Framing a murder: Twitter influencers and the Jamal Khashoggi incident

ABSTRACT

Social media have played a significant role in political discourse across the Mediterranean in recent years. In this research note, we showcase the usefulness of social media data for political analysis by focusing on the main Arabic Twitter hashtag following the assassination of Jamal Khashoggi in Istanbul, October 2018. We collect a sample of almost 2.4 million tweets posted by nearly 370,000 Twitter accounts. We show that just 281 accounts drove 80% of the discourse, and that these accounts can be reliably clustered into separate ideological camps representing different social forces of Egyptian, Turkish, European, and Gulf origin, arrayed against or in support of Saudi Arabia’s regional agenda.

KEYWORDS:

• Social media
• Polarization
• Community Detection
• Process Tracing
• Mixed Methods

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Scholars should beware, however, of platform usage trends, and the relative popularity of alternative platforms. In Egypt, for example, Facebook is far and away more popular than Twitter (http://www.mideastmedia.org/survey/2018/chapter/online-and-social-edia/).

2. For the Arab Spring in general, see Steinert-Threlkeld, Mocanu, Vespignani, and Fowler (2015) and Steinert-Threlkeld (2017). For Tunisia’s ‘Jasmine Revolution’, see Breuer, Landman, and Farquhar (2015). For Egypt’s ‘January 25ʹ revolution, see Tufekci and Wilson (2012) or Acemoglu, Hassan, and Tahoun (2017).

3. For recent work using Twitter to study politics in the Mediterranean and broader MENA region, see Siegel, Tucker, Nagler, and Bonneau (2019) or Kubinec and Owen (2019).

4. Leber and Abrahams (2019).

5. To access the Twitter API, open a Twitter developer account at https://developer.twitter.com. We recommend using a Python or R package to interface with the API. The following article walks the reader through all of these steps: https://towardsdatascience.com/access-data-from-twitter-api-using-r-and-or-python-b8ac342d3efe. For further details on using Twitter data for political science research, we forward the reader to Steinert-Threlkeld (2018).

6. Lorenz curves are often used in the economics literature to highlight wealth inequality. For their application to measuring inequality on social media, see Abrahams and van der Weide (2020).

7. We use the community detection algorithm from Blondel, Guillaume, Lambiotte, and Lefebvre (2008).

8. Influencer tweets and retweets from ‘Egypt’ accounts were entirely from those either living in exile or with no disclosed location.

9. Dickinson (2016).

10. Compare, for example, the tenor of CNN’s main (English) site with the Arabic site during October, 2018. The main site carried articles such as Sam Kiley’s ‘How Saudi Arabia’s media is covering the Jamal Khashoggi disappearance,’ (19 October 2018), while CNN Arabic touched upon the Khashoggi incident more circumspectly: ‘B’ad i’afa’ihi bi ’amr maliki dimn at-tahqiq bi qadiyat Khashoggi … man huwa S’aud al-Qahtani? [After being relieved by royal decree in Khashoggi investigation … who is Saud al-Qahtani?]’ (20 October 2018a) or ‘ ’Ileykum ma n’arifuhu ‘an 3 min aladhin yashtabih bihum fi qadiyat Khashoggi [To you, what we know on 3 of those suspected in Khashoggi incident]’ (18 October 2018b).

11. Rezaian (2018).

12. Fahim (2018).

13. CNN (2018).

14. Wintour (2018).

15. Sullivan, Loveday, and El-Ghobashy (2018).

16. TRT World (2018).

17. Smith-Spark and Alkhshali (2018).

18. Mounting numbers of tweets and retweets on October 9–10, for example, were likely driven by reporting that identified a ‘hit squad’ of Saudi security officers and Crown Prince Muhammad bin Salman himself as playing key roles in the disappearance. Morris, Mekhennet, and Fahim (2018), Harris (2018).

19. Lynch, Freelon, and Aday (2017).

20. Gökçe, Hatipoğlu, Göktürk, Luetgert, and Saygin (2014).

Additional information

Funding

This work was supported by the Middle East Initiative, Harvard Kennedy School.