key: cord-0644397-xjnko0h6
authors: Panda, Anmol; Siddarth, Divya; Pal, Joyojeet
title: COVID, BLM, and the polarization of US politicians on Twitter
date: 2020-08-07
journal: nan
DOI: nan
sha: 527f99f198eca4296d999ce2938185deb6bbf390
doc_id: 644397
cord_uid: xjnko0h6

We mapped the tweets of 520 US Congress members, focusing on analyzing their engagement with two broad topics: first, the COVID-19 pandemic, and second, the recent wave of anti-racist protest. We find that, in discussing COVID-19, Democrats frame the issue in terms of public health, while Republicans are more likely to focus on small businesses and the economy. When looking at the discourse around anti-Black violence, we find that Democrats are far more likely to name police brutality as a specific concern. In contrast, Republicans not only discuss the issue far less, but also keep their terms more general, as well as criticizing perceived protest violence.

For years, politicians have been using social media to communicate directly with voters, and researchers have studied this phenomenon [6, 8, 9, 11, 12, 14] . During this time, the study of politicians' social media use has covered a wide range of topics, with focus on partisan messaging [7] , engagement with constituents, polarization, [1, 10] , volunteer mobilization [4] , targeting of opponents, and propaganda as observed in Germany [2] , Austria [15] , Italy [5] , France [3] and India [13] .

In the present article, we contribute to this space with an analysis of tweets shared by the official handles of 519 members of the US Congress (ninety-seven Senators and four hundred and twenty two members of the House of Representatives who have active and verifiable accounts on Twitter). The period of interest was from Jan 1, 2020 to June 29, 2020. This period was chosen specifically to study the politicians' framing of the COVID-19 pandemic, and accompanying public health and economic crises, as well to capture discourse around the recent wave of anti-racist protest across the country.

In this study, we intend to answer the following questions:

(1) How did the two parties frame their discussion around three key topics: Covid-19, the George Floyd murder, and the Black Lives Matter movement? (2) What does this framing say about the willingness to address or avoid specific issues?

The primary list of politicians was hand-drawn, seeding from a list of members of Congress in late 2019 and searching for their handles on Twitter. It was complemented with more recent data and metadata from Github 1 . In all, we collected 310,686 tweets of 520 legislators of the US Congress using Twitter's public Search API.

With over 10 million followers, the @SenSanders handle of Senator Bernie Sanders (I-VT) was the most followed account in Congress, while @GuamCongressman, the official handle of Michael F.Q. San Nicolas, member of the House from the US territory of Guam, was the least followed with 628 followers at the time of the study. The accounts of @Jim_Jordan (R-OH) and @RepAdamSchiff (D-CA) were the most retweeted for the two parties in the study period, with median retweet rates of 4485 and 11194 respectively. We labelled both original and retweeted tweets for each of our three issues of interest: Covid19, George Floyd, and Black Lives Matter. We use the following lists of string literals to match for topic relevance:

• We use two approaches for this labelling. In the first approach, we consider occurrence of the aforementioned terms in the tweet text of all tweets. For retweeted tweets, we extract the complete text of the original tweet. We call this the InText category. For the second approach, we match the terms only in hashtags of tweets. We call this the InHashtag category. In both cases, the matching is caseless and considers substrings i.e. 'covid' will match to #Covid19, #CovidPandemic, etc. in the InHashtag category, while it will match additionally to terms like 'Covid19', 'CovidCrisis', etc. in the InText category.

In general, Democrats tweeted more on a weekly basis. When considering InText tweets, they tweeted more across categories, with a greater number of mentions of Covid19, George Floyd and BLM than Republicans. The difference in Covid19 tweet counts was small, but significant. On the other hand, the difference in BLM and George Floyd-related tweets was substantial. When considering InHashtag tweets, difference in Covid19 hashtag use by the two parties was not significant. For BLM and George Floyd, it was significant. We controlled for the user that posted the tweet, the week and the number of total tweets they posted in that week.

We plotted the partisan divide in framing by the US Congress on the three key topics using two methods -by estimating how many and which politicians posted about those topics, and by analysing co-occurring terms in tweets about those topics. For the former, we plot the number of tweets posted by politicians and the median retweets they received in three interactive visualizations 2 . In these visualizations, the x-axis plots the number of tweets (for InText category) posted by a member of US Congress, y-axis plots the median retweets they received, and the size of the bubble corresponds to the number of their Twitter followers. Figures 1, 2, and 3 show excerpts from these with the top ten most followed and top ten most active 3 

The parties diverge substantially on all three issues of interest. Firstly, on COVID-19, we see that Democrats characterize the crisis through narratives emphasizing health care and public health, which is consistent with the party's recent, albeit tumultuous, shift towards prioritizing health care in their political platform. Republicans, on the other hand, highlight the effect of the shutdowns on the economy through an emphasis on small businesses, possibly building a narrative to justify re-opening, a recent political priority. This can be seen in context of the Trump administration's concerted efforts to push state and local governments to reopen businesses and schools. We also note that Republicans consistently mention China in their tweets about COVID-19, which is consistent with the party's recent anti-China rhetoric.

On the issue of George Floyd's murder, we observe that both parties mention his name alongside words like justice and police. However, while Democrats linked this murder to the killings of other Black individuals, and specifically named police brutality, we see that Republicans speak much more of protests and violence. Delving into the Republican discussion around 'violence', we see that, in the two weeks immediately after the murder, Republicans posted 272 tweets that included the term 'violence'. Seventy-three of these referred to Floyd by name. These tweets largely included condemnations of the murder, but would then pivot to criticizing perceived rioting, violence and looting. In the three subsequent weeks, there were 154 tweets about violence, but only six referred to Floyd by name. This indicates a shift in narrative, away from an emphasis on Floyd's murder and towards a wider discussion of systemic racism and American heritage, including the defacing and removal of Confederate symbols and statues.

With regards to Black Lives Matter, we see that there are a number of Democratic tweets containing the hashtag, but no significant co-occurrences .This may point to the lack of a clear agenda on how to address or respond to calls for anti-racist action.While there are insufficient Republican tweets to draw meaningful conclusions. This in itself is suggestive, indicating that the phrase in the hashtag #BlackLivesMatter is not one with which Republicans want to associate.

Finally, we analyzed tweets with terms 'defund' and 'police' to examine the Republican position further. Of the 321 tweets posted after May 25 that contained these terms, 307 were posted by Republicans, either criticizing the growing '#DefundThePolice' campaign, defending law enforcement, or criticizing Democratic opponents for supporting police defunding. The only tweets by Democrats that contained these terms appeared to distance themselves from the campaign -defending themselves from allegations by Republicans that they were in favor of it, or bluntly stating that they were against defunding the police. One exception was a tweet by Congresswoman Rashida Tlaib in full support, made significant by its rarity. Thus, we see that politicians at large, across both parties, did not show support for this campaign, despite its centrality in the anti-racist discourse. 

Virtual uprisings: On the interaction of new social media, traditional media coverage and urban space during the âĂŸArab SpringâĂŹ

The afd: Finally a successful right-wing populist eurosceptic party for germany?

Far-right movements in France: The principal role of Front National and the Rise of Islamophobia

Social media use and participation: A meta-analysis of current research

The dark side of the web: Italian rightwing extremist groups and the Internet

Twitter as arena for the authentic outsider: exploring the social media campaigns of Trump and Clinton in the 2016 US presidential election

Framing in Social Media: How the US Congress uses Twitter hashtags to frame political issues

Obama and the power of social media and technology

The social media president: Barack Obama and the politics of digital engagement

Role of the new media in the Arab Spring

Acting in the public sphere: The 2008 Obama campaign's strategic use of new media to shape narratives of the presidential race

Social media and the 2008 US presidential election

Locating the âĂIJInternet HinduâĂİ Political Speech and Performance in Indian Cyberspace

How Obama really did it

Right-wing populism in Europe & USA