected for the study.Video 1 (3'05'') was ''Woman's change of heart brings out her inner guardian angel'' by CBS Evening News, and Video 2 (3'03'') was ''Boston begins removing parking space savers'' from CBS This Morning.Video 1 received 755 comments and Video 2 received 2161 comments by the time the experiment was conducted.The linguistic and paralinguistic criteria set for the selection of videos were those proposed by Garza (1991) for the selection of L2 research video materials: situational appropriateness, grammatical and lexical complexity, and inherent interest value to university-level students.As long as comments were the core part of the study's investigation, a further criteria component was added: availability of relevant, rich comments.
Three experienced English language teachers were consulted who checked the videos and comments against these criteria.They reached a consensus that the two selected videos met the proposed criteria at a very acceptable level.For example, research shows that 90% lexical coverage (known words in a text) is adequate for listening/viewing comprehension of informal texts (Durbahn et al., 2020;Noreillie et al., 2018;Van Zeeland & Schmitt, 2013); the two videos reached 90% coverage at the 1,000-word level, as measured by Cobb's Lextutor's VocabProfile.This assured that the input would not be difficult for the participants to understand.
Regarding the relevance and richness of comments, it was also assured, for instance, that some comments contain direct quotes from the videos that are believed to support comprehension.Also, other types of comments were found to restate, emphasize, paraphrase, or elaborate on the video content.Samples of comments believed to provide potential assistance for listening comprehension can be found in Table 1.

Written Summary and Scoring
Written recall protocol provides more insightful information about listeners' comprehension of the text, as they are unmediated by question prompts (Markham et al., 2001;Yeldham, 2017).For each video, participants were required to produce one written summary that reflected as much of the video as they could remember; they were allowed to write in Arabic.Because written recall protocol is affected by memory constraints, students were allowed to take notes to assist their recall of what they have comprehended (Ca´rdenas-Claros & Gruba, 2014).Markham et al.'s (2001) scoring system of the written summaries was adopted, in which three raters counted the number of idea units, elaborations, and distortions produced by the participants.An idea unit refers to any thought or idea that occurred in the video.Hence, a sentence containing two clauses often includes two idea units.Elaborations are those accurate additional statements (not mentioned in the video) that are related to the video content, whereas distortions are inaccurate statements.Overall, the researchers achieved an interrater reliability of .93 for counting idea units and .88 for counting elaborations and distortions.Because the videos were short, very few elaborations and distortions were found that were excluded from the final analysis for that reason.

Questionnaire and Analysis
The experimental participants completed a brief questionnaire containing 11 Likert-scale items (see Appendix), tapping into students' opinions of comments as a listening help option over three constructs: relevance, effectiveness, and feasibility.The same questionnaire was administered twice (after each video).There were four items related to the relevance aspect, three items related to the effectiveness aspect, and four items related to the feasibility aspect.Two different scales were used: an agreement scale ranging from strongly disagree to strongly agree (items: 1, 2, 3, 5, 6, 8, and 9) and an amount scale ranging from very few to very many (items: 4, 7, 10, and 11).
The questionnaire agreement scale was formulated as follows: strongly disagree = 1, disagree = 2, neutral = 3, agree = 4, and strongly agree = 5; the amount scale was designed as follows: very few = 1, few = 2, average = 3, many = 4, and very many = 5.The analysis procedures for scales were as follows: from 1.00 to 2.60 means negative perceptions/few comments, 2.61 to 3.40 means neutral perceptions/average comments, and 3.41 to 5.00 means positive perceptions/many comments.The scores for the items addressing the same construct will be averaged.The Cronbach's alpha p coefficient will be reported, and then prominent results will be highlighted.Next, basic statistical data, namely mean and standard deviation, will be provided in a table.

Research Procedures
The study took place in the laboratory.The experimental participants watched the original videos on YouTube in order to access the comments.On the other hand, the control students watched downloaded video versions using video player software.Each student worked individually at a computer with a headset.Adhering to the tradition of playing a text at least twice (Field, 2008), participants in both groups were told to watch each video no more than twice.
Given that time constraints might affect student's interaction with help options-since they are more worried about finishing the task during the given time than in making use of the enhanced input (e.g., Cross, 2017)-the experimental students were allowed to read as many comments as they chose.They were asked to record approximately how many comments they read to inform our results.Immediately after viewing each video, students had 15 minutes to write a summary, and additionally, the experimental students also filled out the questionnaire.

Written Summaries
Data were analyzed using the Statistical Package for the Social Sciences (SPSS) software.To determine which statistical technique should be used, the collected data were tested for normality of the dependent variables with the Shapiro-Wilk normality test (p ..05).According to the test results, the scores obtained from the two tests' scales presented a normal distribution, so the parametric independent sample t-tests were used.
The T-tests concerning the number of idea units generated by the participants on the two videos yielded highly significant results, as demonstrated in Table 2. Specifically, data revealed that the participants who read comments (experimental) produced longer and more detailed written summaries than their peers (control) on both videos.The effect sizes (0.91 and 1.02) are considered large, according to Plonsky and Oswald (2014).
Although we were certain about the benefit of comments on listening comprehension, this massive gain difference between the groups was not expected; a further explanation of this unexpected result is provided in the Discussion.The average numbers of comments read by the experimental students were 17.07 (SD = 10.67) and 16.21 (SD = 9.86) for Video 1 and Video 2, respectively.

Questionnaires
A measure of the internal consistency reliability was performed by computing the most common measure of reliability, that is, Cronbach's alpha, which is based on a correlation matrix.Do¨rnyei (2010) suggests that a Cronbach a value of no less than .60 is necessary to have confidence that the items are measuring what they are intended to measure.The Cronbach's alpha coefficient of the total scale is .71on the first administration (Video 1) and .70 on the second administration (Video 2), which are both considered good.As mentioned earlier, the same questionnaire was administered twice.
Table 3 lists the descriptive statistics (means and standard deviations) for three constructs of the comments on each video: relevance, effectiveness, and feasibility.Regarding the relevance of comments, responses to the questionnaires showed that students' perceptions were positive, indicating that comments were relevant to the videos, as the means across the items are 3.70 and 3.77 on Video 1 and Video 2, respectively.With regard to the effectiveness of comments on comprehension, students' perceptions were neutral, yet closer to being positive than negative, as the means are 3.08 and 3.14.Regarding feasibility, the means (3.89) and (3.72) indicate that students believe that most comments were interesting and easy to read.According to the total mean (3.60), the overall perception of comments is positive.

Discussion
Our results suggest that comments in social media such as YouTube have a tangible, positive impact on L2 listening comprehension.However, these results are not surprising, given the rich content of comments.Data obtained from the written protocol instrument sufficiently provides some initial evidence for the benefit of YouTube comments on listening comprehension, as the experimental group generated significantly more unit ideas in both summaries.This finding aligns with the general findings of previous help option studies, which suggest that any kind of modified input of L2 listening materials, be it captions, subtitles, dictionaries, and glosses, is beneficial to L2 listening comprehension in one way or another.The rationale for the provision of a variety of options is to offer learners alternative ways of exploiting aural texts, yet help options differ largely on the quality of help they offer, such as transcript vs. cultural notes (Ca´rdenas-Claros & Gruba, 2013).That being said, it should not be assumed that comments are as effective as carefully created help options such as captions or transcripts.Rather, the best way to look at comments as a help option is to think of the quality of such help options as a continuum, and comments probably fall in the middle.
Although we expected to be a somewhat obvious influence of comments on listening comprehension, the massive gain, as our result showed, was surprising.A possible explanation for this is that reading comments allowed learners to spend more time with the overall content of the video (i.e., watching a video and reading comments on it), which presumably helped them reflect more on the video and get more engaged; consequently, comprehension increased.
The rhetorical functions of comments emphasized by discourse analysis (Dubovi & Tabak, 2020;Madden et al., 2013) seem to have a huge influence on listening comprehension.We assume that the comments posted on both videos were very effective and informative and, by reading them, experimental students were able to increase or adjust their comprehension, resulting in significantly better and richer summaries.By being exposed to such input (i.e., comments that extend, restate, emphasize, paraphrase, and elaborate on the video content by quoting words and phrases), students were able to consolidate what they watched if unsure, or correct themselves about what they thought they had heard.That is confirmed by Khan (2017): ''YouTube users not only seek information through watching videos, but also through browsing through comments which offer valuable information to users' ' (p. 243).
Also, our participants were allowed to read as many comments as they chose, and our results indicate that they read a significant number of comments; approximately 17 is the average number of comments read for each video.Researchers suggest that the design of help options should help learners stay on task and minimize potential distractions (e.g., Ca´rdenas-Claros & Gruba, 2013); the large number of comments read by participants suggests that YouTube comments enjoy a userfriendly, simple, and intuitive design, given that YouTube allows viewers to browse comments while the video plays at the top.
In parallel with Schultes et al.'s (2013) conclusion, ''the main characteristics of a comment that creates added value for the users: fair, substantial and relevant for the underlying video are some of them'' (p.672), we assume that listening comprehension can be best assisted when a YouTube video accumulates a rich repertoire of relevant, meaningful comments touching on multiple parts of the video.However, we do not know exactly how comprehension was assisted, given that a wide range of comment types was posted on the two videos, nor do we know which comments were more beneficial.However, when we consider the biggest listening problem to L2 listeners, which is the lexical segmentation-the ability to identify where each word begins and ends (Field, 2003)-we assume that quote-type comments are very useful, as quotations allow learners to see word boundaries.
Finally, the perceptions of comments as a listening help option were positive.Comments on the two videos were perceived as relevant and feasible.On the other

Pedagogical Implications
Captions were originally developed for the deaf and hard-of-hearing in the 1980s, but later, CALL practitioners and researchers discovered the benefits of captions for L2 learners.Likewise, the comments facility was intended to be a section of text for users ''to interact and be in touch with other, often unknown, YouTubers'' (Bou-Franch et al., 2012, p. 502), but our results suggest that social media comments are also beneficial for L2 listening comprehension.That being said, and because ''language learners may need guidance in navigating the options available to them'' (Vandergrift & Goh, 2012, p. 232), L2 learners need to be familiar with the affordances of comments, given that some frequent L2 learners/users of YouTube do not pay any attention to comments (Chen, 2020).With proper training that focuses on the appealing features of comments, such as reading enjoyment and flexibility (i.e., checking comments while watching), learners should be able to recognize the fundamental benefits of comments.This is confirmed by Chen (2020), ''in the process of engaging in making sense of online texts, students also discovered the features of online language, such as informality, irony, and common uses of abbreviations and capitalizations'' (p.445).With the growing interest in informal, self-regulated learning, reading YouTube comments is particularly suited for autonomous and informal L2 listening.

Limitations of the Study and Future Research
This study has several limitations, and we will be highlighting the most important ones.First, only short videos were used, and both were in a similar category, that is, both videos are considered news or story-telling videos.
With longer videos or those in different categories, the potential for comments to impact comprehension might be affected.Second, the difficulty level of the videos was maximally minimized with the proposed criteria, that is, no issues in the situational appropriateness of the videos and grammatical and lexical complexity were minimized.Thus, it is not known how reading comments could assist comprehension if more difficult videos were used.Third, only one type of measure of comprehension was used, that is, a written summary.Other types of comprehension measures such as multiple-choice or fill-in-the-blank tests might yield different results.Forth, only male participants were recruited as they were the only students available to the researcher.We cannot generalize our results to all English language learners in Saudi Arabia.
Future studies might consider examining other social media platforms that allow both videos and comments (e.g., Facebook), different characteristics of videos (different languages, longer duration, different content, and different accents), learners of different demographic backgrounds (different L1s, ages, gender, levels of education, and second language context), and different types of comments (L1 comments and paragraph-length comments).This study examined the benefit of L2 comments on the listening comprehension of L2 videos.However, it is not uncommon that a YouTube video receives many comments written in different languages, given that some YouTube channels are watched by a considerable number of non-native speakers.The point is that the value of the comment section is not limited to comments written in the video's original language; comments written in the reader's language may, in fact, be even more beneficial.Finally, keeping pace with captioning research, more study is certainly warranted to determine the effectiveness of reading comments on vocabulary acquisition.

Conclusion
In this study, we explored the impact of social media comments on understanding online videos.We found that reading a rich repertoire of meaningful, relevant comments facilitated EFL listeners' comprehension of YouTube videos.Participants who read comments wrote better summaries of the content of the videos than those students who only watched the videos.In addition to the benefits of comments on listening comprehension, students had positive perceptions of comments as a listening help option.

Questionnaire
Directions: There are two types of statements in this questionnaire; for some statements, you need to choose your degree of agreement, and for others, you need to indicate your evaluation of an amount.

Table 1 .
Comment Samples From Video 1 and Video 2. Rap game bob barker ''I have a FRIEND named Ginger, she's on her way.''How sweet and pure is that!One Bright Star ''Life is messy, but if you're gonna love other people, you're gonna have to step into their mess'' I love this quote.

Table 2 .
Means, Standard Deviations, and Summaries of T-Tests + Effect Sizes for Written Summaries (Unit Ideas).

Table 3 .
Perceptions of Comments' Relevance, Effectiveness, and Feasibility.