Opinion Leaders Of #MarchforOurLives On Twitter – Who Are They And What They Want?

March for Our Lives was a student-led demonstration in support of tighter gun control that took place on March 24, 2018, in Washington, D.C., with over 800 sibling events throughout the United States and around the world. The event followed the Stoneman Douglas High School shooting, which was described by many media outlets as a possible tipping point for gun control legislation. The aim of this research is to visualize the Twitter sphere under #MarchforOurLives movement. Because of the limitation of the data collection, we add keyword “law” to observe the possible opinions towards the protest under the legislative issue in the discussion of this movement. So the research question would be:

  • What is the salient opinion towards the protest?
  • Who are the key opinion leaders (KOLs)?
  • Is there any organization behind the discussion? If any, what is the aim?

 

An overview of Data

We choose #MarchforOurlives AND law as the keywords to filter and collect the tweets with TAGS (https://tags.hawksey.info/get-tags/). The script ran from March 24th, 04:38:41 to March 24th, 23:59:59. The raw database has 11,758 unique tweets. The data cleaning procedure excluded tweets of non-English source, leaving 11,613 tweets as data pool. Using Tableau, we compared the most followed users and the user visibility to see if there are any difference of users’ popularity and visibility. The element most followed refers to the followers’ number while the visibility is the ranking according to the mentioned and retweeted times in the tweets using “@”as the symbol.

屏幕快照 2018-04-25 下午12.17.11

After the comparison and some further research on the background information which is public on every users’ twitter account page, it is clear that the accounts with most followers are different from the accounts which are most visible. To be specific, among the top five most followed users, four of them are news agencies, however the most five visible users are all individuals or at least shown on the Twitter as individual run account. That is to say, tweets from individual users are more likely to be retweeted or mentioned at, thus gain more visibility than news agencies. But does that mean the normal people have the same opportunity to spread their opinions like the news agencies instead of always being influenced? To study this, we compare the retweets amount ranking with the Tweet reach ranking. Tweet reach indicates that how many people has seen your tweets directly. It is calculated in an assuming concept that ideally there has 10% of the followers have actually receive and see the tweets. Although this is an ideal estimation, the ranking will not be influenced.

re and reach

The result shows that the top five tweets, whether in the ranking of retweeted amount or tweet reach, are actually the same content, which means the tweets with larger reach can gains more retweet opportunities. This is the tricky thing. Social media like Twitter claims that everyone has the chance to be heard or seen. In fact, at least in this case, with small amount of followers, it is very hard to bring your opinions on the table. Impact could be mostly described by your amount of followers. That is how the key opinion leaders gain their visibility.

This result is supported by Marwick and Boyd (2011) and by Page (2012), saying that providing newcomers and listeners opportunity to have a “conversational environment” on Twitter does not mean that participation and influence opportunity are equally distributed. In fact, the two step flow theory, which firstly acknowledged by Katz and Lazarsfeld (1955), could be employed to explain this phenomenon. In their study, the function of person-to-person communication is more likely depending on the role of opinion leaders in influencing others. Now as social media like Twitter has become new arena for various personal influence and opinion leadership to grow, two step flow still echoes in this new environment. In the analysis of 2010 US midterm election, Vaccari and Nielsen (2013) found that most politicians fail to achieve large-scale direct communication with citizens on social media in spite of the large followers on these platforms. Which means influence does not equal to large audiences.

Moreover, in our finding, the visibility of the KOLs through time is also interesting. Comparing the tweets amount over time and the top users’ visibility over time, we can tell that the top user’s retweet amount suddenly surges to the top, then fade out through time. However, the overall tweets amount rise after descending at the tail part, which probably is because the other KOLs’ tweets are gaining more visibility but are diluted by the big figures, which means that the KOLs’ influence may not be stable.

 

Terms structure analysis using Voyant

To better understand the attitude and opinions towards the protest, we put all the text of our data pool into the Voyant () which is used to explain how the words organized in visualization form.

屏幕快照 2018-04-23 上午11.51.50屏幕快照 2018-04-19 下午8.12.26屏幕快照 2018-04-23 下午12.05.38屏幕快照 2018-04-23 下午12.08.22

The frequency map shows the size of the words in degree of their appearance frequency. We can see that the words like gun, enforcement, people, stop, abiding are the most shown in all tweets. The total words association scattering graph further illustrates the distance of the terms. But it is not clear what is the meaning behind them. When we look at the top five related terms in the allocation form, the word groups are more meaningful. Then we put them into the context analysis. We want to study the attitude towards the protest. So we choose the word “protest”. The result shows that the top in frequency is the supporting attitude. Second is the news report which is neutral. The third and fifth are against the protest, focusing on the unjustified aim and approach of it.

Content analysis

To further study the content meaning, the content analysis is employed. Breaking the attitude into seven parts, we randomly selected 100 tweets from the data pool and manually put them into the seven categories. The definition and example of the categories are shown in the coding scheme (Appendix 1). The result shows that 44% of tweets are negative towards the protest, while 35% are positive. Among the negative attitudes, 22% thinks the protest is unjustified, larger than the dangerous or useless categories, which verifies the above Voyant words structure analysis. Furthermore, the neutral opinions only appear to take 16%, which shows that the opinions have polarized into mainly two camps, not mention most of the neutrals are news reports without any unfolding of opinions.

屏幕快照 2018-04-15 下午10.12.27

 

Closer look at the key opinion leaders

Given the overall attitudes being into two camps and the two step flow seems to play a critical role. It is better to have a closer look at the situation of the top opinion leaders who are assumed to be the source of the opinion polarization. To run it in Gephi, we further cleaning the data, leaving 10263 retweets. With ForceAtlas 2 algorithm which illustrates the force-directed relative position, the nodes and edges are positioned to simulate the motion of them. After filtering by the modularity degree at 70 in order to show the pattern clearly, 7421 nodes and 7693 edges are left in the graph. The top 15 in-degree users have been shown with user ID remarks. The blue remarks indicate that the political background of the user is probably liberal, while the red means conservative. The green are the news agencies or related staffs. The black is the neutrals. All political spectrums are subjectively recogonized according to their personal introduction and tweets content.

final

As illustrated in the picture, @ProudResister is at the bottom of the picture, with all other directions surrounded by conservatives. If we only look at the top 5 users, the force directed distance between the liberals and conservatives is quite far. With nearly non neutrals in the middle of the picture, two camps’ related users are actually departing from each other. But who are they and what kind of people are supporting them? Let’s have a deeper look to the activities of these top players.

Twitter ID In-degree Social Identity Political implication
@ProudResister 1067 Activist New Left, #TheResistance
@RodStryker 581 Indiana State Director New right.
@TomFitton 566 Author Conservative
@CoreyLMJones 466 VA State Director for @NewRightUS New Right
@NatShupe 413 N/A Conservative
@StephenGutowski 351 Free Beacon Staff Writer Conservative
@AmericanHotLips 276 N/A Conservative
@pinkk9lover 187 N/A Conservative
@grantstern 183 Mortgage Broker, columnist Neutral
@Doodisgirl 168 Wife & Mom Conservative
@PoliticalEmilia 165 14 year old commentator Left
@LisaMirandoCNN 161 Executive Producer of CNN Neutral
@RealJack 135 Commentator New right
@davidwebbshow 128 The Hill columnist, N/A

In the graph above, we listed the user ID according to the in-degree value provided by Gephi. Clearly, the top part of this graph is dominated by activist and writers, while for the bottom part the commentators and columnists seem to take control. The hashtags in their personal introductions indicate that for the conservatives, the most frequently mentioned is #2a and the term “new right”, while for liberals the term “resistant” or other related hashtags or terms are most visible. Then we use tableau to look for the users who have top amount of followers and relay the top retweet for both side.

KOLs ranking (Conservatives)KOLs ranking (Liberals)

屏幕快照 2018-04-25 下午12.31.49

In picture, we can see that for the liberal top retweet, many of the zealous users are activists working for “the wave” or “blue wave” movement. On the other side, the most passionate conservatives, without mentioning their political identity, are more likely to describe themselves as patriots and use lots of US national flag emoji.

After checking the hyperlink of the website of the top users, we find that for both side, the political mobilizing sites all call for normal people’s support, highlighting grassroots movement, and call for voting for two parties. which probably meets the aim of their retweeting participation.

屏幕快照 2018-04-25 下午12.32.22屏幕快照 2018-04-25 下午12.32.36

Discussion

On the issue of law, march for our lives movement is criticized by conservatives mostly for its lack of credibility and morality. For liberals, the request for gun control is the salient theme of retweeting.

The traditional media is less influential than the KOLs in this case. Two step flow theory works. Among the KOLs, the activists and writers play a relatively important role. The influential tweets are originally from activists, writers, columnists and commentators. The most influential retweet users are of difference between two camps. For conservatives, key retweet users tend to describe themselves as patriots and conceal their offline social identity, while liberal key retweet users are more likely to be activists and do not mask the promotion purpose for their organizations. Although media is less powerful, the retweet amount and the followers amount are still positive correlated. Users with more followers are more likely to gain more retweets, thus accumulate more visibility.

In this study the result supports the observation that recently the US politics landscape has further polarized. In US politics podcast “选美”, host You Tianlong highlights that part of the reason is the passionate willingness for the extreme camps to speak and advertise on the social media, bringing others the impression that the two camps are divided. Furthermore, President Donald Trump is used by both sides for the personalized image of the rising or worsen politic landscape thus simplify everything into pro or con Trump Administration as what we see above in the political organization websites’ campaign. Donald Trump himself, although mentioned by some users, do not personally participate in this discussion. Given that Trump is in some way at the center of the discussion, the disappearance of his opinion is quite appealing.

This study has several limitations. First, limited on the capability of online Twitter data spider (TAGS). The tweets are released by Twitter through specific algorithm whose criteria is not available publicly, which means that the collected data may not be the whole original data. Besides, the TAGS allows maximum 20,000 tweets in one collecting sheet, far less than which the complete number of the #marchforourlives movement could reach, which is part of the reason that we add “law” to restrain the collection range. It may help to specify the research topic while undermine the data integrity at the same time. Secondly, it should be beard in mind that the influence does not equal to the in-degree or retweeted amount, or the tweet reach amount. When we say visibility, it is mostly used to describe the retweeted amount. But the influence is a more complicated concept. In their study of 2013 Italian general election campaign, Cristian and Augusto (2015) arguing that the influence of political communication on Twitter should not be limited to the ones that could be reached directly, but must also include the secondary ones that could be reached indirectly through the direct audience.

Future work should consider more comprehensive measure for the political influence as well as the condition through days or months which could possibly show more information. Besides, the opinions of the subgroups could be paid attentions to because they may be the explanation of the flexibility of the KOLs’ influence.

 

 

References:

  1. Wikipedia March For Our Lives https://en.wikipedia.org/wiki/March_for_Our_Lives
  2. Force Atlas2 Algorithm https://github.com/gephi/gephi/wiki/Force-Atlas-2
  3. 选美Podcast Episode 60  http://xuanmei.us/60
  4. Katz E and Lazarsfeld P (1955) Personal Influence: The Part Played by People in the Flow of Mass Communication. Glencoe, NY: Free Press.
  5. Marwick AE and boyd D (2011) I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society 13(1): 114–133.
  6. Page R (2012) The linguistics of self-branding and micro-celebrity in Twitter: the role of hashtags. Discourse & Communication 6(2): 181–201.
  7. Vaccari C and Nielsen RK (2013) What drives politicians’ online popularity? An analysis of the 2010 U.S. Midterm Elections. Journal of Information Technology & Politics 10(2): 202–222.
  8. Cristian V and Augusto V (2015) Follow the leader! Direct and indirect flows of political communication during the 2013 Italian general election campaign. new media & society 2015, Vol. 17(7) 1025–1042

 

Appendix 1:

Themes Definition Example
Dangerous protest Opinions criticizing the protest for the result of it could be very dangerous like more people killed, destruction of US democracy etc. RT @NatShupe: Socialists have always loved exploiting children and using them for propaganda.

Especially when you can use them to disarm a nation.

The end goal for #MarchForOurLives organizers/Democrats is to eliminate the 2nd amendment & to disarm all law abiding citizens. https://t.co/IVXq2BPZPA

Useless protest Opinions that is pessimistic on the protest for it is useless and make no change to the reality. RT @MitchBehna: “Oh damn, they just passed that new anti-gun law? I better turn in these guns I have illegally.”

-Said no criminal ever

#MarchForOurLives #2A

Unjustified protest Opinions criticizing the protest for the unqualified motivation that makes it suspicious like Russian penetration, politician propaganda etc. RT @Gogoette: @genthemartian Sure and the armed guard did too and did NOTHING. In Maryland the guard DID SOMETHING and saved lives.  This march is one-sided agenda based on emotion not on law and rights. #MarchForOurLives
Neutral Neutral opinion like news report.. RT @natsecaction: Today, over forty former national security, homeland security, law enforcement, military, and intelligence officials stand with #MarchForOurLives 👇 https://t.co/vqI9QYzejy
Themes Definition Example
supporting Opinions supporting this protest RT @ProudResister: #MarchForOurLives Demands:

1. Pass law Banning Assault Weapons.

2. Stop sale of High-Capacity Magazines.

3. Close loopholes for Background Checks.

Request People request the politician or related people in words. RT @rustinore: So impressed with today’s #MarchForOurLives.  All the young speakers were so articulate, expressing incredibly powerful & well reasoned arguments for significant gun law change.  Don’t stop!

@NRA loving politicians will not be eligible for my vote.

@Emma4Change

@davidhogg111

Call on People call for action like voting or further protest RT @4everNeverTrump: Donald Trump declaring he’ll ban bump-stocks yesterday is a joke. It’s not supported by existing law and it’ll be struck down by the courts.

He and the GOP have made no effort to buck the NRA.

#MarchForOurLives today. And march to the voting booth on November 6.

Advertisements