Session 3 assignments

[simonsimson]

This is the assignment for session 3 taking place on 19 Nov 2020.

1. Read the article “Mapping Controversies with Social Media: The Case for Symmetry” by Noortje Marres and David Moats Open access: https://journals.sagepub.com/doi/10.1177/2056305115604176

2. Drawing from your material in assignment 1, describe media practices fueling controversy in the debate and experiment with different modes of enquiry (precautionary, affirmative, empiricist) according to (Marres and Moats 2015).

Please submit the assignment as a reply to this issue. The exercise can either be tackled alone or as a team of two. Scope: 150 words (alone) or 250 - 300 words (as a group of two students) Due date: 18 Nov 12h00

PS: If you want to dive deeper into conceptualizations of media practices, I recommend this article: Pentzold, Christian. „Conceptualizing the Doings and Sayings of Media Practices: Expressive Performance, Communicative Understanding, and Epistemic Discourse“, 2020, 21. https://www.researchgate.net/publication/341553352_Conceptualizing_the_Doings_and_Sayings_of_Media_Practices_Expressive_Performance_Communicative_Understanding_and_Epistemic_Discourse We will also discuss these more in detail in a later stage of the seminar.

[satorus]

Video title: Naomi Seibt and Her Journey to Climate Realism Link: https://www.youtube.com/watch?v=v8dXpe1Pp6Q

Looking at the comments below the video, one practice that fueled the debate and provoked a rather large discussion (in compraison to other comment chains) was provoking with a statement against the content shown in the video. Here, the comment author asked that we should get real scientists in front of the camera instead of kids, directly criticizing the appearance of Naomi Seibt in a science related video channel in a provoking way (of calling her "just a kid"). This sparked a debate about why scientists shouldn't or do not want to be shown in front of a camera in social media. From a precautious point of view, this is not really a issue driven controversy as most of the discussion revolves around a "person cult", often found on youtube and may be a platform specific phenomen, so it should probably be disregarded for analyzing the issue based controversies. When looking from an affirmative point, having these "person cults" in social media is also a part of the issue at hand, as today many scientific controversies revolve around persons many people flock too, listen to and defend vehemently. This behaviour is fueled by social media platform characteristics but can not be disregarded when looking at the issue based controversies. Having a look from the empricist mode whe could start with removing comments mostly regarding the "persona cult" and analyze if comments fueling the issue based controversy remain. This seems to be not the case, as a large (if not most) commments are just about these persons in question and when eliminating all of them, no substantial discussion about the issue at hand (the climate change) remains.

[SabinaP17]

Video title: Neil deGrasse Tyson scolds cherry picking climate science Link: https://www.youtube.com/watch?v=y1MZ8U8C9c8 Views: 1,516,153 Comments: 13,576 Published: 17.09.2017

One of the main controversy issues in the comments was whether the subject of climate change should be addressed in the US “politically” or not. This debate topic was triggered by a user’s comment on the video, in which he stated that “making climate change 'political' was the worst thing we have done to handle the issue”. The opinion of other users on this matter was rather divergent. Analyzing the issue at hand using a precautionary approach, the positions of the users within this debate could be mapped as following: some users argued that climate change and the discussion around it were from the beginning related to politics and are therefore not true or not proven (“climate change was political to begin and we now trying to make it about science. But, the politics seem to keep getting in the way.”; “It's politicized nonsense. The current fear mongering is created by the political left because they fear the science.”, etc.). Other users, on the other hand, pointed out that climate change was scientifically proven and future policy should be therefore built on the consensus between the scientists, rather than on misleading data (“there is nothing to combat listen to what scientists say and build policy on what is the consensus, not on what you want to believe based on saving money.”; “97% of scientists believe climate change is due to human activity, that's the consensus”, etc.). From an affirmative approach, digital bias should also be taken into account when analyzing the data. Thus, there were many comments which focused exclusively on “conspiracy theories” regarding politics, world order, and so, which might be regarded as a phenomenon specific for the platform. Also, the “Like” and “Dislike” buttons on Youtube might have had an influence on the mediation of this controversial debate. Analyzing the data using an empirical approach, the inclusion of “conspiracy theories” in some comments, as well as the number of “likes” or “dislikes” that a comment had, might have played a role in keeping the intensity of the debate at a high level or might have served as a basis for supporting or dismissing certain comments. The issue debated had as effect the formation of two groups, so to say, one supporting science and scientific proofs and another arguing that it is all political and “fake”. Both the discussed issue and the Youtube technologies lead to a high divergence between the two groups mentioned earlier, which paved the ground for verbal aggression, name-calling, insults, etc. in later comments.

[Aylin-00]

Video Title: Senator Rand Paul Talks COVID Response Link: https://www.youtube.com/watch?v=iAJypmx0d5U Views: 10.478 Comments: 152 Published: 23.07.2020

Summary: The video posted by the Republican US Senator Rand Paul from Kentucky deals with the reaction of different states to the Corona Virus. He criticizes government enforced lockdowns especially in New York, claiming that herd immunity is the best way to address the isssue of Covid19 until a vaccine is ready. His argument for that is that lockdowns do not show any positive effect for infection/death rates.

Precautionary approach: I could not detect any bot behaviour or spam, maybe because the video was not that popular and therefore less interesting for hijacking. The main controversy was about the question if Senator Rand Paul is speaking truth to the public by critizing the lockdowns, praising herd immunity and a vaccine or if he is also part of the "Deep State" and therefore a lier.

Especially people believing in a Covid19-Conspiracy attacked other users who seemed to believe that the virus exist and that it was of natural origin but still opposed any government action.

• "You mean “released”. And yeah listen to Rand Paul pushing propaganda"
• "SammySay because of fake news. Rand is fake news too. Listen to him say that we need contact tracing at 4:00"

Some kind of user interaction took place when the conspiracy theorists tried to explain why the government was enforcing lockdowns. Common arguments were that the Marxists wanted to secretly take over:

• "To destroy the economy and nationalize it."
• They have to weaken the economy to the point where when they burst the bubble, the maximum # of dependents will look to the government to for their food housing etc.. Who will fight a tyrannical government, upon which they rely to keep their families from starving to death?"

Affirmative approach: Discussion often evolved around comments that had many likes. Also these comments were ranked higher by the YouTube Algorithm so that more people saw and reacted to it by commenting. Therefore this process enhaced itself. While users agreeing with Rand Paul seemed to like other praising comments without leaving a comment, people opposing him generally replied to the comments.

Empirical approach: Most of the users wrote only one comment, while others were more active and commented on a lot of other post. These users often commented on high ranked, liked comments. Also they seemed to use the same statements over and over again. One example is the user "Fascism is Right Wing, and God saw that it was good". He often points to Senator Rand Pauls agreement with contract tracing.

• "Yeah they both think we need contact tracing... listen to Rand at 4:00 say that contact tracing is a good idea."

Curbing popular discussions in ones own favor also seemed to take place. If someone brings up a topic unrelated to the popular comment, other users generally jump onto the new topic. This can be used to bring something completly different to attention.

[mrtobie]

Video title: Bill Nye Takes On Climate Change Deniers | Machh | NBC News Link: https://www.youtube.com/watch?v=S7Lu-R4EwQw

In the comments there is one user “Rick TD” that is fueling the debate quit harshly. He has a quite rough tone of voice and claims believing in climate change is “ignorant”. He also states that the whole climate change debate is a “scam”.

Precautionary Approach:

Since there are no obvious bots there is nothing to cut out. There is one user in most of the discussions where he is the only quite active representative of the man-made climate change deniers. But since he doesn't seem to be a bot, he should not be cut out of the data. But on the other hand one man alone should ne be treated as the whole opposition. A precautionary approach would not be purposeful when analyzing the debate in general.

Affirmative Approach:

An affirmative approach by algorithms could be very interesting. Nearly all of the users that reply to a comment and therefore participate in a debate answer directly to somebody by tagging them with an “@”. This data could be used to see if there is a deeper conversation between two/three or more parties. From looking at it, it seems that if a user responds a certain interlocutor, they write quite often answers to each other so that could be interpreted as a somewhat deeper converstation. Meanwhile if others just respond once to state their point, they often don't tag an adressee.

Empirical Approach:

It seems that the side of the debate that deny climate change often use other YouTube-URLs as sources to back up their statement while the other side of the argument uses mostly other sites with often a .gov or .org domain.

[adrianapintod]

Video title: Naomi Seibt Joins One America News Network Link: https://www.youtube.com/watch?v=rcHCretMIZU

One of the most debated topics in the video comments is about the intelligence and leadership of the girl Naomi.

Precautionary mode: Taking into account that the video does not have many views and therefore not much interaction I could not identify any "platform bias" like bots publishing generic content that will affect the empirical viability of the analysis. There were also no comments that were not related to the issue.

Affirmative mode: According to this approach, the participation of digital platforms has a more positive appreciation. In this case, youtube shows the top comments of the video, organizing them by the number of responses and the likes obtained, which could facilitate the analysis by focusing on the issue. Some users made comments where a debate was started while others responded to comments from other users generating a thread that extended the debate.

Empirical mode: From the empirical point of view, the interaction of users with certain videos with likes and replies as well as the algorithm for sorting YouTube comments can play an important role in the continuity and relevance of the debate.

[Cloudz333]

Video title: Climate "realists" want U.S. to stop spending money on climate change Link: https://www.youtube.com/watch?v=bKd30j1HX48 Views: 8370 Comments: 73 Published: 23 Apr 2017 Summary: In this CBS Evening News report, a journalist interviews the former president and CEO of the Heartland Institute (a conservative think tank) Joseph Lee Bast, who defines himself a "climate realist". He supports the thesis that the efforts to stop global warming or slowing it down are way disproportional to what the science suggests it would be necessary, and thinks that it's not worth investing money in climate change.

---

Reviewing the comments section for a second time, I was able to focus more on the popularity of each type of debates, and reconfirm that the most popular discussion revolves around whether global warming is caused by humans or not.

By adopting a precautionary approach I could think to filter out non meaningful data points that are not related with the debates taken into consideration. In this regard I could identify a cluster of comments that mainly focuses on whether the interviewee has the necessary competences to make certain statements or not. For instance: “Who gave this man a degree”, or “Not a scientist - OMG are you freaking kidding me?! Thanks for the fake news.”.
Since the video contains a limited number of comments, I was unable to identify any type of bot related actions. On the other hand, I could Identify one meaningless comment that can be considered as spam, in which an user wrote: "1".

In order to analyze the dispute using an affirmative approach, it is possible to use some interesting features made available by youtube such as the number of likes, the number of comments and the comment sorting (e.g. by popularity). Through these functions I could identify immediately that the two main clusters of commenters, (those who believe in human’s contribution to global warming and those who don't) are pretty equally distributed across this video and trigger pretty the same amount of reactions.

However, by adopting an empirical approach in order to investigate which effects belong to the media technology, it turned out that the comments ranking algorithm of youtube can play an important role. Here, by taking as reference an unofficial source, it seems that the algorithm also takes into consideration who has posted a certain comment, rewarding users who have a higher average like/dislike ratio on total comments.

[iraari]

Video title: Tucker vs. Bill Nye the Science Guy Link: https://www.youtube.com/watch?v=qN5L2q6hfWo Views: 4,475,831 Comments: 50,765 Published: 27 Feb 2017

Summary: Bill Nye the Science Guy is invited to Fox News for a discussion about climate change and tries to convince Tucker Carlson, who hosts the nightly political talk show, that climate change is caused by human activity. The discussion escalates into an argument in which the parties interrupt each other.

This video clearly shows that climate change is a problem not only about epistemic issues but political ones as well. Most of the comments evaluate the behavior of the interlocutors and their political views and, depending on this, support or disparage one of them.

Therefore, to focus on the topic discussion I would exclude the comments containing quotes (“...”) from the video, that people use to highlight their opinion about the host and the guest, as well as all the comments with the names of both Carlson and Nye (Precautionary mode).

There is a pattern that comments on the topic are longer and often someone replies to them. It’s possible to distinguish this kind of comments thanks to the "View replies" option. Sorting comments by popularity shows the main opinions (that will be repeated in other words) and the wittiest comments (Affirmative mode).

In the default mode the top comments are shown. As mentioned above, they do not facilitate an effective subject discussion, but rather concentrate on the judgments on Tucker and Nye. I think these comments contribute to distraction from the climate change issue and encourage people to join the polemic against the host or the guest (Empirical mode).

[Rahaf66]

Video Title: Climate Change Activists vs Skeptics: Can They See Eye To Eye? Link: https://www.youtube.com/watch?v=6EFHZfISGp4 Views: 1,774,864 Comments: 22913 Published: 28.07.2019

Summary: Three climate change activists and three skeptics were brought together to debate their differences and similarities. The main points of the discussion are the reality of climate change, the responsibility of human activities for it, and how the concern on climate change could be against human rights and freedom.

The precautionary mode: Since the video is an episode of the series “Middle Ground” which aims to bridge people together and enable personal discussions, so the video itself shows different views, all the commentaries were pertained to the main issue “Climate Change” without bots, including debates almost about all the introduced points and arguments. Here we can see the role of the platform and its goals in determining the users and directing their commentaries.

The affirmative mode: The debates concerned different topics such as dealing with scientific truths and denying them by skeptics, The answer of one of the skeptics: “I am not a climate scientist but I know a lot of them” bothered many users, and caused controversy. Most of the comments have many replies and a lot of likes, and most of the users commented once. The commentaries on this video provided a perfect example of the symmetrical approach between science and society.

The empirical mode: The commentaries show a high interaction between users, where the sorting algorithm of YouTube plays a role, and also the power law. In other words, the commentaries with a large number of likes and replies will attract more and more users.

[Moritzw]

Video title: Trump slammed for denying climate change's role in US wildfires Link: https://www.youtube.com/watch?v=VMXcbWwzeY8

using the precautionary mode, i detected no obvious bots. There are some comments i would consider spam, since they are not part of a discussion and have little relevance to the topic shown in the video nor entice reply's, for instance comment „AND SO WHAT DUDE?“ from „JOHN JESUS“ or „Here we go “ from „Kris Roberts“

for the Affirmative mode YouTube provides the feature to sort comments by popularity and shows the likes and dislikes of comments. By using these features i could identify that the most discussion and longest comment chains happened because of the statement „I don’t think science knows“ made by trump in the shown video.

In empirical mode, YouTube shows by default the most popular comments. Those comments contain the most reply's too, with later comments only occasionally showing more than one reply if they have any at all. This seems to be the result of a self enhancing progress, where more popular comments are displayed at the top and are therefore more likely to garner additional reply's.

[yaozheng600]

another teammate: @ChristyLau Video title: Causes and Effects of Climate Change | National Geographic Link: https://www.youtube.com/watch?v=G4H1N_yXBiA Views: 2,082,496 Comments: 2,398

Summary: In this YouTube video, the narrator used a series of assertive sentences accusing human beings of causing climate changes. And the impacts of climate change, which shown in the video, are too far away for most of the audiences. Thus, they focus more on other things rather than what they were expected to learn from this video. They turned to argue about the truth of the statements or put the blame on others. The comments targeting on some countries or someone have also fuelled the debate in the comment area.

The precautionary approach: When studying the social media controversy, the comments related to 'home school', 'online school' and 'test/exam' should be deleted at first, since it has nothing to do with the target problem. Any type of bot related actions was not identified.

The affirmative mode: YouTube filters popular comments and sorts them according to the number of likes and dislikes, which allows us to filter out valuable information simply and directly. And comments with a large number of responses are usually more controversial arguments. And these replies not only debate existing arguments but also generate new arguments, which expands the scope of discussion.

The empirical approach: We should also do the same towards the comments with less than (maybe) 5 words. This will help filter interfering signals like 'gg' or 'wow'. This decision is made based on the empiricist mode. In fact, most comments with few words will not have a lot of likes, and these comments will sink to the bottom of the list.

[yuxin16]

Video Title: How climate sceptics tricked the public - BBC News night Link: https://www.youtube.com/watch?v=_lslIyHo6zg Views: 20,724 Comments: 444 Published: Jul 10, 2019

Short summary: The video describes how the climate sceptics come out and how it influences public opinions on the climate issue. It has been ten years after the "climate gate" but it raised many skeptics and still has influences nowadays.

Precautionary mode: After viewing through the comments, I personally think that no bots can be detected. Most of the user comments are personal expressions, even some of them are off topic (off topic regarding the video, but still on topic of climate change), but there are no generic behavior hidden behind. Somehow, some comments such as "Donald Trump denied global warning" will be considered as useless information (spam) from my point of view. Even though it expresses something, but it can rarely be considered as opinions/ideas, but rather as "thoughtless facts". Otherwise, emotional/biased off topic comments such as "BBC is a propaganda source and nothing more" will also be excluded from analysis.

Affirmative mode: For analysis based on affirmative mode, the comments can be sorted by popularity (I am not sure is it sorted by the number of comments or likes or weighted combination) or recency. Generally the youtube functions are nice if the comments are gathered onto certain debate and discussed as replying interaction. However, the user discussions were most scattered and cannot be categorized into linked comments. The top 2 concentrated debates are: "what was the context of those revealed emails" and "BBC was not trustworthy". The second debate was clearly off topic but we can see gathered discussion in the top 1 popular debate but still somehow the conversation was led to subsidiary direction by someone.

Empirical mode: From an empirical perspective, the youtube features could influence the user interaction in a significant way. But it also depends on how the users use the youtube functions. Some users just watch the video and randomly push their comments/thoughts in comments whereas other people , who may enjoy a discussion, would rather sort the comments and join the debate.

[budmil]

Video Title: The REAL Truth Behind 'Chemtrails' Link: https://www.youtube.com/watch?v=CEii1_kEoZs Views: 354.139 Comments: 4,967

Short summary: Even though the title might lead us to other conclusions, in a 4min long explanatory video, a guy debunks the chemtrail conspiracy theory in a very harsh and mocking manner, but still fact supported.

Comment: According to Marres and Moates' practices, when talking about the precautionary approach analyzing controversial discussions and phenomena on social media in the video I watched it is hard to distinct possible platform driven "irregularities" such as bots. Most of the comments do indeed look too simplified and one-way in the way of negating and opposing what's being said in the video. But I still suppose they are for real. When observing the comment section from an affirmative point, we would probably have the comment replies and likes as the only way of social interacting on YouTube (if we exclude newly introduced features like posts and stories), thus the parameters which feed the algorithm for which comment should be up. The empirical approach would focus on massive threads below some of the comments, which spark the discussion and ignite ugly ad hominem sentences which would should not be dismissed from the dataset, but rather give more insights on the "hotness" and importance of the topic to people on a personal level.

[JuliusBro]

Video title: The Problem With Environmental Alarmism Link: https://www.youtube.com/watch?v=AdDHIsjZ7wY Views: 26,898 Comments: 409 Published: Aug 11, 2019

Summary: In a talk with Ben Shapiro novelist Jonathan Safran Foer makes an argument against radical and panicked measures against climate change, but for well-informed and necessary ones. He compares the symptoms of climate change and the diagnosis by the scientific experts to the symptoms of cancer and its diagnosis by doctors. If the first doctor recommended chemotherapy he would get multiple opinions by other doctors and if, but only if enough of them agreed, he would go through with the therapy. Similarily, 97% of scientific experts agree that there is climate change and therefore people should start to adapt.

Comment: Using the precautionary approach at first, there are no easily identifiable bots. There are a lot of users repeating very similar phrases and sentiments, however this seems to be more because of the issue.

Going by the affirmative approach the best platform-specific metrics are the likes and dislikes of the comments and of the video itself, as well as the recency of the comments. The like to dislike ratio of the video is quite positive, seemingly because the video is hosted by a far-right youtube channel, so the viewer base is rather homogenous. Sorting by number of likes gives a very climate denialist picture with some few interactions from the opposite side of the issue. Sorting by newest first however actually shows that there are a good amount of comments going against the climate denialism, without any further interaction by other commenters.

Lastly, using the empirical approach the way most users interact with this controversy is going to be by seeing the top comments, as this is the default sorting, and either agreeing with them and upvoting or disagreeing and starting a comment chain that quickly devolves into name-calling. Some other users might sort by new or just give their thoughts on the video and write a completely new comment that is unlikely to be seen, unless it really stays as the newest comment.

[raupy]

Video title: Die Lüge vom menschengemachten Klimawandel! - Michael Espendiller - AfD-Fraktion im Bundestag URL: https://www.youtube.com/watch?v=dtlOUMEz4y8 Views: 71.300 Comments: 948 Published: 16.01.2020

About: It's a four minute long Bundestag speech from AfD politician Michael Espendiller. He is commenting on a proposal for mobility research that was made by the government parties and that the AfD doesn't support. He says that the human made climate change is a lie, so that the Great Transition towards green energy and green mobility is based on a lie. It will limit the freedom, restrict the economy and destroy the current human lifestyle. He says that the suffering of the weak German automotive industry is caused by the restrictive governmental defined fantasy threshould values that no manufacturer can reach.

I seem to be the only student who is struggling with the different analysis approaches.. Let's try. From the precautionary point of view we have to clean the data and remove platform artifacts. I couldn't identify any bot comments. Do we also remove the hate comments from our data because they are typical for YouTube or do we find hate on every social network and they are just a mirror of the society? I actually don't quite understand how much cleansing we do from the precautionary point of view. Is it equal to the data cleaning step before you do your actual analysis or is it something additional? From the affirmative point of view we focus on platform specific features like Likes and replies to other comments and use them to identify the most relevant opinions and discussions on the climate change topic. For my video though, this reveals a lot of hate comments and AfD fan comments (or are they already removed from the data after the precautionary approach? or are the approaches stand-alone?). In fact, there is very little discussion on climate change under my video. From the empiricist approach, we can say that because YouTube shows by default the most popular comments, it is probably a self enhancing progress, that they are more likely to attract other people to like them or comment on them. "Funny" comments (at least they seem to be funny to most of the viewers) are therefore more popular than critical or serious comments.

[Francosinus]

Title: What Will Happen In One Billion Years? URL: https://www.youtube.com/watch?v=FVVT8EBqK3g Views: 8,549,192 Comments: 18,534 Published: May 19, 2018

About: This video is about how the earth will change and what could happen to earth and humanity in certain time steps. Some things that are mentioned:

• Earth will be hit by an asteroid
• super volcanic eruption
• supernova -> earth extinction
• sea level falls -> no photosynthesis -> no oxygen -> no life on earth

Reading the comments once again, I dove a bit deeper into the comments and responses from other people. The top comments with the most replies are jokes. Taking a comment as an example: " Weathermen can’t even predict the weather tomorrow". In the reply section a discussion stared about how difficult a weather forecast prediction can be in contrast to predictions about space and earth. So if one analyzes fun comments a little more closely, it is still possible to find controversial discussions.

The precautionary approach:

This approach focuses on data cleaning/filtering to remove platform bias. This is done in the early phase of the research. Bots were mentioned as an example in the paper, but I'm not sure if this is a common thing on youtube videos as well, especially in the area of climate topics. Still I think the first step would be to detect if there is any bias. For example counting how often a user commented (which could then be detected as a bot).

The affirmative approach:

The affirmative approach grasp dynamics that operate across content and medium, analyzing platform-specific dynamics in order to detect issue activity. For that approach it could be useful to have the number of replies and likes of a comment, which are provided by youtube. Taking top comments with a lot of likes and user discussion makes sense to figure out what the people are discussing about and if there are (as I mentioned above) discussions hidden behind jokes.

The empiricist approach:

This approach focuses on the qualification of the empirical object as an objective of social media research. The method to analyze the twitter data used in the paper is, to have a look at what changes over time. But this is about a topic that can have different interactions over time due to events. In case of my video it would be also possible to analyze how the comments change over time. I detected e.g. the appearance of the corona virus as an event where the comments changed. My idea for this approach would be to use topic modelling for example to detect certain keywords which appear in a certain time range. This would make it possible to detect events and how the comments changed due to that.

[Kosmopy]

Video Title: The Anti-Greta, Naomi Seibt, Talks Climate Realism with Dana Perino Link: https://www.youtube.com/watch?v=9zrHs3dv7LQ Views: 150,719 Comments: 1639 Published: Feb 28, 2020

The precautionary approach: When focusing on possible ways to clean the dataset, bot activties seem the most obvious. Viewing over the comments I could not detect anything bot or advertisment related. Another option could be to clean the dataset from non-English comments as it will be difficult to consider different languages when applying AI techniques.

The affirmative approach: Likes, dislikes, date of comment release and number of replies are platform specific features of interest. By sorting the comments by date, we can already identify distinct periods of increased activity. The total number of views, comments and replies can give us information about the popularity and polarity of the comments (and of the video).

The empiricist approach: Social media experience is highly individualized. Which comments I see, which videos next to the video are recommended, which ad is played before the beginning of the video or inbetween differs for every user and affects their experience and platform interactions. It could be interesting to identify paths that led users to the video (e.g. frequently mentioned links) or setting the topics or polarity of comments in relation to the time period they were posted.

[milanbargiel]

Video Title: Global warming: why you should not worry 
Link: https://www.youtube.com/watch?v=pwvVephTIHU Views: 316.445 Comments: 4.186 Published: 14.05.2010

Richard Lindzen, a former MIT professor of meteorology, argues that there is no evidence that global warming is man-made. Commenters are overwhelmingly on the side of the climate skeptics. There are various ways to fuel controversies and a lot of them were mentioned in the previous lecture: People humorise, praise, and politicize in their comments. However, it seems that controversies are especially fueled by absolute statements that can't be negotiated (Example: „there is such a thing as consensus and we definitely have that with climate change" or „Fear is a very effective control device. The government wants you afraid. Then you give up your rights willingly”). Most comments seem to be written by humans. I could neither find bot content nor platform artifacts in this comment section. Depending on the research question I would consider the comments as substantial to the matter.

While analyzing the comment section in an affirmative way, there were several things of interest to me:

• Youtube added an information box on the subject of climate change underneath the video that redirects to Wikipedia
• Climate skeptic comments have more likes than critical ones
• Highly liked comments tend to be discussed more

When taking into account these platform characteristics and mixing them empirically with substantive issues it seems that controversies are not only fueled by (provocative) content but also by likes, number of answers, and the general placement of the content within the comment section.

[isaschm]

Video: Climate Alarmism Isn't Rational Link: https://www.youtube.com/watch?v=tUR0LrSadkg
 Views: 630.387 
Comments: 2.545 
Published: 06.09.2019

The comment section of this PragerU video allows for analysis from several angles as outlined by Marres and Moats (2015). To highlight the viewpoint of each, I will use comments regarding real and perceived platform intervention.

The video has a banner underneath informing viewers of climate change. The banner was ostensibly put up by YouTube itself. If one were to adopt a negative stance on the contribution of platform intervention to the debate, all comments regarding the banner and a suspected future ban of the video would be considered as not related to the controversy. This would be the precautionary perspective’s stance. However, YouTube employing the banner as a signifier for its campaign against misinformation is political and specific to the issue of climate change conspiracies and their dissemination on YouTube. Yet, a significant number of viewers observe the banner and seemingly understand it as an intervention. In the affirmative way of analyzing the comment section, it would be interesting to explore if comments regarding the platform changed over time. In the empirical approach, I would assert that YouTube, as a social media, intervenes in the controversy by putting up a banner in an intentional way. The viewers commenting on the banner and suspecting a coming ban of the video, are equally “mediating” on the role of the platform in the controversy.

[jtembrockhaus]

Video Title: Why Some TV Meteorologists Are Still Climate Skeptics (HBO) Link: https://www.youtube.com/watch?v=LNt-JsBMj4k Views: 87,758 Comments: 838 Published: Feb 17, 2018

Precautionary approach: By a simple screening of the comment section I was not able to observe any signs for bot activity. Thinking of other approaches to clean the data set could be to filter out clear off-topic comment. There are for example comments like “Can someone explain why the dinosaurs are extinct?”, “science is a hoax!”, single emojis or an entire sub-debate whether the interviewing reporter is male or female. A data cleaning regarding this kind of noise could ease the analysis of the actual debate.

Affirmative approach: The affirmative approach investigates how platform-specific dynamics may be an indicator of issue activity. Here, the order in which comments are displayed must definitely be considered since it does not follow a comprehensible sorting. Often liked, replied and newer comments are surely shown at the start of the comment section, but the exact order seems to follow an underlying algorithm that takes more than these properties into account. And since many users do not like to scroll far down, it can steer the general course of the debate by that platform property.

Empirical approach: From an empirical perceptive, it could be very interesting to analyze if the debate may have changed over time. The video was published at the beginning of 2018 and since then, the social awareness for the climate change and whether it should be discussed in TV weather forecasting shows may have changed. There could be a different proportion of users arguing for the pro and contra side in recent comments compared to early comments in 2018.

[DanielKirchner]

Video: ‘Don’t lie about climate change’ in Pacific

Link: https://www.youtube.com/watch?v=d9EqGa3PvNo

Views: 38.395

Comments: 520

Published: 13.08.2019

Precautionary:

There doesn't appear to be an overwhelming (if any) amount of non-genuine comments (bots, spam), and all of the comments appear to be about the video in some way. Only some comments carry no information for the discussed topic ("Is the a main channel?").

Affirmative:

The video seems to have gotten all of its views in a period of about half a year (last comment was made 10 months ago). YouTube seems to value upvotes of comments more their freshness, since all of the suggested top comments are from around the time the video was uploaded, while having many upvotes. Long comment chains seem to attract more comments, as there are extreme outliers in comment chain length (52), while the average comment chain length must be very close to one.

Empirical:

The YouTube comment sorting algorithm seems to play a role in how the video is perceived by viewers that find the video after some time. Since new information on the subject will be very unlikely to be high up in the top comments, because it will not be seen, the public consensus about the video will be very static after some time.

[travela]

Video Title: Alex Jones: Hookers descend on Copenhagen 
Link: https://www.youtube.com/watch?v=7Az05gIzEYM Views: 22.247 Comments: 71
 Published: Dec 7, 2009

Precautionary approach: The overall number of comments is quite low and there is no indication that any comments were left by a bot. Especially given that the video is more than 10 years old, a time when bots were less frequent and established. However, it is to say that some users are following their agenda, i.e. namely promoting conspiracy theories such as free energy suppression and new world order, for which they specifically seem to target this video as it supports their view.

Affirmative approach: It seems like most comments were made roughly around the time when the video was uploaded. There appears to be a small cluster of comments that appeared two years later, with no indication of media platform influences or any major underlying news events. (This is hard to tell as the the cluster is simply to small.)

Empirical approach: Albeit it is not YouTube itself that stirs the discussion into any direction, it is the channel that uploaded the video. The channel being from a Russian broadcaster, a substantial amount of the comments see the content of the video in the light of this and start a meta discussion.

[xixuanzh]

Video: Global warming: why you should not worry

Link: https://www.youtube.com/watch?v=pwvVephTIHU

Views: 316.456

Comments: 4.067

Published: 14.05.2010

Summary: A MIT scientists (climate change skeptic) claimed that it is hard to prove the climate change scientifically.

In Precautionary Mode: The video I chose is quite old (posted more than ten years ago), and it was listed on the first page of my search query of "climate lie." The first 200 comments are relatively new here (two months to one year old). Why did the video well placed on Youtube? We should think about content modulation and platform regulation here.

In Affirmative Mode: The order of comments displayed (platform affordance) and the positions towards climate change in the comments (posting pattern) would be the research object. Also, how do the chains of responses look like (denier posted --> supporter replied-> denier replied-> supporter replied--> supporter replied) would also be worthy of scrutinizing. Do old comments also trigger the same cascade of responses as the new ones?

In Empiricist Mode: We can take a look at the dynamics of the commenting pattern from 2010 to 2020. For instance, to see whether people are mentioning new facts found by scientists and to observe if the current issues related to climate debate such as environmental governance, social movement, the raising attention on disinformation change the discourse dynamics on Youtube.

[adrigru]

Video title: Why humans are so bad at thinking about climate change Link: https://www.youtube.com/watch?v=DkZ7BJQupVA Views: 1,924,825 Comments: 6,197

There are no noticeable signs of bot-generated content in the commentary section. The discussion relates mostly to humans not doing anything against climate change. Many users point out what the primary causes of global warming are, and that, a single individual cannot do much himself against it. There are a few not relevant comments about the video itself like "1:18The background lol" that I would exclude.

Thanks to YouTube upvote function for comments as well as the view reply option, I was able to identify the most active discussions. Usually, the comments with the most replies are the most controversial ones. Moreover, the platform allows sorting the responses based on popularity.

Following the empirical approach, it might be interesting to analyze how the users interact with the platform. Since the default sorting is by popularity, many might engage in discussing one point over and over, and other arguments might remain undiscovered.

[Erdnaf]

Video Title: Geoengineering: A Horrible Idea We Might Have to Do 
Link: https://www.youtube.com/watch?v=dSu5sXmsur4 Views: 3.584.639 Comments: 18.173 Published: 27.10.2020

In Precautionary Mode:

Despite there beeing many comments to the Video, there is no clear sign for Bots. Therefore we don't have to filter the data.

In Affirmative Mode:

Most of the dicussion happened in the "higher-ranked" comments. I Noticed taht the ranking combines the 'Upvotes of the comments and how recent they have been pulished. For Example there were comments that had much more upvotes than others, but despite that they were rankesd lower by the algorithm than them becase they were older. this Sorting via Top comment- Withc is the default setting, schould be considered when analysing to get a good result. Because the video is not very old, there are still a lot of new comments

In Empiricist Mode

From an Empriristic perspective one can look at the channbel witch uploaded the video. It is an well establishes scientific channle and known for precice Research. In the video description ia a link to a webpage with all the sources braken down by topic. This could maybe lead to a healthy discussion, because everyone can research the things stated in the video by himself

[Alioio]

Summary:

Video title: AfD-Physiker sprengt Klimahype mit einer Grafik! – Dr. Christian Blex (AfD) URL: https://www.youtube.com/watch?v=XGVUV2tpR5g Views: 73.774 Comments: 762 Published: 27.10.2020

About: A german politician holds a speech at a regional parliament about the history of climate change. He is presenting a diagram to the audience where temperature changes of the past 11T years is visualized as an area chart on a printed piece of paper.

He is referring to warmer periods in the history and is pointing out that climate change was normal in the history. While doing so he sharply criticize and attack the opposition .

---

Precautionary Approach:

The precautionary approach is focusing on data cleaning and removing platform bias eg. comments which are obviously produced by bots and/ or comments which are not related to the topic. By reviewing the comments from my video I couldn't find any comments which could be produced by bots. Although there are many short meaningless comments like (Some of them from the same user repeatingly) I believe that these are comments which are praising the speaker in the video and expressing belongingness to the party and the attitude they represent.

Affirmative Approach:

With this approach the platform-specific dynamics are investigated to find indicators for issue activity. Platform specific attributes I could find are: Number of comments, amount of answers to a comment, number of likes for a comment, number of likes to the video itself, possibility of addressing directly someone a commenter want to respond to by adding a "@" as prefix to a user name. In the comments I analyzed there were many threads were disputes formed where users make use of the "@" prefix to directly respond to each others comment. What can be noted too is that youtube's default sorting of the comments is Top comments by the amount of answers to a comment and the number of likes. The user has the ability to switch between the sorting "Top comments" and sorting by date "Newest first".

Empirical approach:

In the empirical approach one thing I think needs to be considered is that the default sorting algorithm of youtube is maybe steering the attention of the users to some comments with many likes in one or the other direction. I think this is intended by youtube to generate fuel discussions on comments which tend to be more popular then others. However it would be good to put some lesser rated comments in between to avoid self-reinforcing popularity on some discussions. By taking a look into the content of the comments and reply I can notice that there are many threads formed over time which seems to be off topic on climate change but somehow still related in matters of the rhetoric and attitude the speaker and the party he belongs is representing. Therefore maybe a topic modelling could be beneficial.