Early Public Responses to the Zika-Virus on YouTube: Prevalence of and Differences Between Conspiracy Theory and Informational Videos

Summary:

As YouTube emerged as a popular social media platform, it also became the favorite playground of fake news and conspiracy theory videos. Surprisingly, most of the videos on this website are health-related. This creates a concern that fake news and conspiracy theory content can mislead and install fear among the audience thus affecting the spread of epidemics. The researchers collected the content of the 35 most widespread YouTube videos during the first phase of the Zika-virus outbreak – from December 30, 2015, to March 30, 2016 – and user reaction to those videos. This paper aims to find the dissimilarities in terms of user activity (number of comments, shares, likes, and dislikes), and the attitude and content of user responses between informational and conspiracy videos.

Unexpectedly, there is no significant difference between these two types of videos in most cases. Both informational and conspiracy videos share not only the same amount of responses and unique users per view but also a low rate of additional responding per unique user. Furthermore, the user comments on these two types of videos are slightly negative, which challenging Vousoughi, Roy and Aral’s conclusion that false news provokes more negative sentiments than true news. However, there is a dissimilarity in the content of user response between informational and conspiracy videos. Informational videos user responses focus on the consequences of the virus. Whereas, the conspiracy theory videos user responses are more interested in finding out who is responsible for the outbreak.

Reflection:

    Even though the studying of misleading health-related content on YouTube might be worth researching, the authors’ choice to limit their research to the topic Zika-virus is, sadly, not a wise decision. First of all, despite the fact that YouTube has over one billion highly active users watching around six billion hours of videos per month, it is also true that distribution of user age and interest are extremely skewed. Also, a large percentage of YouTube users are teenagers and young adults who have little or no interest in Zika-virus. In fact, there is a huge gap in user activity and engagement between popular topics (gaming and beauty) and other videos. Much of the popular health-related content (anorexia, tanning, vaccines, etc.) is also linking to beauty trends. In other words, YouTube is not an ideal social media forum for studying user response to different type of videos of Zika-virus or most health issues. The perfect proof for this is the extremely small dataset the authors were able to acquire on the topic. So, if the same question was asked about videos related to vaccines, the result might have been more interesting.

    Another concern about the dataset is that there is a lack of explanation and inspection of the data collected. What are the user statistics: age range, gender, prior behavior, etc.? Do the dataset includes videos that were reported as many conspiracy theory videos can be flagged as inappropriate? Why did the authors choose 40,000 as the cut off for their dataset? How popular are Zika-virus related videos compared to other videos posted on the same day? 

    As mentioned above, the most active group of YouTube users are not attracted to the topic of Zika-virus. Older users are less likely to show reactions on social media forums, especially on controversial topics. Therefore, the user inactivity and passiveness can make the difference in user activity between informational and conspiracy theory videos seem insignificant. Also, the low rate of additional responding can also be the result of the user behavior on social media forums rather. There needs to be more analysis of the user behavior before concluding that the user activity is similar across the two types of videos.

    The most interesting analysis of this research is the semantic maps of user comments between informational and conspiracy theory videos. The clusters of the informational videos are bigger and more concentrated. Surprisingly, offensive words (shit, ass, etc.) are used more frequently in informational videos. Moreover, the comments in conspiracy theory videos are more concerned with foreign forces such as Brazil, Africa, etc. Meanwhile, the audience of informational videos focuses more on the nation’s eternal conflicts between parties and religions. What might be the cause of the difference in interest and usage of offensive language? The fact that Bill Gates, the only businessman, is not only mentioned but also has his name misspelled frequently is interesting. Why did he appear as much as presidents of the United States in the comment? Does the common misspelling of his name indicate the education level of the audience?    

Automated Hate Speech Detection and the Problem of Offensive Language

Summary:

    The paper address one of the biggest challenges of automatic hate-speech detection: distinguish hate speech from offensive language. The authors define hate speech “as a language that is used to expresses hatred towards a targeted group or is intended to be derogatory, to humiliate, or to insult the members of the group.” Even with a stricter definition of hate speech than previous studies, the authors still find the distinction between these two types of expression too ambiguous and case specific. The raw data contains tweets with terms from the hate-speech lexicon, which is compiled by the website Hatebase.org. Then, the tweets were reviewed and grouped into three categories: hate-speech, offensive language, and none of the above. After trials and errors, the authors decided to build a logistic regression model. Even though the algorithm achieves a relatively high overall precision of 91%, it does not accurately differentiate hate-speech from offensive languages with a misclassification rate of 40%.

    Despite the improvement in precision and accuracy compared to previous models, the research acknowledges that there are plenty of issues with the current model. The algorithm often flags tweets as less hateful or offensive than human coders. The heavy reliance on the presence of particular terms and neglect of the tone and context of the tweets result in a high misclassification. Moreover, it is noticeable that sexist tweets are much less likely to be considered as hateful than racist or homophobic tweets. The authors suggest future research should consider social context and conversations in which hate-speech arises. Also, studying the behavior and motivation of hate speakers can also provide more insight into the characteristics of these expressions. On the other hand, another pitfall of the current model is that it performs poorly on less common types of hate-speech such as those targeting Chinese immigrants.

Reflection:

    It is not an exaggeration to say that even humans are struggling to differentiate hateful speeches from those that are merely offensive, let alone an algorithm. This research shows that the presence and frequency of keywords are by far sufficient in distinguishing hate speech. Indeed, in most cases, the decisive factors determining whether a tweet is hateful or offensive are the context of the conversation, the sentiment of the tweet, and the user speaking pattern. However, analyzing these factors is much easier said than done.

    The authors suggest user attributes such as gender and ethnicity can be helpful in the classification of hate-speech. But acquiring these data might be impossible. Even in the few cases in which users agree to provide their demographic information, the trustworthiness is unguaranteed. However, these attributes can still be derived from analyzing the user behavior in the forum. For example, a user who often likes and retweets cooking recipes and cat pictures have a higher chance to be a woman. Therefore, future studies can consider looking into a user prior behavior to predict their tendency of expressing hate-speech and classify their tweets’ nature.

    One of the main causes of the high misclassification of hate-speech is that the word choices and writing styles vary greatly among users. Even though most hateful expressions are aggressive and contain many curse or offensive terms, there are plentiful exceptions. An expression can still “intended to be derogatory, to humiliate, or to insult” disadvantaged groups without using a single offensive term. In fact, this is the most dangerous form of hate-speech as it often disguises itself as a statement of truth or solid argument. On the other side, there are statements using flagged terms yet convey positive messages. Therefore, rather than focusing on the vocabulary of the tweets, it might be better to analyze the messages the users want to convey.  

Read More

Benjamin D. Horne and Sibel Adali. “This Just In: Fake News Packs a Lot in Title, Uses Simpler, Repetitive Content in Text Body, More Similar to Satire than Real News.”

Summary:

The main goal of the research is to build a model to differentiate fake news from real news. The authors also analyze satire news, which is considered a type of fake news in this paper. The data is collected from 3 data sets: Buzzfeed 2016 election data set; Burfoot and Baldwin data set; and a data set containing fake, real, and satire news created by the authors. The articles are analyzed and compared with each other based on three main feature categories: stylistic, complexity, and psychological. A Support Vehicle Machine classification model is built with the four most significant features selected from ANOVA and Wilcoxon rank sum test. 

The result of the research is in sync with previous studies of fake news, which can be summarized in three main points:

1.    Fake news articles have more lexical redundancy, more adverbs, more personal pronouns, fewer nouns, fewer analytic words, and fewer quotes. This means that fake news articles is much less informative and require a lower educational level to read than real news articles.

2.    Real news articles convince readers through solid reasoning while fake news supports their claim through heuristics.

3.    The content and title of fake news and satire articles are extremely similar.

Reflection:

First of all, the authors definitely practiced what they preached. The title of the paper is packed with verb phrases yet contains zero stop word. Moreover, all three main conclusions are included in the title, which hopefully will increase the chance of this article, or at least its main points, being read. Despite it might sound like a half-hearted solution, the authors’ suggestion of transforming real news articles’ titles to resemble that of fake news articles is actually a good idea. What will happen if we create the title of real news articles using the formula for fake news articles? Will people be more likely to read them? Will people classify them as fake news based on their titles?

    In spite of successfully building a classification model with a relatively high accuracy distinguishing fake and satire from real news articles, the research fails to deliver any new findings. Indeed, the result is nothing but a reconfirmation of previous studies. Furthermore, the difference between fake and real news articles seems obvious to most people. Similar to clickbait, detecting fake news articles is relatively easy. But the bigger question is how can we improve people willing to read real news articles instead of scrolling through a list of fake news titles?

    One interesting finding is the different title features of fake news articles between BuzzFeed 2016 election news dataset and the political news dataset collected by the authors. The former one uses significantly more analytical words. Nevertheless, the later one has more verb phrases and past tense words. That suggests that there is more than one type of fake news. Their difference in word choices also suggests they might be targeting different groups or trying to provoke different reactions. It can be interesting to study the cause of the distinction in feature between fake news articles.

    In addition, it is surprising that the SVM model produces much more accurate classification with satire articles than with fake news articles. It “achieve a 91% cross-validation accuracy over a 50% baseline on separating satire from real articles.” On the other hand, the model only “achieve a 71% cross-validation accuracy over a 50% baseline when separating the body texts of real and fake news articles.” Is that because of the mocking tone that distinguishes satire from real news articles? Furthermore, the model has a low accuracy when separating satire from fake news articles. This might post an issue as we might want to treat satire articles differently than fake news articles.

    Lastly, the distinction between clickbait’s and fake news articles’ title is quite intriguing. Because of their similarity in lack of validity, ethics, and valuable information; many people put clickbait and fake news in the same category. Yet, these two types of articles serve completely different purposes. Clickbait encourages readers to visit the web page thus the titles “have many more function words, more stop words, more hyperbolic words (extremely positive), more internet slangs, and more possessive nouns rather than proper nouns.” Fake news, meanwhile, wants to deliver their messages even if the majority of the links are never clicked. Hence, their titles, loaded with claims about people and entities, are an extremely concise summary of the whole articles. One way or the other, both fake news and clickbait have found their strategy to attract readers’ attention and engagement. So why real news articles are failing so far behind? Is it because they are not aware of the tricks fake news and clickbait are using? Or that they are too proud to give up their formal and boring titles?

Read More

Analyzing Right-wing YouTube Channels:
Hate, Violence and Discrimination S

Summary:

The paper investigates the occurrence of hateful content and discriminatory bias in right-wing channels. First, the authors examine the similarities and dissimilarities between comments and video content on YouTube channels to see whether content creators express less, equally, or more hatred and discrimination than their audience. In addition, the results of right-wing channels are compared to those of baseline channels to check if extreme content is common on social media platforms or more prominent in right-wing channels.

The dataset consists of videos and comments from 12 right-wing channels and 10 most subscribed channels on YouTube. The authors conduct a three-layered analysis through which they evaluate lexicon, topics and discriminatory bias in videos and comments from the collected channels.

The results suggest that right-wing channels’ topics are more concentrated, focusing on terrorism and war. Moreover, these channels have a higher percentage of negative word categories, such as aggression and violence compared to baseline channels. Surprisingly, the level of bias against immigrants and LGBT people is comparable between right-wing and baseline channels. However, the negative bias against the Muslim community is more noticeable in right-wing channels. The analysis also shows that right-wing video hosts use negative words as often as commenters, yet the actual semantic fields vary. On the other hand, compared to the content creators, the viewers of right-wing channels express a higher bias against LGBT people and lower bias against Muslims in the comment section.

Reflection:

It is obvious that the research, as a whole, is partial and flawed. Not only that the title and the research questions reflect the authors’ strong presumption of the conservative section but also the authors’ choice of data is questionable. Why did they choose the website InfoWars and Social Blade as a seed? Furthermore, the phrase “according to our understanding” suggests their description of right-wing is not only ambiguous but also potentially biased. The confusion carries over to their selection of baseline channels. Despite being selected from the “news and politics” category, the fact that the majority of the comments are related to gaming (RiceGum, PewDiePie, Minecraft) means those channels cover a much wider range of topics. One of the channels is even called “DramaAlert”, which contributes more than one-third of the total number of comments of baseline channels. Hence, there is no surprise that entertaining videos contain more positive words than those discussing serious, controversial topics. Rather, the research should have studied the extremist behavior in political predominant YouTube channels. It will be interesting to see if right-wing channels can be separated from the mix.

Even though the authors’ bias discredit many of their findings, some analysis might still provide some insight and inspiration for future research. One worth noticing observation is that Muslims and terrorism overshadow other topics such as LGBT, illegal immigrants, anti-abortion, etc. Is terrorism always the prominent focus? Has this topic gained more attention because of the series of recent attacks and the President’s election campaign? It might be interesting to see if the change in the trend of topics and people opinion through time, especially after a major incident.

Another surprising finding is the dissimilarity between the video hosts and the commenters. Despite the comparable level of negativity, the variance in the semantic fields present in the caption and in the comments in right-wing channels varies greatly across videos. In another word, the viewers’ response to the conservative videos is more unpredictable and less affected by the sentiment of the hosts. The authors mentioned previous studies show hateful users “more negative, more profane and, counter-intuitively, use fewer words associated with topics … ” Does this suggest right-wing channels’ audience is more hateful and aggressive? Do their comments remain as offensive and negative in other videos? Does a user’s tendency to leave a hateful comment correlate to his/her subscription?

Read More