Information Discovery and Delivery

InformationTwitter Dialogue: Discovery An analysis of Pakistani and Politicians’ Delivery Information Sharing

Journal: Information Discovery and Delivery

Manuscript ID IDD-03-2020-0031.R5

Manuscript Type: Original Article

Twitter, Information sharing, Politicians, Political communication, Social Keywords: media,

http://mc.manuscriptcentral.com/ilds Page 1 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 8 Information Discovery and Delivery 9 10 Twitter Dialogue: An analysis of Pakistani Politicians’ 11 Information Sharing 12 13 14 Abstract 15 Purpose: The use of Twitter by political parties and politicians has been well studied 16 in developed countries. However, there is a lack of empirical work which has 17 examined the use of Twitter in developing countries. The present study aims to 18 explore the information sharing patterns of Pakistani politicians through Twitter 19 accounts during the pre-election campaign of 2018. 20 21 Design/Methods/Methodology: Data of three weeks of the official party accounts 22 and the politicians running for prime minister was analysed. The mixed methods 23 approach has been employed to analyze quantitative and qualitative data retrieved 24 through Twitonomy. 25 26 Findings: It was found that the most active Twitter account belonged to the winning 27 party. The prominent Twitter account functions were call to vote, promotional 28 Tweets, promises and tweeting about party developments. The present study 29 provides evidence that there is a difference between the tweeting behaviour of 30 established and emerging parties. The emerging party heavily posted about changing 31 traditional norms/culture/practices. 32 33 Practical Implications: The study contributed to existing knowledge and has 34 practical implications for politicians, citizens and social media planners. 35 36 Originality/Value: The present study was designed carefully and based on empirical 37 research. The study is unique in its nature to fill the research and knowledge gap by 38 adding variety of Twitter functions used by politicians. 39 40 41 42 Background 43 44 Social networking sites (SNS) have emerged as significant communication 45 platforms within 21st century society and enable citizens to connect with each other 46 from all around the world. Recent technological innovations have seen the wide-spread 47 usage and adoption of the platforms by developed and developing nations. These novel 48 communication services allow users to share personal updates, general information, 49 discussions, shopping choices, educational information, political updates and so on. 50 51 The use of social media during election campaigns is increasing worldwide and 52 politicians have successfully utilised this medium for their popularity (Casero-Ripollés, 53 2017;Lopez-Meri, Marcos-García Stieglitz and Dang-xuan, 2013; Weaver et al., 2018 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 2 of 26

1 2 3 4 5 6 7 2 8 Information Discovery and Delivery 9 ; Lee et al., 2020). To communicate on Twitter with 280 characters may seem 10 challenging, however, politicians and other groups are extensively using this medium 11 for developing online connections. Recent studies have identified significant use of 12 Twitter by global politicians, more specifically during election campaigns. It has been 13 argued that the use of digital platforms such as Twitter during the US 2008 elections 14 led Barack Obama to victory. Politicians use Twitter in a number of ways to mobilize 15 voters, share private messages, sometimes to share images to stimulate retweets 16 (Lopez-Meri, Marcos-García and Casero-Ripollés, 2017; Tumasjan, Sprenger, Sandner 17 and Welpe, 2010; Stier et al., 2018; Sun et al., 2019; Sundstrom & Levenshus, 2017). 18 Extensive use of Twitter has been witnessed globally, specifically within the USA, UK, 19 Sweden, Finland, Australia and New Zealand (Ahmed and Skoric, 2014). Most 20 recently, study evidenced that Donald J. Trump and his supporters mobilized their 21 following through Twitter and overcome traditional Republican party during US 22 elections 2016 (Bryden and Silverman, 2019). Twitter is a silent political tool of 23 Trump’s success and he utilized this to spread politicized information more efficiently 24 and widely (Hollinger, 2017). Research has also highlighted how misinformation can 25 exist around popular topics on Twitter (Ahmed, Vidal-Alaball, Downing and Seguí, 26 2020). 27 28 The focus of this paper is related to major political parties and political leaders during 29 the 2018 elections in Pakistan. In previous years corruption has plagued Pakistani 30 politics and the 2018 election is likely to have been closely followed around the world 31 including the United States which works closely with Pakistan related to the war on 32 terror. The first time Pakistani politicians used Twitter to share information and to 33 mobilize voters was during the 2013 elections. A significant reason of this exponential 34 growth of online information sharing is terrorism threats in the country. 35 36 Twitter played a significant role in engaging Pakistani youth in politics. Statistics 37 show that male population and, specifically, youth (18-24 years) is more actively 38 engaged on social media than female (Digital 2019 Pakistan, 2019). Previous research 39 has highlighted how online campaigns increased voters’ turnout (Ahmed and Skoric, 40 2014). Research has also explored political information sharing through social media 41 by educated Pakistani youth and an emerging party Pakistan Tehreek-i-Insaf (PTI) 42 played substantial role in increasing social media trafficking (Eijaz, 2013; Ahmed and 43 Skoric, 2014). Presently, there are three popular political parties in Pakistan, Pakistan 44 Muslim League Nawaz (PMLN), (PPP) and Pakistan Tehreek- 45 i-Insaf (PTI). The present study aims to analyze popular political leaders of selected 46 parties and their official Twitter account during the pre-election campaign in 2018. This 47 research project has potential for practical and theoretical implications for politicians, 48 social media policy makers and information providers. To the best of the authors’ 49 knowledge, this is the first empirical study to examine information sharing patterns 50 during the 2018 elections in Pakistan. 51 52 The overall research aim of this project is to explore the information sharing patterns 53 of Pakistani politicians through their and their parties’ Twitter accounts during pre- 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 3 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 3 8 Information Discovery and Delivery 9 election campaigns of 2018. The study will attempt to answer the following research 10 questions: which party Twitter account was 1) most active 2) most replied to 3) which 11 hashtags were used 4) most retweeted and 5) most mentioned users. Moreover, the 12 study highlighted content appeared on Twitter (functions/categories) during election 13 campaign, and, 7) Word cloud of Politicians’ Twitter accounts. The study findings will 14 highlight distinct tweeting behaviour of politicians, particularly in the context of 15 developing countries. The results will have theoretical and practical implications for 16 politicians, social media analysts and researchers. Although previous research has 17 examined the Pakistan election this study is among the first to conduct an in-depth 18 analysis of tweets sent by those running for election. 19

20 21 22 Literature Review 23 24 Many studies have been published globally on social media political communication. 25 There has been a significant amount of research which has investigated Twitter’s 26 political usage and information sharing by candidates and citizens (Ahmed and Skoric, 27 2013; Graham, Broersma, Hazelhoff and Guido van 't Haar, 2013; Graham, Jackson 28 and Broersma, 2014; Ahmed, Jaidka, and Cho, 2016). Related research highlights 29 Twitter use, usage & information sharing patterns, Twitter functions, comparison of 30 parties and political leaders and comparison of usage pattern from politicians of 31 different countries (Larsson and Moe, 2014; Graham, Jackson and Broersma, 2014). 32 Fewer studies have focused on the type of published Twitter content and how fast 33 information is spread and retweeted (Stieglitz and Dang-xuan, 2013). 34 35 Twitter as a communication tool is successful in connecting ordinary and influential 36 people (Ahmed, Bath, Sbaffi and Demartini, 2018). Twitter provides more enhanced 37 opportunities to connect with general public than traditional communication means 38 (Kim and Lee, 2020). It also becomes a source of information for journalists and other 39 media professionals (Ahmed and Lugovic, 2019). To be popular and more powerful, 40 less established and emerging political leaders use Twitter more than established 41 leaders. Some have argued that emerging leaders receive less coverage on traditional 42 media, therefore, they use new medium to mobilize their voters (Ahmed, Choa and 43 Jaidka, 2017). However, candidates’ age and gender have no impact on their usage of 44 Twitter account (Grant, Moon and Grant, 2010). Social media platforms are providing 45 opportunities to newly established parties to engage and interact with their voters 46 (Larsson and Moe, 2014; Stier et al., 2018; Sundstrom & Levenshus, 2017; Weaver et 47 al., 2018) . It has also been evidenced that Twitter usage is significantly related with 48 the victory of the winning party (Ahmed, Jaidka and Cho, 2016; Singh et al., 2020). 49 Interestingly, this study found that being straight forward and having open approach 50 distinguished Trump from other politicians on Twitter. Researchers argued that Trump’ 51 political stance and judging opponents’ capabilities made him popular and mobilize his 52 victory in US elections 2016 (Ross and Caldwell, 2019). 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 4 of 26

1 2 3 4 5 6 7 4 8 Information Discovery and Delivery 9 A number of empirical researches focus on Twitter’s function or information sharing 10 behaviour of political leaders and their parties. Most of the studies conducted content 11 analysis and quantitative analysis to highlight these aspects. Generally, politicians and 12 political parties’ Twitter accounts share campaign updates, promotional content, 13 criticism on opponents, call to vote, political news, acknowledgements, advice giving, 14 other news and party details (Graham, Jackson and Broersma, 2014; Lopez-Meri, 15 Marcos-García and Casero-Ripollés, 2017; Stieglitz and Dang-xuan, 2013;Tumasjan, 16 Sprenger, Sandner and Welpe, 2010). Frequent tweeting behaviour involves, tweeting 17 about campaigns, party details and criticism of opponents. Political retweets can be 18 associated with positive and negative sentiments. The quantity of retweeting of negative 19 messages is more than positive messages (Stieglitz and Dang-xuan, 2013). 20 21 Studies have also examined that the majority of political tweets address the general 22 public, other politicians, journalists, party activists, experts and celebrities (Graham, 23 Broersma, Hazelhoff and Guido van 't Haar, 2013; Ahmed and Skoric, 2013). It has 24 also been established that politicians’ Twitter image trigger their social media 25 campaigns more positive than their traditional communication (Lee, Lee and Choi, 26 2020) A comparative study of political candidates of Norway during two elections of 27 2011 and 2013 classified Twitter user groups into three classes, Retweeters, who 28 retweet only, Eite, who do not retweet but message their own and Networkers, who 29 retweet and tweet simultaneously (Larsson and Moe, 2014). Another study investigated 30 Twitter messages’ impact during the German federal election and found that important 31 messages could be delivered even in 140 characters and citizens are more engaged in 32 political dialogue on Twitter than other social media platforms (Tumasjan, Sprenger, 33 Sandner and Welpe, 2010). The research also identified that politicians involved in 34 Twitter propagation networks during election campaigns are more likely to win 35 elections (Sun, Ch’ng and See, 2019). The rapid growth of social media through a large 36 number of community engagements helps in identifying real world problems (Singh et 37 al., 2020). 38 39 Abundant use of Twitter in political campaigns is increasing. One of the significant 40 reasons of Twitter’s popularity is that social media may select stories which may be 41 ignored by traditional media. Studies evidence politicians’ Twitter blunders or high 42 claims influence significantly on their public image (Lee, Lee and Choi, 2020). Mainly, 43 in Pakistan individuals’ can launch free campaigns and run movements on social media 44 to engage others. Also humanitarian aspects can be highlighted more on social media 45 than traditional media (Eijaz, 2013). Few studies shed light on social media and Twitter 46 usage of Pakistani politicians during 2013 elections (Ahmed and Skoric, 2013; Eijaz, 47 2013). 48 49 A study investigated the social media usage of Pakistani political parties. The study 50 reported as a popular leader on Twitter having more followers than others. 51 And his party, Pakistan Tehreek-e-Insaaf (PTI) involved more in writing blogs and 52 using social media during 2013 elections (Ahmed and Skoric, 2013; Eijaz, 2013). 53 Another study confirmed these facts by investigating supporters’ networks through 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 5 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 5 8 Information Discovery and Delivery 9 Social Network and Semantic analysis (Khan, et. al. 2020). The study used pre-election 10 2017 dataset and found that Imran Khan (IK) is more vigilant in posting his comments 11 and popular leader and sentiments of supporters indicate more positive comments for 12 IK and neutral and negative for other politicians. The latest elections were held in 13 Pakistan during 2018 on July 25th. There is a need to examine recent election trends on 14 Twitter and popular leaders’ information sharing patterns in comparison with previous 15 elections. Previously, studies examined politicians’ popularity in terms of followers’ 16 statistics, however, analysis of content posted on Twitter is a big research gap, which 17 needs to be addressed. 18 19 20 Methodology 21 22 The methodology of the present study is mixed methods in nature, quantitative and 23 a qualitative content analysis of tweets published by Pakistani popular political parties 24 and their popular leaders during pre-election campaigns 2018 was conducted. The 25 criterion to select popular political leaders was based on the number of Twitter 26 followers they had. Three popular political parties, Pakistan Muslim League Nawaz 27 (PMLN), Pakistan Peoples Party (PPP) and Pakistan Tehreek-i-Insaf (PTI), and the 28 respective leaders for these parties were also selected. These included Imran Khan (PTI 29 leader), Shehbaz Sharif (PMLN), and Bilawal Bhutto Zardari (PPP Leader). A dataset 30 was built using Twitter data for each of the aforementioned political parties and their 31 respective leaders (6 datasets in total). Each dataset consisted of 3200 of the account’s 32 latest tweets which is the maximum that can be retrieved using Twitter’s Search API. 33 Tweets analysed in this study were published three weeks before the 2018 elections 34 including Election Day (5th July-25th July, 2018). Data was retrieved using free version 35 of Twitonomy. The free version of Twitonomy had custom dates and data modification 36 limitations, however, it is easy to use, fast and was helpful to obtain basic level data. 37 38 The first phase of data exploration involves statistical descriptive analysis of 39 Pakistani politician tweets revealing basic data about tweeting behavior, most 40 mentioned user, most used #hashtag and most retweeted users. The second data analysis 41 phase involves conceptual content analysis which can be used to construct replicable 42 and valid inferences from text (Krippendorff, 2018). More specifically, the process 43 involved reviewing the data in order to develop the initial coding scheme in line with 44 previous seminal research on Twitter (Chew and Eysenbach, 2010). The study drew on 45 code-frames developed by previous studies 46 (Lopez-Meri, Marcos-Garcia and Casero-Ripolles, 2017; Ahmed, Jaidka and Cho, 47 2016; Graham, Jackson and Broersma, 2014). Once a code frame was developed tweets 48 were then manually coded into the 14 categories. 49 50 The text used within in tweets was carefully read and reduced to meaningful and 51 specific categories and patterns. We utilized a deductive approach to fit in words, 52 sentences, phrases into already derived categories from previous studies. 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 6 of 26

1 2 3 4 5 6 7 6 8 Information Discovery and Delivery 9 The process was done carefully to avoid irrelevant and inaccurate connections. The 10 final 14 categories emerged as a result of data extraction and constant comparison with 11 already defined themes from previous studies. 12 13 14 However, few new categories were introduced based on the politicians’ and their 15 parties’ Tweet content and could not fit in any available category (Appendix). The 16 following new categories (a) Condemning terrorism acts and/or country’s situation/link 17 with country’s safety, (b) Own position, (c) Religious, (d) Energizing party workers, 18 (e) Targeting youth, and (f) Political report/Highlighting efforts for developments were 19 introduced as an addition into literature. These categories show information sharing 20 patterns of Pakistani politicians and their parties’ Twitter accounts. 21 22 23 24 Findings 25 26 Table 1 and 2 highlight tweeting statistics three weeks before the 2018 election 27 including Election Day (5th July-25th July, 2018). Tweets per day refer to the number 28 of Tweets which were sent by a Twitter account. Retweets by users refer to the number 29 of retweets which a user performed. Hashtags used by a user simply refer to the number 30 of times a hashtag was used in a Tweet. The user-mentioned statistic examines the 31 number of times a user mentioned another user. 32 33 Insert Table 1 here 34 35 The results in Table 1 indicate that the most active Twitter account in the lead up to the 36 elections belonged to the PTI which sent on average 119.62 Tweets per day. The 37 account with the least amount of Tweets per day was related to the account of the PPP 38 with an average of 31.1 Tweets. The party which used the most hashtags related to the 39 PTI account which had a total of 1821 Tweets. The party with the least hashtags used 40 belonged to the PMLN which had a total of 235 hashtags which were used in tweets. 41 The Twitter account of PTI was the account which would mention most other users 42 with 763 mentions. The account which mentioned other accounts the least was that of 43 the PMLN with 44 Tweets. 44 45 Insert Table 2 here 46 The results in Table 2 highlight that the most active Twitter user in the lead up to the 47 elections was that of Bilawal Bhutto Zardari with an average of 6.29 Tweets per day. 48 The least active Twitter user was Shehbaz Sharif whom had a total of 2.43 Tweets per 49 day. An interesting finding here is that the political leaders of parties were tweeting 50 significantly less than the official party accounts. There were no significant differences 51 in the number of users that each of the accounts were mentioning. 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 7 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 7 8 Information Discovery and Delivery 9 Insert Table 3 here 10 What stands out in Table 3 is that the most mentioned users by the political leaders 11 during the election. The most interesting aspect is that Imran Khan mentioned 12 Actor/Singer mostly in his tweets as he is one of his big supporters and mentioned 13 Imran Khan in his tweets as well. This actor/singer is emerging and mostly appears 14 within TV dramas and advertisements these days. 15 16 The Tweets data (Figure 1) also shows that Imran Khan intends to target youngsters, 17 therefore, he may have mentioned the young actor/singer in his tweets tactically. 18 From Table 3, it can also be seen that Shehbaz Sharif mentioned three users 19 ‘@youtube’, a party supporter and his niece. His niece is a popular and active leader 20 of his own party and was actively engaging in the political campaign. The party 21 supporter which is mentioned by Shehbaz Sharif, has had his account suspended by 22 Twitter. Twitter suspends accounts if they spam, if there is a security risk with the 23 account and on the basis of abusive behavior. It seems that this account was spam or 24 created by the leader or his media because the user appears as among the most 25 mentioned users. The abusive behavior towards competitors could also be a reason for 26 the account suspension. 27 28 29 Bilawal Bhutto’s most mentioned user is @YouTube. YouTube is mentioned because 30 Bilawal would often Tweet YouTube videos which were critical of opponents. 31 Shehbaz Sharif also mentioned @youtube to show some sentimental videos related to 32 his brother , who was the party leader and sentenced to imprisonment. 33 34 Twitter is emerging in Pakistan, specifically to communicate political messages. 35 Political leaders mentioned only one user in their communication, it is because 36 Facebook users are majority in Pakistan. Further analysis shows that political leaders 37 mostly used #hashtags were their party slogans, symbol and campaign hashtags 38 created by them or their media teams. 39 40 The political leaders were mostly retweeted by their own party supporters and their 41 party’s official Twitter account. The overall findings demonstrate that political leaders 42 were very much central to their Twitter account activities and sharing information and 43 mentioning only their supporters and own party’s political leader (such as niece). 44 45 46 47 Tweet Content Analysis 48 49 Figure 1 shows the total number of Tweet counts per category and the results of the 50 tweet classification will now be discussed. The 14 categories (Appendix) were a result 51 of conceptual content analysis of the Tweets of politicians and their party account. A 52 manual process was adopted to analyse single Tweet as a unit of analysis. As mentioned 53 in the methodology, the coding scheme has been adopted from literature (Ahmed, 54 Jaidka and Cho, 2016; Graham, Jackson and Broersma, 2014; Lopez-Meri, Marcos- 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 8 of 26

1 2 3 4 5 6 7 8 8 Information Discovery and Delivery 9 Garcia and Casero-Ripolles, 2017). Replies came on the Tweets, website links, video 10 links and pictures were excluded from the analysis. Only text Tweets, including 11 and English languages, were examined. Urdu language Tweets were not translated into 12 English to retain the context, only important extracted codes were translated into 13 English language. In case, the Tweet falls into two categories, the dominant category is 14 considered for analysis purpose. 15 16 However, there was a need to introduce a few new categories due to specific 17 situations in the country and religious values that had arisen such as (a) condemning 18 terrorism acts and/or country’s situation/link with country’s safety(b) Own position, 19 (c) Religious, (d) Energizing party workers, (e) Targeting youth, and (f) Political 20 report/Highlighting efforts for developments. A few categories which the present 21 study introduced were merged with their main categories, however, the present study 22 used individual categories due to their dominance in Tweets for instance, (d) 23 energizing party workers and (f) political report. 24 25 This section will discuss which and how new categories emerged from data (Appendix). 26 Interestingly, the category “condemning terrorism acts and/or country’s situation/link 27 with country’s safety” specifically emerged from the Tweets of Imran Khan. This 28 political leader has never been in the government and his party was emerging in 29 Pakistani politics. Imran Khan demonstrated his concern on current country situation 30 (terrorism) and put this responsibility on the previous government (led by Shehbaz 31 Sharif). While condemning on terrorists attacks, the other politicians showed their 32 concern about country’s present condition (Shehbaz Sharif and Bilawal Bhutto). Due 33 to the prevailing situation in the country, this specific category appeared from data. The 34 following Tweets evidence this fact: 35 36 Strongly condemn the terrorist attack on X; his convoy. There seems to be a conspiracy 37 to sabotage the 25 July elections but the people of Pakistan will not allow any design 38 intended to target these historic elections to succeed (Imran Khan). 39 40 Y family has suffered the worst terrorist violence & rendered sacrifices for 41 Pakistan…..Pray for his & others' early recovery (Shehbaz Sharif). 42 43 Devastated by continued attacks on innocent ….. must ensure this mindset is 44 defeated. Have suspended my political activities in solidarity (Bilawal Bhutto). 45 46 The second significant category was “own position,” which was derived only from 47 Imran Khan’s Tweets. Imran Khan heavily Tweeted about his personal efforts for the 48 country as he was a famous cricketer in past (Pakistan’s team’s captain who led them 49 to world cup victory in 1992). Imran Khan is also a humanitarian, who built the first 50 cancer hospital in Pakistan through donations. Therefore, he reminded the nation about 51 his personal efforts through Tweets. The following Tweet is an example of this 52 expression 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 9 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 9 8 Information Discovery and Delivery 9 By Monday night I had done 60 different jalsas in the most difficult times under serious 10 terrorist threats… (Imran Khan). 11 12 Pakistani politicians used Twitter to voice their religious beliefs, connection and trust 13 on God. Interestingly, this was also visible from Imran Khan’s Tweets. He and his party 14 has never been in government, so he demonstrated his strong belief in God that his 15 efforts will not go in vain and God will reward him this time. 16 17 To energize party workers, these politicians also used Twitter as a communication tool. 18 Imran Khan and Shehbaz Sharif most Tweeted messages for their party workers, while 19 addressing to a young supporter Imran Tweeted: “Proud of @farhan_saeed for this 20 composition for our Movement for Justice. Election 2018”. 21 22 Shahbaz Sharif Tweeted about the torture on party supporters and senior party leaders 23 and saluted to their passion as ……they were tortured without any reason & FIR 24 was also registered against them including against the senior leaders ….. 25 26 Imran Khan and his party were mostly targeting to youth of Pakistan through their 27 Tweets, assuming that Twitter must be a commonly used medium among youth. It is 28 their strategy to target youth during elections being the future of Pakistan. Khan 29 Tweeted as: 30 31 Setting a new trend as they launch PTI merchandise and encourage the youth to play 32 their part in the elections and come out to vote on 25 July (Imran Khan). 33 34 35 Politicians’ (Shehbaz Sharif and Bilawal Bhutto) tweeting behaviour also highlighted 36 their messages about efforts and developments for the country. The parties of these 37 leaders remained in government several times and Tweeted about the best projects for 38 the uplift and progress of Pakistan. They also claimed to resolve major issues. 39 40 For example: ShehbazSharif Tweeted: 41 42 Establishment of peace in Karachi, Pakistan's financial capital, was the result of 43 sustained efforts by former PM Nawaz Sharif …(Shehbaz Sharif). 44 In another tweet Shehbaz Sharif noted: 45 46 After months of hard work by the PPP government Phase 1 of S3 - 77 million gallons 47 of sewerage treated daily - is online. This is a big step forward for #Karachi and will 48 help meet its #water needs (Bilawal Bhutto). 49 50 51 Insert Figure 1 here 52 An overall finding is that courtesy messages (condolence, appreciation, 53 anniversaries) were tweeted more. Imran Khan Tweeted more of these messages to 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 10 of 26

1 2 3 4 5 6 7 10 8 Information Discovery and Delivery 9 potentially show him as a more connected and vigilant politician. He was a famous 10 cricketer, keeping in view his previous affiliation, he Tweeted: 11 12 “Congratulations to Fakhar Zaman for becoming the first Pakistani to score a double 13 century in One Day cricket” (Imran Khan). 14 15 Shehbaz Sharif also Tweeted some courtesy messages, such as: “Remembering Edhi 16 sahib on his second death anniversary today. What an iconic figure he was who taught 17 us how service of the people over…” (Shehbaz Sharif) 18 19 The second most Tweeted category was “call to vote /link with criticism of adversary. 20 Bilawal Bhutto had the most number of tweets in this category call to vote messages. 21 I ask for your support in eradicating this menace from our society. Together, we can 22 fight extremism and create a peaceful, progressive and prosperous Pakistan…(Bilawal 23 Bhutto). 24 25 However, Imran Khan mostly linked call to vote with criticism of adversary. 26 27 People of Pakistan must come out and vote tomorrow in this historic election…the 28 nation has a chance to defeat the entrenched status quo…(Imran Khan) 29 30 Bilalwal Bhutto heavily focused on posting messages about “political report/efforts for 31 development,” future “promises” and “promotion.” Shehbaz Shariif also posted about 32 promotional content. This includes messages about huge participation and 33 overwhelming response of the nation in their rallies and processions. Tweets in this 34 category include the following: 35 36 37 Party has a history of proving shelter to the homeless and I promise you that 38 following our mandate, even this time we will provide homes to those in deed (Bilawal 39 Bhutto, Translated from Urdu language) 40 41 Heartening to see the hordes of people lining the streets today in the cities I passed 42 through…(Shehbaz Sharif) 43 44 Bilawal Bhutto and Shehbaz Sharif also Tweeted more in the Personal life backstage 45 political campaign category. A big reason to Tweet such messages is that both 46 politicians belong to political dynasty and inherited family politics. Therefore, both 47 leaders focused on posting about former leaders’ sacrifices and used the names of those 48 leaders/ family members emotionally. Shehbaz Sharif also tweeted about his brother 49 who was former Prime Minister. 50 51 “People are aware of the injustice done to Nawaz Sharif… keen to take revenge on July 52 25. No corruption charges proved against NS … there is no doubt about his integrity…” 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 11 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 11 8 Information Discovery and Delivery 9 Shehbaz Sharif used Twitter to post about “campaign updates”, “promotional stuff”, 10 “energizing party workers”, and “call to vote”. 11 12 Bilawal Bhutto also used Twitter to spread messages about their “political efforts”, 13 “party position” and “promises” to work on remaining issues in future. Bilawal Bhutto 14 belonged to one of the traditional and dynasty party of the country. Therefore, he was 15 tweeting more about his family and wanted to trigger people’s sentiments by posting 16 about party position and promises. 17 18 Politicians mostly used words were highlighted (Figures 2, 3, 4) through word clouds 19 in order to see influential words used by them. Word clouds were generated using free 20 website, wordclouds.com. Word clouds were created by analyzing English language 21 tweets only. The tweets showing pictures, videos and website links were not included 22 in this process. Also articles, names of procession places, most mentioned users and 23 most used hashtags were excluded from this analysis. Only important words were 24 extracted and their frequencies were noted for this purpose. 25 26 Insert Figure 2 here 27 28 29 An overview of Figure 2 shows that this politician’ mostly used words are “martyr,” 30 “Bhutto (family name),” “Benazir (mother’s name),” “Pakistan” and “mandate.” Most 31 of his family members were killed and sentenced to death on the cost of politics. 32 Therefore, this man wanted to warm people’s sentiments about his family’ sacrifices 33 by using the words, martyr, Bhutto and Benazir. 34 35 36 Insert Figure 3 here 37 38 39 From Figure 3, it is evidenced that Shehbaz Sharif used his brother’s name (Nawaz) 40 and his family name (Sharif) in most of his Twitter messages. He also used his party 41 name (PMLN) and the word ‘Pakistan’, ‘people’ and interestingly the word ‘mother’ 42 in his Twitter posts. 43 44 45 Insert Figure 4 here 46 47 The winning party leader Imran Khan’s messages were dominant with words ‘new’, 48 ‘change’, ‘justice’, ‘youth’, ‘overwhelming’, his party name (PTI), ‘hot’ and ‘money’. 49 Imran Khan, stimulated people with the dream of new Pakistan and change in the 50 country. He targeted youth and induced people with the promise of ‘justice in new 51 Pakistan’. Similarly, he posted promotional messages to highlight that there is 52 overwhelming response from public in his processions. He also mentioned in his 53 messages that he is a man, who is still in public in hot weather and even with danger of 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 12 of 26

1 2 3 4 5 6 7 12 8 Information Discovery and Delivery 9 terrorism. His posts about that he would bring change may have given him an edge 10 over other contenders. Our results have highlighted how the content of tweets varied 11 across the different accounts studied and that the most popular types of tweets sent 12 related to courtesy-related messages. 13 14 15 16 Discussion 17 18 Our preliminary analysis revealed an insight into the frequency of tweets sent 19 by major political parties and their leader’s personal accounts during the 2018 20 Pakistan elections. Based on the initial exploration of content it was found that the 21 political Twitter account with the most activity belonged to PTI Official account 22 which was the eventual winner of the 2018 elections. It is interesting to note that 23 the winning party’s account was more active, Tweeted more and retweeted than 24 the opposing parties’ accounts. Previous studies also evidenced that there is a 25 significant relationship between increased tweeting behaviour and their winning 26 position (Ahmed, Jaidka and Cho, 2016; Khan, et.al., 2020). In contrast (Bright, 27 et.al., 2019) researchers found Twitter campaigns didn’t impact election outcomes 28 during 2015 and 2017 UK elections. 29 30 Another interesting finding highlights that emerging political parties for 31 example Pakistan Tehreek-e-insaf (PTI) Tweeted more than established parties 32 (Pakistan Muslim League Nawaz (PMLN), and the Pakistan Peoples Party (PPP). 33 Previous research has also found this to be the case (Larsson and Moe, 2014; 34 Ahmed, Choa and Jaidka, 2017; Kelm, 2020). 35 36 Another important finding relates to most mentioned users, it was found in 37 line with previous studies (Ahmed, 2016; Graham, Broersma and Hazelhoff, 2013) 38 that politicians mostly interact with political leaders, party activist/supporter and 39 celebrities. Imran Khan mostly mentioned a celebrity who was also his party 40 activist and Shehbaz mentioned a party supporter (fake account/self-managed) and 41 own party political leader (his niece). Unlike previous research, we found that the 42 two established party leaders (Shehbaz Sharif and Bilawal Bhutto) most mentioned 43 user was @YouTube. The reason for this is that they would direct their followers 44 towards videos which were critical of their opponents. The most noticeable finding 45 is that these leaders used Twitter in a very limited manner by mentioning one or 46 two users only and by using hashtags related to party slogans, symbols and 47 campaign. In contrast to present study findings, Grant (2010) explored that 48 politicians mentioned users in thousands and minimum in hundreds. 49 50 The tweeting behavior of politicians was generally in line with the results of 51 previous studies (Ahmed, 2016; Graham, Broersma and Hazelhoff, 2013; Larson 52 and Moe, 2011; Khan, et.al., 2020). However, out of 14, six (Condemning 53 terrorism acts and/or country’s situation/link with country’s safety, Own position, 54 Religious, Energizing party workers, targeting youth and Political 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 13 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 13 8 Information Discovery and Delivery 9 report/Highlighting efforts for developments) categories were introduced due to 10 the dominance of content. Graham, Broersma and Hazelhoff (2013) also reported 11 that politicians tweeted about their own position or stance. This is because of the 12 specific country’s safety situation (Terrorism), different political background and 13 religious values. In accordance with the present findings, one may conclude that 14 tweeting behaviour and information sharing patterns vary with the country’s own 15 situation and political context. Politicians facing war and conflict situations in their 16 countries mostly focused on statements related to peace (Graham, Broersma and 17 Hazelhoff, 2013; Noa, 2012). 18 19 The candidates heavily focused on posting about courtesy, call to vote and 20 promotional material, this also accords with observations reported in previous 21 studies (Noa, 2012; Ahmed, Jaidka, and Cho, 2016; Ahmed, 2016). Another 22 interesting fact is that winning party leader and his supporters established about 23 oponents as old and traditional by mentioning themselves as new/change and 24 choice of youngsters. Imran Khan’s word cloud visualization gives prominence to 25 the word ‘change’, ‘new Pakistan’ and ‘youth’. This fact is also consistent with the 26 findings reported by Ahmed (2016) about Indian politicians. It is significant to 27 highlight that both countries (Pakistan and India) have almost similar cultural, 28 social and political values. The present study’s results further support the literature 29 findings (Ahmed, 2016; Larsson and Moe, 2011; Stier et al., 2018; Xeller & Atkin, 30 2019; Ross and Caldwell, 2019).) that politicians used Twitter traditionally which 31 is very much linked with their offline activities (speeches, rallies, events, 32 meetings). However, researchers identified new and interesting findings not 33 reported in previous studies. Namely, that politicians, who belonged to a 34 political dynasty (Shehbaz Sharif and Bilawal Bhutto) heavily posted about their 35 family members (who are or were political leaders) and personal life to make 36 people sentimental about the legacy and values they have. 37

38 39 40 41 Conclusion & Implications 42 43 Our study has outlined results of a research project which sought to develop an 44 insight into political parties’ use of Twitter during the 2018 election. We found that the 45 winning party in 2018 had engaged with Twitter significantly more by sending more 46 tweets and retweets and which had a larger amount of Twitter followers. Our findings 47 are in line with previous studies which found a significant relationship between winning 48 an election and having sent the most Tweets with a larger follower base. The present 49 study also found that less established parties used more social media platforms than 50 established political parties. We also defined two classification code frames; one which 51 will derive insight into interaction habits and another into the content of tweets. 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 14 of 26

1 2 3 4 5 6 7 14 8 Information Discovery and Delivery 9 The tweeting behaviour of politicians and their party’s accounts was in line with 10 published literature, however, few categories emerged based on specific elements for 11 instance country’s safety situation, political situation, cultural and religious values. 12 One may conclude that politicians’ tweeting trends are more similar than different. 13 Another implication of these results is that politicians heavily link their offline activities 14 with their messages on Twitter. These results also suggest that there is a difference 15 between information sharing patterns of established political parties’ leaders and 16 emerging leader. This study also established the fact that the winning political party or 17 leader heavily focused on messages about courtesy, to bring change, foster 18 development of new country and devalue traditional political practices. These results 19 encourage politicians to make active participation on social media and directly answer 20 to public twitter accounts on their issues, problems and challenges. As the number of 21 tweets and retweets matter, politicians must understand that their role is extended to be 22 active, concerned and vigilant on social media platforms. Another important 23 implication of our study indicate that social media campaigns are more effective than 24 traditional ones, so information sharing through these channels enhanced visibility and 25 timely communication. The results improve our understanding that continuous tweeting 26 behaviour encouraged to have more following and likes on twitter. Our finding in 27 accordance with mentioning young activists and celebrities, those have already a good 28 number following on twitter, increased politician’s visibility and information sharing 29 circle. Targeting youngsters and famous people help politicians to easily spread their 30 words, is another significant finding which has implication for policy makers and 31 concerned stakeholders. 32 Moreover, the present study contributes to literature by adding new Twitter 33 categories and functions. The results also provide an insight into the understanding 34 about tweeting behaviour of politicians in different contexts. The results also have 35 practical implications for politicians, researchers focusing on people’s information 36 engagement with social media and social media planners and managers. 37 The current study was limited by a small sample size and examined three weeks of 38 data with a limit of 3200 tweets across the 6 datasets. In addition, study did not analyze 39 posts on other social media platforms such as Facebook. Moreover, the study was not 40 specifically designed to investigate comments on the tweets and retweeting behavior of 41 Twitter users. Advance techniques such as Social Network Analysis and Semantic 42 Analysis have not been used in this study to analyze the data. Considering this, future 43 research is needed to analyze general public Twitter accounts before election and after 44 election days/weeks to determine the real indicators of politicians’ popularity. The 45 Twitter information sharing patterns of different population groups including 46 youngsters, educationists, journalists, knowledge workers and socialists must be 47 analyzed to examine different tweeting behaviours, the results would be helpful for 48 active social media campaigns. A longitudinal study is also needed to validate the 49 present study findings. Additional methods such as interviews and surveys are needed 50 to complement the twitter analysis method and so on. 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 15 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 15 8 Information Discovery and Delivery 9 References 10 11 12 Ahmed, S., Cho,J. and Jaidka, K. (2017) “Leveling the playing field: The use of Twitter 13 by politicians during the 2014 Indian general election campaign”, Telemat. Informatics, 14 Vol. 34 No.7, pp 1377–1386. 15 16 Ahmed, S. and Skoric, M. (2014), “My name is Khan: The use of twitter in the 17 campaign for 2013 Pakistan general election”, in Proc. Annu. Hawaii Int. Conf. Syst. 18 Sci., pp. 2242–2251. 19 20 Ahmed, W., Bath, P. A., Sbaffi, L., and Demartini, G. (2018). “Measuring the Effect of 21 Public Health Campaigns on Twitter: The Case of World Autism Awareness Day”, 22 in International Conference on Information, pp. 10-16, Springer, Cham. 23 24 Ahmed, W., and Lugovic, S. (2019), “Social media analytics: analysis and visualisation 25 of news diffusion using NodeXL”, Online Information Review, Vol 43 No 4, pp 149- 26 160. 27 28 Ahmed, W., Vidal-Alaball, J., Downing, J., and Seguí, F. L. (2020), “COVID-19 and 29 the 5G conspiracy theory: social network analysis of Twitter data”, Journal of Medical 30 Internet Research, Vol 22 No 5, e19458. 31 32 Bright, J., Hale, S., Ganesh, B., Bulovsky, A., Margetts, H. and Howard, P., (2020), 33 “Does Campaigning on Social Media Make a Difference? Evidence from candidate use 34 of Twitter during the 2015 and 2017 UK Elections”, Communication Research, Vol 47 35 No7, pp.988-1009. 36 37 Bryden, J. and Silverman, E. (2019), “Underlying socio-political processes behind the 38 2016 US election”, PLoS ONE, Vol 14 No 4, e0214854. 39 https://doi.org/10.1371/journal.pone.0214854 40 41 Chew, C., and Eysenbach, G. (2010), “Pandemics in the age of Twitter: content analysis 42 of Tweets during the 2009 H1N1 outbreak”, PloS one, Vol 5 No11, e14118. 43 44 45 46 “Digital Pakistan 2019 v02” (2019), Available at 47 https://www.slideshare.net/DataReportal/digital-2019-pakistan-january-2019-v02 48 49 Eijaz, A. (2013), “Impact of New Media on Dynamics of Pakistan Politics”, J. Polit. 50 Stud.Vol 20 No1, pp 113–130. 51 52 Graham, T., Broersma, M., Hazelhoff, K., and van ’t Haar, G. (2013), “Between 53 Broadcasting Political Messages and Interacting with Voters: The use of Twitter during 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 16 of 26

1 2 3 4 5 6 7 16 8 Information Discovery and Delivery 9 the 2010 UK general election campaign”, Inf. Commun. Soc., Vol 16 No 5, pp 692– 10 716. 11 12 Graham, T., Jackson, D., and Broersma, M. (2014), “New platform, old habits? 13 Candidates’ use of Twitter during the 2010 British and Dutch general election 14 campaigns” New Media Soc. Vol 18 No 5, pp 765–783. 15 16 Grant, W. J., Moon, B. and Grant, J. (2010), “Digital dialogue? Australian politicians’ 17 use of the social network tool twitter”, Aust. J. Polit. Sci.Vol 45 No 4, pp 579–604. 18 19 Hollinger, J. (2017), “Trump, social media and the first Twitter based presidency. 20 Diggit Magazine”, available at https://www.diggitmagazine.com/articles/Trump- 21 Twitter-Based-Presidency 22 23 Khan, A., Zhang, H., Shang, J., Boudjellal, N., Ahmad, A., Ali, A. and Dai, L., (2020), 24 “Predicting Politician’s Supporters’ Network on Twitter Using Social Network 25 Analysis and Semantic Analysis”, Scientific Programming, No 1, pp 1-17. 26 27 Kelm, O., (2020), “Why do politicians use Facebook and Twitter the way they do? The 28 influence of perceived audience expectations”, SCM Studies in Communication and 29 Media, Vol 9, No1, pp.8-34. 30 31 Kim, C. and Lee, S., (2020), “Innovation vs. Normalization: Politicians’ Twitter use at 32 the early majority stage of its diffusion in the Korean assembly”, The Social Science 33 Journal, pp.1-13, DOI: 10.1080/03623319.2020.1782634 34 35 Larsson, O. A. and Moe, H. (2014), “Triumph of the underdogs? comparing twitter use 36 by political actors during two Norwegian election campaigns”, SAGE Open Vol 4 No 37 4, https://doi.org/10.1177/2158244014559015. 38 39 Lee, E. J., Lee, H. Y. and Choi, S. (2020), “Is the message the medium? How politicians' 40 Twitter blunders affect perceived authenticity of Twitter communication” Computers 41 in Human Behavior, Vol 104, doi: 10.1016/j.chb.2019.106188. 42 43 López-Meri,A., Marcos-García, S. and Casero-Ripollés, A. (2017), “What do 44 politicians do on Twitter? Functions and communication strategies in the Spanish 45 electoral campaign of 2016”, El Prof. la Inf. Vol 26 No 5, pp 795-804. 46 47 Ross, A.S. and Caldwell, D., (2020), “Going negative: An appraisal analysis of the 48 rhetoric of Donald Trump on Twitter”, Language & communication, Vol 70, pp.13-27. 49 50 Singh, P. et al. (2020), “Can twitter analytics predict election outcome? An insight from 51 2017 Punjab assembly elections” Government Information Quarterly, Vol 37 No 2, doi: 52 10.1016/j.giq.2019.101444. 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 17 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 17 8 Information Discovery and Delivery 9 Stier, S. et al. (2018), “Election Campaigning on Social Media: Politicians, Audiences, 10 and the Mediation of Political Communication on Facebook and Twitter”, Political 11 Communication, Vol 35 No 1, pp. 50–74, doi: 12 10.1080/10584609.2017.1334728. 13 14 Stieglitz, S. and Dang-Xuan, L. (2013), “Emotions and Information Diffusion in Social 15 Media—Sentiment of Microblogs and Sharing Behavior”, J. Manag. Inf. Syst., Vol 29 16 No 4, pp 217–248. 17 18 19 Sun, H. liang, Ch’ng, E. and See, S. (2019), “Influential spreaders in the political 20 Twitter sphere of the 2013 Malaysian general election”, Industrial Management and 21 Data Systems, Vol 119 No 1, pp. 54–68, doi: 10.1108/IMDS-09-2017-0409. 22 23 Sundstrom, B. and Levenshus, A. B. (2017), “The art of engagement: dialogic strategies 24 on Twitter”. Journal of Communication Management, Vol 21, No1, pp. 17–33, 25 doi:10.1108/JCOM-07-2015-0057. 26 27 Tumasjan, A., Sprenger, T., Sandner, P., and Welpe, I. (2010), “Predicting elections 28 with Twitter: What 140 characters reveal about political sentiment” in Proc. Fourth Int. 29 AAAI Conf. Weblogs Soc. Media, pp. 178–185. 30 31 Lee, E. J., Lee, H. Y. and Choi, S. (2020), "Is the message the medium? How 32 politicians’ Twitter blunders affect perceived authenticity of Twitter communication", 33 Computers in Human Behavior, doi: 10.1016/j.chb.2019.106188. 34 35 Krippendorff, K. (2018), "Content analysis: An introduction to its methodology", Sage 36 publications. 37 38 Singh, P. et al. (2020), "Can twitter analytics predict election outcome? An insight from 39 2017 Punjab assembly elections", Government Information Quarterly, Vol 37 No 2, 40 doi: 10.1016/j.giq.2019.101444. 41 42 Stier, S. et al. (2018), "Election Campaigning on Social Media: Politicians, Audiences, 43 and the Mediation of Political Communication on Facebook and Twitter", Political 44 Communication. Vol 35 No 1, pp. 50–74, doi: 10.1080/10584609.2017.1334728. 45 46 Sun, H. liang, Ch’ng, E. and See, S. (2019), "Influential spreaders in the political 47 Twitter sphere of the 2013 Malaysian general election 48 ", Industrial Management and Data System, Vol 119 No 1, pp. 54–68, doi: 49 10.1108/IMDS-09-2017-0409. 50 51 Sundstrom, B. and Levenshus, A. B. (2017), ‘The art of engagement: dialogic strategies 52 on Twitter’, Journal of Communication Management, Vol 21 No1, pp. 17–33, doi: 53 10.1108/JCOM-07-2015-0057. 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 18 of 26

1 2 3 4 5 6 7 18 8 Information Discovery and Delivery 9 10 Weaver, I. S. et al. (2018), "Dynamic social media affiliations among UK politicians", 11 Social Networks, Vol 54, pp. 132–144, doi: 10.1016/j.socnet.2018.01.008. 12 13 Xeller, J. and Atkin, D.J. (2019), "The First Twitter Handle(s) of the United States: An 14 Information Processing Perspective on Twitter use by the President of the United States 15 and Its Effect on Millennials", Schulz, J., Robinson, L., Khilnani, A., Baldwin, J., Pait, 16 H., Williams, A.A., Davis, J. and Ignatow, G. (Ed.) Mediated Millennials (Studies in 17 Media and Communications, Vol. 19), Emerald Publishing Limited, pp. 129- 18 146. https://doi.org/10.1108/S2050-206020190000019008 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 19 of 26 Information Discovery and Delivery

1 2 3 4 5 6 7 19 8 Information Discovery and Delivery 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 20 of 26

1 2 3 Tables 4 5 6 7 8 TableInformation I. Tweet Statistics Associated Discovery with Political Parties and during the Delivery 2018 Election in 9 Pakistan from 5th July to 25th July 2018 10 11 12 Political Party Tweets Retwe Hashtag User Position 13 Twitter account per day ets by used by mentioned during 14 user user Elections 15 16 PTI Official 1821 763 1st 119.62 815 17 account 18 19 PMLN Official 240 44 2nd 20 46.52 235 21 account 22 PPP Official 1456 323 3rd 23 31.1 251 24 account 25 26 27 28 Table II. Tweet Statistics Associated with Political Leaders during the 2018 Election in 29 th th 30 Pakistan from 5 July to 25 July 2018 31 Leaders of Political Party Twee Retweet Hashtag User 32 33 ts per s by used by mentioned 34 day user user 35 36 PTI Leader: Imran Khan 4 0 3 1 37 PMLN Leader: Shehbaz 9 3 38 2.43 1 39 Sharif 40 41 PPP Leader: Bilawal 200 1 6.29 5 42 Bhutto 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 21 of 26 Information Discovery and Delivery

1 2 3 4 5 Table III. Twitter accounts that were most mentioned by each of the political leaders during 6 7 the election 8 Information Discovery and Delivery 9 Imran Khan Shehbaz Sharif Bilawal 10 Bhutto 11 Zardari 12 Mostly Actor/Singer @youtube @youtube 13 14 mentioned Party Supporter 15 users Famous party leader 16 (niece) 17 18 Mostly used #awaskafaisl #pmlncampaign2018 #jeetkanishaa 19 #hashtags apti nteer 20 21 22 Users most - Party supporter Party’s 23 retweeted Twitter 24 Account 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 22 of 26

1 2 3 Figure 1. Politicians’ Tweets count per category 4 5 6 7 Personal life backstage political campaign 10 8 Information Discovery and Delivery 9 Political report/Highlighting efforts for... 20 10 18 11 Promises 12 Party position 10 13 14 Updates 15 Targeting youth 16 17 Criticism on media Bilawal Bhutto N=133 18 Promotion 19 19 Shahbaz Sharif N=51 20 Acknowledgement 10 Imran Khan N=84 21 22 Energizing party workers 23 Religious 24 25 Courtesy (Appreciation, condolence etc.) 26 Call to vote/link with criticism of... 36 27 28 Own position 29 Condemn on terrorism acts or present... 10 30 31 0 10 20 30 40 32 33 34 35 36 37 38 39 Figure 2: Word cloud of Bilawal Bhutto 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Page 23 of 26 Information Discovery and Delivery

1 2 3 Figure 3: Word cloud of Shehbaz Sharif 4 5 6 7 8 Information Discovery and Delivery 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Figure 4: Word cloud of Imran Khan 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 24 of 26

1 2 3 Appendix 4 Analysis protocol used in study with examples 5 6 Categories Description Tweets 7 8 InformationCondemning terrorism Discovery acts Politicians Tweeted andBeginning Delivery to wonder why 9 and/or country’s about current country whenever Nawaz Sharif is 10 situation/link with country’s situation in trouble, there is 11 safety (inflation/economic increasing tension along 12 crisis/terrorism act to Pakistan's borders and a 13 highlight opponents rise in terrorist acts? Is it a 14 mere coincidence? (Imran 15 Khan) 16 Own position Highlighting personal This has been the 17 efforts and sacrifices culmination of 22 years of 18 struggle. I can honestly say 19 I have given my best for 20 Pakistan. (Imran Khan) 21 Call to vote/link with Calling people to come People of Thatta! I am 22 criticism of adversary out for vote or also here for the first time on 23 relating vote call with election campaign, to ask 24 criticism of adversary you to vote for PPP. 25 (Bilawal Bhutto) 26 Courtesy (Appreciation, Messages including My condolences to the 27 condolence etc.) condolence, bereaved families suffered 28 appreciation, in terrorism. (Shehbaz 29 anniversaries etc. Sharif) 30 31 Religious Building connection Now I leave the rest to 32 with God, Prayers Allah. (Imran Khan) 33 Energizing party workers Messages to mobilize 34 party workers Our relationship is deep 35 rooted in several 36 generations. I salute all 37 those workers who stand 38 by me. ( Translated from 39 Urdu to English, Bilawal 40 Bhutto) 41 Acknowledgement Acknowledging Thank you to the people of 42 supporters’ #Gojra for coming out in 43 participation in rallies droves to support Pakistan 44 and processions Peoples Party. You have 45 shown that you stand for a 46 peaceful, prosperous and 47 progressive Pakistan. 48 (Bilawal Bhutto) 49 50 Promotion Highlighting huge Biggest jalsa (procession) 51 participation of nation ever in history, Karachi 52 in their processions hasn't seen such a large 53 and rallies jalsa in six years. (Imran 54 Khan) 55 Criticism on media Highlighting private The blatant bias of Dawn 56 media’s bias (newspaper) against PTI 57 has now come out in the 58 open. So much for Dawn's 59 neutral and liberal 60 credentials! Complete

http://mc.manuscriptcentral.com/ilds Page 25 of 26 Information Discovery and Delivery

1 2 3 farce! Full marks to Stevan 4 Sackur for exposing Dawn 5 in his BBC HardTalk 6 interview. (Imran Khan) 7 Targeting youth Energizing youth to Another first for our young 8 Information Discoveryparticipate in election andteam of volunteerDelivery activists. 9 campaign Setting a new trend as they 10 launch PTI merchandise 11 and encourage the youth to 12 play their part in the 13 elections and come out to 14 vote on 25 July. (Imran 15 Khan) 16 Updates Updates about rallies, These hurdles cannot stop 17 processions, timing us, I have reached Lahori 18 and place Gate, we will start our 19 peaceful journey soon 20 from Muslim Mosque 21 Lahori Gate right after 22 jumma prayers IA. 23 (Shehbaz Sharif) 24 Party position/details Sharing Mandate and 25 Party’s Values and We do not believe in 26 Vision making promises and tall 27 claims, but actually doing 28 meaningful welfare work; 29 the evidence of this lies in 30 the establishment of 31 hospitals of international 32 standards and other 33 medical institutions within 34 Sindh in the last five years, 35 where quality and free 36 medical care is being 37 provided. (Translated from 38 Urdu to English, Bilawal 39 Bhutto) 40 Promises Future plan and It is our history that we 41 promises always supported homeless 42 people, I promise that we 43 will act on our and we will 44 provide home to all. 45 (Translated from Urdu to 46 English, Bilawal Bhutto) 47 Political report/Highlighting Highlighting PEEF is the largest 48 efforts for developments developments when endowment fund in Pak 49 they were in under which 375,000 50 government students belonging to the 51 poorest families got 52 scholarships. 1000s of 53 them have become Doctors 54 & Engineers. 55 Scholarships worth Rs.17B 56 have been distributed. 57 PEEF will become 58 flagship project of PMLN 59 (Shehbaz Sharif) 60

http://mc.manuscriptcentral.com/ilds Information Discovery and Delivery Page 26 of 26

1 2 3 Personal life backstage Aspects of personal Motherly love: the best 4 political campaign life (i.e.family) to feeling in the world! I was 5 show love with family in Raiwind for a Jalsa 6 backstage political & just dropped in to 7 campaign see my mother at Jati 8 Information Discovery andUmra & Delivery seek her 9 blessings ( Shehbaz Sharif) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

http://mc.manuscriptcentral.com/ilds