Voices in Videos: How YouTube Is Used in #BLM and #StopAAPIHate Movements

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript presents quality research on a meaningful topic. The analysis was somewhat fair and comprehensive. Some comments are as follows:

Line 32 – the research gap is not clear

RQ2 – could be break down into two subsections. Could have mention Kitts work briefly before the RQs so that people know the logic behind.

The paragraph for line 51-61 could have come too soon.

Section 3 > Is it referring to a pilot study?

Section 4 > why even sample is needed? If so shouldn’t the authors ensure equal number of video clips after excluding unrelated content? How to sample 298 video clips for BLM from 1250?

Table 1 > the title of the table could be improved. Also, should it be “storytelling” instead of “story”?

Is the paragraph for line 249-258 presenting the same information with the paragraph for line 51-61. Is it the Youtuber or Public figures constituting the most video content creator?

Line 272> just wonder if certain type of identify leads to more of certain information. News reporter would be news/ artist performing arts…

Wrong spelling in line 483 and 485?

Line 493-496 > why youtube should offer support to religious figures / ordinary people? In what form?

Line 532 > further concrete suggestion need for first implication.

The reference list need to go through final editing. Some references shows the doi twice. Somehow when I was trying to see Kitts work, it shows a different year of publication. Please have a check.

Author Response

Line 32 – the research gap is not clear

> Thank you for bringing attention to this part, we reworded the last paragraph in the introduction to make research gap clearer

RQ2 – could be break down into two subsections. Could have mention Kitts work briefly before the RQs so that people know the logic behind.

> We agreed with this suggestion and broke up RQ2 into 2 sections. We also added a line in the first paragraph briefly and explain how Kitts guides the identification of the RQs.

The paragraph for line 51-61 could have come too soon.

> We updated this section to provide a more introduction of the findings with a more general description rather than introducing the numbers.

Section 3 > Is it referring to a pilot study?

> It is not a pilot study. This refers to the qualitative analysis of the sample to derive the codebook. We have corrected this session in the first two paragraphs of 3. Data Collection and Analysis Method.

Section 4 > why even sample is needed? If so shouldn’t the authors ensure equal number of video clips after excluding unrelated content? How to sample 298 video clips for BLM from 1250?

> The decision to balance the datasets was made after identifying an unequal representation between the two movements during initial data collection. Given that our data processing included manual filtering, we balanced the datasets prior to this step to reduce the amount of unnecessary manual checking. This approach also helped minimize potential biases toward the #BLM movement. This information is updated in the  first two paragraphs of 3. Data Collection and Analysis Method.

Table 1 > the title of the table could be improved. Also, should it be “storytelling” instead of “story”?

> Thank you for pointing both of those out, we changed the title of the table to “Description of Themes and Subthemes”,  and fixed the category to be “storytelling” instead of “story”

Is the paragraph for line 249-258 presenting the same information with the paragraph for line 51-61. Is it the Youtuber or Public figures constituting the most video content creator?

> As per an earlier comment, we updated this section to provide a more introduction of the findings with a more general description rather than introducing the numbers. Public figures and YouTubers (renamed to vloggers according to the review) were the most common video creators with similar presentations, as pointed out in this comment.

Line 272> just wonder if certain type of identify leads to more of certain information. News reporter would be news/ artist performing arts…

> We had conducted an association analysis between information types and creator identities. However, this analysis did not yield any statistically significant associations; thus, we chose not to report these non-significant results.

Wrong spelling in line 483 and 485?

> We thank the reviewer for noticing this wrong spelling, and we corrected the typos

Line 493-496 > why youtube should offer support to religious figures / ordinary people? In what form?

>Our intention was to highlight the unique participation by these groups, as viewers who typically watch their content might not anticipate videos related to racial movements. We have reworded paragraph 3 in Section 5.1 to clarify this point. 

Line 532 > further concrete suggestion need for first implication.

> We agree with this sentiment and reworded the sentence by adding more implications for video searchability and identity concealment methods. See the change in 5.2 paragraph 4

The reference list need to go through final editing. Some references shows the doi twice. Somehow when I was trying to see Kitts work, it shows a different year of publication. Please have a check.

> The reference issues have been addressed.

Reviewer 2 Report

Comments and Suggestions for Authors

The paper focuses on the role of video sharing platforms as a medium of influence for social justice movements, assessing their real strategic and operational potential. The novelty and strength of this work is given by the fact that it chooses the specific role of YouTube and its platform culture in supporting social justice movements. This is a topic that is actually understudied in the social sciences. The author therefore addresses this gap in research by analyzing video content related to two important online social justice movements: #BLM and #StopAAPIHate. From a methodological and analytical point of view, the author works on a dataset of 489 videos containing these hashtags. In the second part of the study, thematic categories were developed to explore the identity of video creators, the information transmitted in the videos, the storytelling techniques and promotional features. The results appear solid and convincing as they suggest that public figures, YouTubers and journalists are the most common groups sharing movement videos. The main use of YouTube is to spread knowledge of the movement and stories of discrimination. The author also adds that most of the video creators only use videos to share their social media accounts and do not actively use the platform features such as live streaming, merchandising, donations or sponsorships to support the movement. Such information is of high interest to the journal and its readers and of course to strengthen research in human and social sciences. For these reasons, we ask you to accept this publication.

Author Response

We sincerely thank R2 for recognizing the value of this work.

Reviewer 3 Report

Comments and Suggestions for Authors

1- it's important to explain the data collection technique better in the abstract; (p.1)

2-on line 18, page 3, you need a more credible source to substantiate questions about youtube 

3- in terms of theoretical research, it is necessary to better clarify the differences between new and old social movements and also the new digital social movements (p.4)

4- clarificate line 192-193 (p.5) the lines need to be clarified; a sentence that is too assertive to be stated without further development  

5- the method should have been made clearer at certain stages (pages 5 and 6) 

6- Table 1, with the categories, is interesting, but there are two of them that are not mutually excluding: public figures and youtubers; artists can also be conflicting. How can they be distinguished from each other? aren't there famous youtubers? aren't there famous artists and youtubers at the same time?

7- as a result of the previous question, the discussion of results and the conclusion should also be changed because the results are dependent on this change of categories (pages 13-15).

 

Author Response

1- it's important to explain the data collection technique better in the abstract; (p.1)

> Thank you for noting this, we added a mention of videos being crawled through youtube data api in the abstract

2-on line 18, page 3, you need a more credible source to substantiate questions about youtube 

> We added more references regarding social movements on YouTube.

3- in terms of theoretical research, it is necessary to better clarify the differences between new and old social movements and also the new digital social movements (p.4)

> At the end of section 2.2, we added an explanation about how YouTube is different from traditional social movements.

4- clarificate line 192-193 (p.5) the lines need to be clarified; a sentence that is too assertive to be stated without further development  

> We thank the reviewer for noting this point, and we have amended the statement and further developed Section 2.3 paragraph 3 by adding the knowledge gap.

5- the method should have been made clearer at certain stages (pages 5 and 6) 

> We thank both this reviewer and Reviewer 4 for this suggestion, and have addressed this comment by clarify the data analysis steps in Section 3 paragraph 1 and 2.

6- Table 1, with the categories, is interesting, but there are two of them that are not mutually excluding: public figures and youtubers; artists can also be conflicting. How can they be distinguished from each other? aren't there famous youtubers? aren't there famous artists and youtubers at the same time?

> We acknowledge the confusion caused by the non-mutually exclusive identity labels. The category "YouTuber" was not intended to imply that artists or public figures could not also be YouTubers. Rather, we used this label to refer specifically to vloggers -- creators who regularly post videos as a way to share personal experiences or thoughts, often in the form of commentary or everyday life content, without a clear affiliation to an authoritative role or specialized artistic skill. To clarify this distinction, we have renamed the category to "vloggers" and added brief descriptions explaining how each identity was differentiated.

7- as a result of the previous question, the discussion of results and the conclusion should also be changed because the results are dependent on this change of categories (pages 13-15).

> Such changes have also be reflected in the discussion section.

Reviewer 4 Report

Comments and Suggestions for Authors

Thank you for your hard work in preparing this manuscript. It covers an intriguing research topic. I am confident that addressing the following suggestions will help improve the quality of your paper. Thank you.

 

1. Strengthen any claim that isn’t directly obvious from your data by providing either data evidence or a literature citation. For example, the discussion suggests that YouTube’s interactive features "may not be suitable or effective" for engaging viewers​. To make this convincing, add reasoning or evidence: perhaps cite studies or user surveys that found low uptake of these features, or clarify that this is a hypothesis based on your observation of creator behavior. Any interpretive leap should be backed by a source or clearly framed as a suggestion for why the trend exists, rather than a definitive conclusion.
   
   
2. Be cautious about introducing new assertions that you haven’t examined directly. For instance, the statement that the monetary aspect of features raises legal and ethical is thought-provoking, but the paper does not explore legality or ethics in depth. Either support this point with a reference to relevant debates or omit it to keep the argument focused. Stick to conclusions that your data can support, and if you want to raise broader issues (like ethics of fundraising via YouTube), frame them as open questions or future work rather than findings.
   
   
3. Ensure each part of the argument flows logically. For example, when transitioning from results to implications, explicitly connect the dots: “Because we found X, this suggests Y.” In the discussion sections (6.1–6.3), make sure each paragraph clearly ties a result from your study to a broader interpretation or implication. If readers can see the step-by-step reasoning, your conclusions will be more persuasive.
   
   
4. Since the study covers two movements, strengthen the comparative aspect of your argument. Highlight any interesting differences or similarities between #BLM and #StopAAPIHate findings and argue why those might exist. For instance, if YouTubers make up a larger share of StopAAPIHate content than BLM content, discuss what that means – perhaps the StopAAPIHate movement relied more on independent content creators due to different media attention. Making these comparisons explicit will reinforce the paper’s insights and show the nuance in your argument.
   
   
5. The description of data collection and sampling needs to be clearer and more consistent. In the abstract you mention 489 videos analyzed, but in the methods you describe collecting 1,548 videos and using a balanced sample of 596. Later, it’s explained that the final analytic sample was 489 videos (after removing overlaps: 266 #BLM, 210 #StopAAPIHate, 13 both. This process is confusing as written. To improve clarity, explicitly outline the steps: e.g., “We retrieved 1,548 videos (1,250 #BLM and 298 #StopAAPIHate). To balance the two movements, we took 298 from each, then removed duplicates and videos unrelated to the movements, yielding a final sample of 489 unique videos.” By clearly explaining each step and why decisions were made (like balancing the sample), readers can trust and replicate your methodology.
   
   
6. Provide details about when and how the videos were collected. For instance, mention the date range or period during which the YouTube Data API was used to crawl videos, and whether any filters (language, region, upload date) were applied. This is important for reproducibility. If videos were identified by searching the hashtags, clarify whether the hashtag had to be in the title, description, or anywhere in metadata. Being specific about criteria (e.g., “videos were included if they had #BLM or #StopAAPIHate in the title or description, and were uploaded between January 2020 and December 2021”) would strengthen the methodology section.
   
   
7. Some research choices should be briefly justified to preempt questions. For example, explain why it was important to use an even number of videos from each movement (to enable comparison without one dominating the sample). Also, note why thematic analysis was suitable for this study – perhaps because you were exploring categories inductively. If any videos were excluded as "irrelevant categories" during initial coding​, give examples of what those were (to reassure readers you didn’t bias the sample). Such justifications make your method rationale transparent.

Comments on the Quality of English Language

1. Ensure that all sentences convey their point clearly without unnecessary complexity. For example, define specialized terms like "parasocial interactions" upon first use so that all readers understand them. Avoid jargon where possible or explain it in simple terms.
   
   
2. There are a few minor errors that can confuse readers, such as "addressess" instead of "addresses"​ and "intnroducing"/"uplaod" instead of "introducing"/"upload. Careful proofreading or copy-editing will catch these issues. Correcting such errors will improve the professionalism and clarity of the text.
   
   
3. Use terms and hashtags consistently throughout the paper. For instance, if the movement is introduced as Stop AAPI Hate (StopAAPIHate) in the introduction, consistently use one format. Similarly, maintain consistency in using or not using the "#" when referring to #BLM and #StopAAPIHate in the prose to avoid confusion.
   
   
4. When making a point, consider adding a brief example to illustrate it in plain language. The paper already does this at times (e.g. describing specific YouTube videos as examples of themes), which is great for clarity. Expanding this practice will help readers grasp abstract concepts. For instance, when you mention different storytelling techniques, you might briefly explain each with an example scenario so readers immediately understand what an "artistic" vs. "conversational" video looks like.

Author Response

Strengthen any claim that isn’t directly obvious from your data by providing either data evidence or a literature citation. For example, the discussion suggests that YouTube’s interactive features "may not be suitable or effective" for engaging viewers​. To make this convincing, add reasoning or evidence: perhaps cite studies or user surveys that found low uptake of these features, or clarify that this is a hypothesis based on your observation of creator behavior. Any interpretive leap should be backed by a source or clearly framed as a suggestion for why the trend exists, rather than a definitive conclusion.

> We thank the reviewer for this detailed comment, and have provided more references in the discussion section to backup claims made for describing the implications.

Be cautious about introducing new assertions that you haven’t examined directly. For instance, the statement that the monetary aspect of features raises legal and ethical is thought-provoking, but the paper does not explore legality or ethics in depth. Either support this point with a reference to relevant debates or omit it to keep the argument focused. Stick to conclusions that your data can support, and if you want to raise broader issues (like ethics of fundraising via YouTube), frame them as open questions or future work rather than findings.

> We thank the reviewer for raising this point and have incorporated it as a future research question rather than presenting it as a finding. Additionally, we have toned down certain claims—such as those related to fundraising and legal considerations—in the discussion section 5.3 to focus more on resource mobilization.

Ensure each part of the argument flows logically. For example, when transitioning from results to implications, explicitly connect the dots: “Because we found X, this suggests Y.” In the discussion sections (6.1–6.3), make sure each paragraph clearly ties a result from your study to a broader interpretation or implication. If readers can see the step-by-step reasoning, your conclusions will be more persuasive.

> We agree with this comment, and reworked the Discussion section, to make the arguments flow better

Since the study covers two movements, strengthen the comparative aspect of your argument. Highlight any interesting differences or similarities between #BLM and #StopAAPIHate findings and argue why those might exist. For instance, if YouTubers make up a larger share of StopAAPIHate content than BLM content, discuss what that means – perhaps the StopAAPIHate movement relied more on independent content creators due to different media attention. Making these comparisons explicit will reinforce the paper’s insights and show the nuance in your argument.

>We have added more comparisons between the two movements in Section 6.2. This includes a new discussion in Section 5.2, paragraph 1, addressing the timing differences between the movements, and in paragraph 2, discussing the theme of denouncement.

The description of data collection and sampling needs to be clearer and more consistent. In the abstract you mention 489 videos analyzed, but in the methods you describe collecting 1,548 videos and using a balanced sample of 596. Later, it’s explained that the final analytic sample was 489 videos (after removing overlaps: 266 #BLM, 210 #StopAAPIHate, 13 both. This process is confusing as written. To improve clarity, explicitly outline the steps: e.g., “We retrieved 1,548 videos (1,250 #BLM and 298 #StopAAPIHate). To balance the two movements, we took 298 from each, then removed duplicates and videos unrelated to the movements, yielding a final sample of 489 unique videos.” By clearly explaining each step and why decisions were made (like balancing the sample), readers can trust and replicate your methodology.

Provide details about when and how the videos were collected. For instance, mention the date range or period during which the YouTube Data API was used to crawl videos, and whether any filters (language, region, upload date) were applied. This is important for reproducibility. If videos were identified by searching the hashtags, clarify whether the hashtag had to be in the title, description, or anywhere in metadata. Being specific about criteria (e.g., “videos were included if they had #BLM or #StopAAPIHate in the title or description, and were uploaded between January 2020 and December 2021”) would strengthen the methodology section.

> We thank both Reviewer 3 and 4 for noting this, and have amended the Methods section. The decision to balance the datasets was made after identifying an unequal representation between the two movements during initial data collection. Given that our data processing included manual filtering, we balanced the datasets prior to this step to reduce the amount of unnecessary manual checking. This approach also helped minimize potential biases toward the #BLM movement. This information is updated in the  first two paragraphs of 3. Data Collection and Analysis Method.

> We also added information about when the data was collected and the time span of the initial data.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for taking in my comments.

Reviewer 4 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this manuscript, "Voices in Videos: How YouTube is Used in #BLM and #StopAAPIHate Movements." The authors have done commendable work in conducting a detailed thematic analysis and clearly presenting their findings. The topic is timely and important, shedding valuable light on how video-sharing platforms shape contemporary social justice movements.

Overall, this is a valuable contribution to scholarship on digital activism and online communities. Thank you again for your hard work—I look forward to seeing the final published version.