1 The Use and Impact of Social Media: Tunisian Elections 2014Andrea Kavanaugh Virginia Tech with Hamida Skandrani University of Manouba, Tunisia and Steven D. Sheetz and Edward A. Fox Liuqing Li and Islam Harb UNAMUNAMUN University of Manouba, Tunisia

2 Media Use and GratificationPeople use different media for different purposes (gratifications) Entertainment Social interaction Information seeking During crises, information seeking is often primary purpose Reliable information difficult to obtain in authoritarian states Information seeking gratification obtained from which sources? With what effects on political efficacy and engagement? Higher political information efficacy->greater political engagement

3 Methods: Survey and Twitter DataOpportunity Sample of Students: U of Tunis, U of Manouba Survey (online and print): Round 1: February (N=183, response rate 65%) Round 2: November 2013 (N=130, response rate 38%) Round 3: February 2015 (N=156, response rate 47%) Survey Data Analysis (SPSS): descriptive statistics, correlations, factor analysis, and regression analysis Twitter Data: Extraction of Election tweets: 307,000 Analyses: LDA, Topic Modeling, Visualization No significant differences on Rounds 1 and 2, so we combined the data for analysis (N=313)

4 Questionnaire: Survey ConstructsMedia Use: Information Sources and Frequency of Use Information Reliability Information Sharing Behavior with Family and Friends Political Information Efficacy Demographics

5 Use of Information Sources during 2011 UprisingYoung, educated Tunisians perceived the Internet and social media to be more reliable than national broadcast media sources for political information Internet news, Facebook, YouTube (plus Al-Jazeera TV) Political information may have diffused broadly despite low penetration of social media because young Tunisians regularly shared political information obtained online with friends and family. High contact society, large youth population, youth use Internet more Use of reliable media and information sharing predicts higher political information efficacy -– the belief a person holds that they are politically well-informed and knowledgeable. Changes in Use of Information Sources for Tunisian Elections 2014?

6 Tunisia: Elections 2014 New Constitution: January 2014First regular Presidential elections and Parliamentary elections Oct-Dec 2014 Founding leader of secularist Nidaa Tounes Party, Beji Caid Essebsi, won in second round election for president, against Mohamed Moncef Marzouki, former Tunisian President , moderate Islamist Ennahda Party. Secular Nidaa Tounes Party also won plurality of seats in Parliament. Since 2011 Revolution: Greater press freedoms and new communication channels Neighbors: Algeria and Libya Independence (1956); Bourghiba dictatorship til 1987; Ben Ali dictatorship til 2011 Population: ~10 million (2010); Adult literacy (80%) Upper middle income country: GDP PC $4K; Youth bulge and higher youth unemployment (~30%)

7 Tunisia: Media Landscape 2010-2015Cell Phone penetration: >100% Internet penetration: Rose to 48.5% (2015) 43.8% (2013) 36.8%(2010) Facebook penetration: Rose to 85% (2015) 33.9% (2013) 17.6% (2010) Twitter penetration: Less than 1% (0.34%) 2015 Similar in 2010 (0.3%)

8 Use of Info Sources: Revolution vs ElectionsInformation Source % Used Source: Revolution % Used Source: Election Internet 99.2% 90.6 SNS/FB 98.7 90.1 TV ~ 90 ~ 92 F2F Friends/Family 84 92.9 Video Sharing Sites 79.1 80.6 Radio 29.6 82.5 Newspaper 26.4 30.3 Twitter 25 24.1 From Revolution to Election period, use of Internet went down, use of Facebook went down – use of radio way up, use of TV similar – important difference with TV is which TV station respondents watched during revolution versus elections.

9 TV Stations Watched Most Often: 2011 Uprising vs. 2014 Elections% Who Watched during Revolution % Who Watched during Elections Al Jazeera 75.6 % 15.5% TV Nationale 40.2 56.5 Al Arabiya 37.0 7.5 Nessma 33.1 57 Other (France 24) 29.9 24 Hannibal 25.4 21 BBC Arabic 7.7 5.5 El-Hiwar Ettounsi NA 68.5

10 Mean for 2011 Revolution (SD) Mean for 2014 Elections (SD)Mean Reliability of Information Sources: 2011 Revolution vs Elections Information Source Mean for Revolution (SD) Mean for Elections (SD) Video Sharing Site 1.81 (.86) 1.40 (.99) Facebook 1.79 (.82) 1.46 (1.06) Face-to-Face 1.73 (.79) 2.02 (.89) Internet News 1.59 (.76) 1.42 (.86) TV 1.38 (.88) 1.66 (.88) Radio 1.33 (.99) 1.60 (.87) Newspapers 1.01 (.93) 1.01 (.88) Government 0.90 (1.01) 1.38 (1.07) Internet and Social media reliability ratings are higher during revolution than election period. Government Sources and TV are higher during elections than revolution. Based on paired-samples t-tests to compare reliability of information sources with each other (for each survey). During revolution: F2F, FB, video-sharing sites, and Internet news sites are significantly more reliable than government sources, newspapers and radio. During elections: (F2F more reliable than all others); FB, video-sharing sites, and Internet news sites are not significantly more reliable than government sources. Substantial shift in the perceptions of the reliability of government sources for information during 2014 elections.

11 Significant Differences in Source ReliabilityPaired-samples t-tests comparing differences in reliability of sources: During revolution: F2F, FB, video-sharing sites, and Internet news sites are all perceived as significantly more reliable than government sources, newspapers and radio. During elections: (F2F more reliable than all others); FB, video-sharing sites, and Internet news sites are not significantly more reliable than government sources. Substantial shift in the perceptions of the reliability of government sources for information during 2014 elections. Internet and Social media reliability ratings are higher during revolution than election period. Government Sources and TV are higher during elections than revolution.

12 Information Sharing BehaviorSharing with friends and family the political information respondents obtained from Internet and social media: “I regularly shared information about the uprising that I obtained from the Internet (including Facebook, YouTube or Twitter) with other family members and friends” (77.9%, 64%) “Other family members and friends regularly shared information with me about the uprising that they obtained from the Internet (including Facebook, YouTube or Twitter)” (81%, 72%)

13 Information Sharing and Political Info EfficacySharing information obtained online with family and friends about both the revolution and the elections was significantly correlated with PIE. (r = .331, p = 014) Receiving political information from family and friends during both the revolution and the elections was significantly correlated with PIE. (r = .325, p = .017)

14 Internet and FB Use Predict PIEDuring both the revolution and the elections, frequency of Internet use and FB use are the strongest media use predictors of political information efficacy adj. R Sq = .25, p < .05 (revolution) adj. R Sq = .19, p < .01 (elections) Using regression analyses we tested different media sources for their prediction of PIE.

15 Twitter Topic 1: Media, Reporting, UpdatesWords used: Media, Reporting, Elections, Tunisia, Marzouki (candidate name): This topic seems to be largely about media reporting, such as, updates on status of competitions among candidates; for example: - Initial Indicators that "Essebsi" will win - "Essebsi" is leading with a score of 90% - Tentative Results for who is the winner in different cities (e.g. Sfax).

16 Twitter Topic 2: Candidates, Essebsi, MarzoukiTopic 2: Elections, Authority, Essebsi, Beji, Caid, Marzouki: This topic largely about the candidates for Presidential elections in Tunisia 2014, for example: the individual words - Essebsi - Beji - Caid > refer to "Beji Caid Essebsi" who is one of the candidates, the winner of the election for president of Tunisia in 2014 - Marzouki---> refers to "Mohamed Moncef Marzouki", former Tunisian President , one of the two top candidates, who lost in the second round of the presidential election 2014.

17 Twitter: Topic 3: Al-Nahda Movement and PartyTopic 3: Al-Nahda, Movement, Elections, Presidential: This topic appears to be about the Islamic “Al-Nahda” Movement and Political Party; for example: Al-Nahda "Uprising" referring to the Islamic party and movement in Tunisia.

18 Twitter: Topic 4: Football, Africa CupTopic 4: Football Matches, Following (Matches), Watching (Matches), Qualifications for African Cup of Nations

19 Conclusions Important shift in media use 2011-2015:TV channels: From Al Jazeera to Tunisian national private TV Government Sources: increased reliability for elections Internet/Social Media: less important by 2014 Elections Internet and FB use and information sharing with social network are strongest predictors of political information efficacy. Increased political information efficacy – known to increase political engagement – bodes well for nascent democracy in Tunisia. Twitter Topic Modeling indicates interest in and expression of politics among users during the 2014 elections.

20 Acknowledgements Collaborators:Donald Shoemaker, John Tedesco, Sunshin Lee, and Mohamed Farag Magdy (Virginia Tech) NSF (III and III )

21 Sehl Mellouli (Laval University)Acknowledgements Collaborators: Donald Shoemaker, Sunshin Lee, Mohamed Magdy Farag (Virginia Tech) Sehl Mellouli (Laval University) NSF (III and SES )