Organizational use of Social Media and its impacts on Politics: Evidence from Pakistan

,


Introduction
Social media platforms constitute the most visible and prominent social space for public communication.With phenomenal worldwide growth of internet (7.3%) and social media (13.2%) over the past one year up to Jan 2021, the cyber space is transforming human interactions in many ways.The use of social media for coordination and communication of mass protests during the Arab Spring in 2011 to its use by the protestors during the occupation of Capitol Hill building in post-2020 US elections indicate that the social media is shaping both the public perceptions and actions to a large extant.
Public and private sector organizations adopt different strategies to shape public opinion using social media and achieve their corporate interests.Past studies indicate that organizations use social media for dissemination of information, interaction with customers, service delivery and public image building.However, public sector organization's use of SM and its impacts on national politics has received no attention in extant literature.We attempt to bring this new insight into the organizational use of social media by investigating that how do a public sector organization such as the military organization in Pakistan, with deep roots in the country's economy and politics can use SM to achieve its political goals.
The debate about the social media's role in democratization is far from conclusive. .One school of thought believes that the social media advances freedom, democracy, transparency and openness.On the other hand, scholars have questioned that how these internet-enabled digital tools strengthen authoritarian structures within the state.In the post authoritarian period, how social media is influencing citizen's political behaviour and how the authoritarian legacies (the military organization in our case) are using these new-found tools to shape public perceptions, influence voting pattern and manipulate political discourse in line with their political agenda.This study attempts to answer these questions using the data from Pakistan.
Our study makes important contributions to the literature of democratization and organizational use of digital media both from the empirical as well as theoretical perspectives.First, it brings new insights about the role of SM in Pakistan's politics.Secondly, it introduces and empirically tests a new conceptual construct of Social Media Activists aligned with authoritarian structures in a state.It also explores citizen's political participation and voting patterns influenced by a public sector organization's use of SM.Lastly, it adds to the debate of online and offline political participation in a unique setting of transitional civilian rule.The study underpins the idea that while studying about the determinants of SM and politics, it's important to consider the remnants of antecedent regimes, that has largely been ignored in literature.

Literature Review
In recent years, the diffusion of social media platforms have witnessed a record growth, making it imperative to study their impact on society from different perspectives.Statistics indicate that, as of Jan 2021, World internet penetration has captured 4.6 billion people (59% of total population) with 4.2 billion (53% of total population) social media active users.The leading Social media platforms as per their usage are; Facebook (2.7 billion), YouTube (2.2 billion), WhatsApp (2 billion), Facebook Messenger (1.3 billion) and Instagram (1.2 billion).

Theoretical Work related to Organizational use of Social Media
Past studies have emphasized on the role of social media in public sector organizations for better management of crises.SM has been found a crucial tool in sharing public awareness information, activating social networks, supporting communities and reducing uncertainties during the crisis periods such as floods, earth quacks, traffic accidents and rescue operations.Business management literature has focused on organizational use of SM related to the topics such as; social customer relationship management, digital advertising, opinion mining, crowdfunding, conducting viral campaigns, sentiment analysis, e-recruitments, and SM metrics.Political dimensions of public sector organization's social media use is a missing link in literature.Public organization's use of social media is a recent phenomenon; therefore, no established theories exist in the field.Some scholars used theory of organizational impression management (IM) to study organizational use of social media.Organizational IM has been defined as "any action that is intentionally designed and carried out to influence an audience's perceptions of the organization" .Organizations are social actors engaged in enhancing their respectability and impressiveness in front of their constituents.An impression management (IM) model usually consists of two main players: an actor, who engages in an "IM behaviour" and an audience, which observes that behaviour.Literature categorizes organizational IM strategies into defensive and assertive.A defensive OIM approach focus about the concealment, downplaying, and underreporting of a negative event.While an assertive strategy highlights about the magnification of a firm's good performance.
Mohamed and Gardner established that the practice of image spoiling is also a legitimate domain of IM theory.Image spoiling occurs when an organization highlights the negatives of other entities in order to build its own impression.Past studies used text mining and sentiment analysis over SM to understand organizational IM strategies .In our case, we attempt to understand how organizations manage SM content in cyber space using IM strategies and how it influences citizen's political choices.Therefore, we add a new dimension to the application of organizational IM theory by linking the organizational impression management strategies with their impacts on citizen's political behaviour.IM explains the strategies an organization employs to manage its impression.We explore further by studying the impact of those impression management tactics on citizen's political participation.We use organizational IM theory to explore the military aligned SMA's in Pakistan.Our study contributes to theory building by investigating that, how a public sector organization can use the social media to influence the political process in a country.

Pakistan's Military Organization as an Authoritarian Legacy
The role of antecedent authoritarian regimes in shaping the nature of government in the post-authoritarian period has been recognized in academic literature.In their ground breaking volume, Hite and Cesarini introduced the concept of "authoritarian legacies" as a new analytical framework to indicate that how it offers fresh insights to the dispositions of contemporary democratic governments in Southern Europe and Latin America.Military organization receives a special emphasis in the concept for being an important actor of the past authoritarian regime.
The recent studies on regime change has paid attention to the challenges that both social elite and society as a whole has to face during transitional governments from the legacies of the past dictatorial regimes.Whitehead accepts that the institutions affiliated with dictators remain a challenge for democracy to take a root long after the dictators relinquish their power, because they refuse to change their attitudes in the new environments.When institutions of the past regime do not change their behaviour patterns and keep on meddling in politics, they mould democratic practice while staying behind the façade of a civilian rule.
Politicized militaries with their perverse influence pose huge challenges for stability of political process during the post authoritarian period.Pinto explains that the more innovative and durable an authoritarian institution, the greater influence it will cast in the post-authoritarian period.When the authoritarian legacies are embedded in the state structures such as the military organization in Pakistan, it is very hard to end their influences as the institutional path dependence process resists any enforcement of external change .Hooghe and Quintelier finds that the low levels of political participation in Eastern and Central Europe as compared to the western part of the continent even decades after the democratic transition is an important determinant of authoritarian legacies.Castaldoa and Pinnab conclude that the quality of Croatian democracy is being affected by the innovative nature of the former hybrid regime.Past literature has paid attention to the authoritarian legacies that influence democratic governments .Yet, how they incorporate social media in preserving their interests in the post-authoritarian period is largely unknown in literature.This paper attempts to bridge this gap in literature from the perspective of authoritarian legacies.

Social Media Activists (SMA's) aligned with Military in Pakistan
Pakistani military's management of media affairs have been extensive, but, surprisingly, little attention has been paid in literature about military's management of the vast spectrum of media, especially the social media.Khan and Pratt find that that the military in Pakistan strategically engages with different segments of society and uses motivational frames of communication to gain public support for its war against terrorism.Directorate of Inter Services Public Relations (ISPR) is the formal communication channel of the military.ISPR has an official twitter handle and an official Facebook page-each drawing millions of following.
Nevertheless, military also operates a vast network of covert media operations as well.Some accounts claim that the intelligence agencies co-opt print and TV journalists to toe official narrative.Others assert that the military runs social media campaigns, FM radio channels and propagates social media content to harness the public opinion in line with its political agenda.In April 2019, Facebook removed "103 pages, accounts and groups engaged in coordinated inauthentic behaviour" originated from Pakistan.These accounts had more than 2.8 million followers.Facebook's head of cyber security revealed that the people involved in this malicious activity attempted to hide their identities; however, the investigation found that they were linked to the ISPR.There are tons of fan pages aligned with military narrative on various social media platforms, and some of them drawing millions of followings.They promote similar nature of content all across them, aimed at military's image building, political commentary, and propagation of a defamatory narratives against the personalities and political parties critical of Pakistan's military.
Military's active involvement in social media is affecting public opinion in many ways.Yet, the current understanding of military's social media involvement is limited.Since 1994, the military's official media wing has produced 7 military oriented dramas, 21 documentaries and 60 songs.In addition to the TV channels, these contents are also broadcasted on Social media platforms for wider public reception.In September 2020, Facebook suspended another "103 Pages, 78 Groups, 453 Facebook accounts, and 107 Instagram accounts for engaging in coordinated inauthentic behaviour" originating from Pakistan.Malhotra argues that one of the objectives of military's narrative building among masses is winning its internal legitimacy Vis-à-Vis civilian government.
The military in Pakistan uses various techniques to shape the public opinion through social media.It runs an internship scheme for the youth as part of its outreach strategy to influence content writers on social media.There is a breed of journalists, TV anchors, columnists and opinion makers on regular and social media, who actively promote military's narrative.Military also "backs a number of analysts and political commentators that toe the military's line".Military provides lists of approved experts to news agencies for inviting them to TV talk shows for propagation of certain narrative.Some analysts claim that through the media agenda setting, military portrays politicians as incompetent and corrupt while presenting itself as a "messiah".We term these social media pages, accounts and groups across SM platforms (both directly and indirectly related to military) as Social Media Activists (SMA's); which are used for military's image building and promoting its political agenda.Literature suggests that the citizens connect with different pages, personalities, groups, and political pundits over social media "to satisfy their curiosity and add to their political knowledge"-known as SM activists within the SM space.These SM communities and political commentators shape citizen's political ideologies and lead them to a certain political (re)alignments.In the past studies these SMAs have been studied in relation to the political parties, however, we extend the discussion to the authoritarian legacies embedded within state structure and introduce a new conceptual construct called "SMAs aligned with military organization" in case of Pakistan.
Mohamed, Gardner, and Paolillo developed a 2x2 taxonomy of OIM behaviours.This taxonomy divides OIM behaviour into four quadrants of direct and indirect as well as defensive and assertive.Taking a lead, we develop a 2x2 taxonomy of how an authoritarian structure can use SMAs to propagate its narrative and organizational impression management.Authoritarian entities engage with both directly and indirectly associated SMAs.We term SM accounts, pages, and groups that are formally affiliated with the organization as directly aligned SMAs, while the indirectly aligned SMAs are the accounts, pages, and groups, which are not formally associated with military, but engaged in military image building and support the military's organizational narrative over SM.Both of these types of SMAs are used defensively and offensively as shown in Table 1.We assert that organizational IM strategies through directly and indirectly aligned SMA's-as we explain in this paper-are equally applicable to politicized militaries, authoritarian regime parties and far-right political parties.We provide a conceptual framework to study organizational image building and image spoiling activities over social media through both the direct and indirect mechanisms.

Directly aligned SMAs
Indirectly aligned SMAs Defensive Strategy(image building) (Highlighting the positive and downplaying the negative) Projection of organizational achievements or downplaying of negative events about organization through SM accounts directly linked with the organization and its leadership Projection of organizational achievements or downplaying of negative events about organization by SM content writers not directly related to the organization.Offensive Strategy(Image spoiling) (highlighting the negative and downplaying the positive) Image spoiling content over SM through the SM accounts, pages and groups directly linked with the organization and its leadership.DG ISPR's tweet of rejecting the government notification about Dawn Leaks is an example of highlighting negative of others .Army Chief's statement about government's economic performance that the "growth has picked up but the debts are sky high" is an example of downplaying the positive of others.
Image spoiling content over SM through the SM content writers not directly related with the organization.These indirectly associated SMA's may include both the real as well as the fake SM accounts used for image spoiling of others.The discovery of hundreds of accounts and pages by Facebook administration-linked indirectly with Pakistan's military-engaged in coordinated unauthentic behaviours, fall in this category.

Voting Realignments
Voters' preferences can change either suddenly or over a long period.These (re)alignments can be caused by changes in political structures, systems, ideologies or the change in national policies and leadership.Key introduced the term voting realignments for such changes in voter's preferences for political parties.The use of SM for influencing the citizen's voting behaviour is a recent phenomenon.SM possesses the capability to influence citizen's voting choices more effectively due to a two-way interaction as opposed to traditional media.
Extent research has found that political commentators participate in online political activities.These activists create and join political groups on social media to share their political opinions.They involve in political and ideological debates, discuss about legislative matters and persuade their audience to align with them. .Tang and Lee argue that the citizens have enduring relationships these political pundits and commentators.In our study, we explore this connection in the context of a politicized military and its allied political commentators in SM space, influencing voting patterns and political participation in a transitional democracy.
The broader phenomenon of connections between politics and SM is mainstream; such as SM impact on political engagement , voter turnout, or voting behaviour.However, the influence of SM on voting realignment is missing from literature as pointed out by Frimpong et al .In this study, we explore citizen's connections with SMA's aligned with a politicized military and its effect on their voting realignments in a transitional democracy.We believe that the more citizens remain connected with military associated SM activists; their voting preferences will align with political parties supported by military.It will increase the military's political influence and increase its legitimacy to indulge in roles beyond its constitutional mandate.
Hypothesis 1 : The citizen's online interaction with SMA's aligned with military is likely to be positively related with their voting realignments.
Hypothesis 2 : The citizen's online interaction with military aligned SMA's is likely to be positively related with their increased online political participation.

Online Political Participation (ONPP) over SM
There have been contrasting evidence about the impacts of SM on politics.Many studies report a positive association between ONPP and use of SM (i.e.optimists); while others provide a contradictory evidence that, the use of SM does not motivate the citizens to participate in online politics (i.e.pessimists).There are other studies indicating that those citizens participate in online politics, which are already engaged in offline politics (i.e.normalizers). .Scherman and Arriagada conclude that citizen's ONPP is positively associated with their intension to vote, however, how it affects their voting realignments is least known in literature.
Lee et al argue that the people who actively participate in online politics over SM are likely to develop more extreme political attitudes over time, as compared to those who did not use SM.Studies suggest that the structure of interaction imposed by features of the platform can lead to the political polarization.The auto generated platform algorithms present SM content based on the likings of the users, thus reinforcing their perceptions.This phenomenon of "echo chamber" effects will reinforce the previous political beliefs of the like-minded individuals strengthening their existing party affiliations.Therefore, online political participation over social media is not likely to influence their voter realignments.
Hypothesis 3 : The citizen's online political participation over social media is not likely to be associated with voting realignments, as the more political content a netizen consumes over SM, their bondage with the political party is likely to be further strengthened.

Offline Political Participation and use of SM
Offline Political participation (OFPP) is defined in literature as the permissible acts of citizens aimed to influence and/or support the selection of persons in government and/or the actions taken by the government officials .Past studies have shown a positive and significant relationship between online SM use and OFPP.Gil de Zúñiga et al observed that the readers of online political blogs showed a strong tendency to participate in offline politics as well.There are various other studies which report a positive relationship between OFPP and the use of SM for politics .Although, there are some studies that do not indicate any significant relationships among the use of SM and offline political participation, but their number is very small. .The use of social media during political campaigns, protest marches and political movements have been widely observed all across the world without distinction of political system, culture, ethnicity or beliefs .Coordination, communication and propagation of protest information using SM during the Arab spring in 2011 to the pro-democracy protest in Hong Kong in 2020 exhibit a strong relationship between ONPP and offline political activity.Based on above discussion we assume that; Hypothesis 4 : The citizen's offline political participation is likely to be positively related with online political participation over SM.

Political Disposition
Disposition is defined as "the dominant or prevalent tendency of one's spirits including natural, mental and emotional condition or mood".It refers to as a "person's intentional state of identification", that how persons intentionally identify themselves .Disposition can be understood as a belief of one's traits, mind-set, determination, inclination or tendency.
Past studies have examined the impact of SM on an individual's political disposition, such as, a person's political knowledge, and efficacy, however, we combine these aspects into a construct to understand how citizen's political disposition affects their tendency to participate in politics and the voting pattern.We assume that a person's belief that they have better understanding of political situation and policy knowledge will motivate them to participate more in politics and consequently, it will deepen their political loyalties with political parties.
Hypothesis 5 : Strong political disposition is likely to be positively related with online political participation, as the citizen's possessing strong political dispositions will participate more in politics.
Hypothesis 6 : Strong political disposition is not likely to be associated with voting realignments, as the citizens with strong political dispositions are less likely to change voting patterns based on the SM interactions.

Measures
The construct measurements are primarily drawn from the past studies and adapted to the current research context of social media and political participation in Pakistan.Five point Likert scales were used for measurement of constructs.
The construct of social media's influence on the citizen's voting realignment was measured with three variables, as to what extent the SM influenced a voter's choice of a certain candidate, motivated them to switch the candidate, and affected the voting preferences among one or more candidates.Online political participation was operationalised as three item construct to understand about respondent's participation in online politics over SM following the study of Kim and Chen and Chen, Francis and Ying .Similarly, offline political participation was also devised by four items about the extent to which the respondents participated in offline political activities as captured by the past studies such as Choi , Lee and Metzga .Social disposition was measured by three items variable following the study of Chen, Francis and Ying .The indicators reflecting the role of social media activists aligned with military in Pakistani politics was adapted from the study of Frimpong, Li, et al and modified to suit the local context.We measured the citizen's engagement with SMA's (both directly and indirectly) aligned with the military organization.The measurements and construct details are given at Appendix A.

Sampling and data collection
A random sampling technique was used while conducting online survey from Pakistan.A self-reported online questionnaire was generated about the respondent's opinion of the role of SM in politics.All the questions in the survey were marked as "required" to avoid the chances of missing values.The researchers used personal contacts, social media pages and groups to send out the survey link to prospective participants.Researcher ensured to send the survey link to the social media users from all the provinces of the country.Literature points out that this convenient sampling technique draws participants mostly from younger age bracket, literate and usually restricted to the economically favoured population.Considering UNDP report , 64% of the population in Pakistan is under the age of 30 years (65 percent of our sample size comes from young age bracket), and therefore, this sampling bias of young participants may not cause significant skewedness in data.Moreover, the study also pertains to the use of social media for politics; therefore, SM users are the appropriate population to collect the data.Total 508 responses were deemed sufficient to make statistical inferences.After data screening and removing significant outliers, 407 responses were selected for further statistical analysis.

Analysis technique
To investigate the hypothesized relationships, structural equation modelling (SEM) (using maximum likelihood estimation) was employed.The research model contains causal relationships at multiple layers, for example, a dependent variable at one stage becomes independent variable at the next level.SEM allows testing of multi-layered models through path analysis in a single run; therefore, it is efficient for such research models as ours.Following Straub (1989), we conducted Confirmatory factor analysis (CFA), convergent and discriminant validity tests to assess the validity of measurement model.To validate appropriateness of our model (5 factors), we compared the model fitness indices of our baseline model (5 factors) with other models and the results (Table 2) indicate that the five-factor model fits best.
We tested the statistical significance of Covariance based (CB) SEM results (bootstrapping with 5000 subsamples) using maximum likelihood estimaations, because, the bootstrapping technique allows estimation of standard errors for non-normality.Additionally, large sample sizes from 100-400 observations can also alleviate the problems associated with non-normality as pointed out by several researchers.

Descriptive statistics
The demographic profile of respondents indicate that Punjab province, which comprises of more than 50% of the country's population draws bulk of the samples (Table 3).The data further revealed that half of the respondents voted for Pakistan Tehreek-e-Insaf (PTI) during the 2018 elections.Pakistan Muslim League-Nawaz (PMLN) supporters comprised the next largest group (37%).These two political parties were winners and runner-up during the elections.The gender distribution of the sample was 55 % and 44% for female vs male, as compared to their representation of 51% and 48.7% respectively in population.

Assessment of measurement model
Before conducting factor analysis, the items showing poor loadings with their constructs were identified and eliminated.Further, the results of normality tests showed that for few items, the critical ratios (c.r) for skewness and kurtosis does not fall within the specified limits, indicating the data is not normally distributed.
Past studies indicate that the large sample size and bootstrapping method takes care of the normality issues in data.
The assessment of measurement model involves examining the indicator loadings, convergent validity, internal consistency reliability, and discriminant validity.The examination of collinearity diagnostics (Table 4) indicate that the value of VIF for all the indicators were within acceptable range (< 3).All the outer loadings of items were between 0.731 and 0.917 .The scores of Cronbach's α and Jöreskog's CR were observed between 0.795 and 0.922 indicating the construct's consistency and reliability.The values of average variance extracted (AVE) was found from 0.564 to 0.748, indicating that there were no concerns of convergent validity.The fit indices indicate that the model is a perfect fit with the data.Although χ2 test is significant (χ2 = 169.42,df = 95, p = .000),however, it is sensitive to large sample sizes, therefore the past studies have not considered it a problem .The value of χ2/df (1.73) is less than the threshold value specified for normed χ2 (χ2/df < 3).The values of GFI (0.95) and TLI (0.97) are above the acceptable limit of 0.90 and CFI (0.98) was also found above the required value of 0.94.The values of RMR (0.034) is below 0.05 and RMSEA (0.044) is below the 0.08 as suggested by the past studies.

Common method bias
In this study, a single sourced self-reported questionnaire was used for collection of data.To minimize the chances of common method bias due to "consistency motif and social desirability", we followed the measures suggested by the past studies .In the first step, deliberate measures were taken to bring clarity in questionnaire constructs and remove the chances of ambiguity.Secondly, Harmen's one-factor test was conducted to examine the common method bias.The value for principal component extracted (PCE) was found 43.48.This results suggest that the Single factor variance is on higher side due to the highly correlated variables of ONPP and OFPP, yet, it's below the cut-off value of 50, indicating that the common method bias is unlikely to be a serious concern in this study .

Model and hypothesis testing
Table 6 reports the summary of the hypothesized relationships (based on SEM Maximum likelihood estimation of 5000 bootstrapped samples), including unstandardized regression weights (coefficients B), standard errors (S.E), critical ratios (t value), the standardized regression weights (coefficient β), significance (ρ values) and results for each hypothesis.
Table 6 shows that the SMAs aligned with the military organization are significantly and positively related with citizen's online political participations (β = 0.255, ρ < 0.000) and voting realignments (β = 0.347, ρ < 0.000).This result indicates that the citizens who consume online content propagated by the military-aligned SMAs may increasingly participate in online political activities and change their voting preferences.These results support our hypothesis 1 and 2. Citizen's Online political participation is significantly and positively associated with their voting realignment (β = 0.263, ρ < 0.050).We hypothesized that the citizen's political participation on SM may not trigger a change in their loyalties in H3; however, our proposition was not supported.One possible explanation can be that the citizens who are politically active in cyberspace can interact various political ideologies including the strong presence of military's narrative in case of Pakistan, and change their voting pattern during the elections.The offline political participation has a significant and a strong positive relationship with online political participation (β = 0.556, ρ < 0.000) validating our hypothesis 4. Political disposition was found insignificantly associated with the voting realignments (β = 0.13, ρ > 0.05), while it was found positively and significantly associated with the online political participation (β = 0.251, ρ < 0.000) proving our hypothesis 5 and 6 correct.The citizens who have a strong disposition towards politics and believe on their political skills are likely to be attracted to participate in online politics and at the same time, stay loyal to the political parties with whom they are associated.Another implication of this finding is that those citizens who do not have sufficient political knowledge and possess weak political disposition will easily change their voting preferences.

Discussion
The discussion about the use of social media for politics in the context of Pakistan is largely missing from academic literature.This study analysed that how military's support present in cyber space-both directly and indirectly associated with military organization-is shaping politics in Pakistan.This research investigates the impact of social media activists aligned with military organization on citizen's political behaviour during a transitional civilian rule in a county, where, the military has been ruling for long spells directly through coup making and indirectly through its informal mechanisms.Past studies indicate that the authoritarian legacies remain relevant in the political process during post-authoritarian period, but how SM is being incorporated by these authoritarian legacies embedded within the state structures to preserve their political interests is least understood in literature.Moreover, we also introduced a new conceptual construct of SMAs aligned with a public sector organization.The study makes extension to organizational IM theory, that, how a public sector organization can influence politics while conducting impression management strategies, especially when the organization is linked with the country's authoritarian past.Other evidence also supports these research results that the military in Pakistan backed PTI during 2018 election.As per the Freedom house report, the military attempted concerted efforts with the help of judicial elements to bring PTI into power.These results also confirm the views of other scholars who studied military's traditional role in electoral politics in Pakistan.Mazhar and Aqil Shah extensively examined that how the military in Pakistan manipulates political process through traditional means.Our study provided an empirical evidence that how the new technologies such as social media is being incorporated by Pakistani military to influence politics.
Similarly, the study results proved the "echo chamber" effect hypothesis incorrect about the relationship between ONPP and VR.Previous studies indicated that the more citizens engage in online political activities, they will strongly align themselves with their existing political beliefs due to the reinforcing algorithms of SM platforms.Our study found it otherwise.Citizens' online political participation and their connectedness with military aligned SMAs were both found positively associated with voting realignments.Possible explanations can be that the presence of military aligned SMAs spreading political narrative against some political parties and in favour of others may have caused this voting shift among citizens.The voters may have got dissatisfied with political parties in line with military narrative and changed voting preferences.These findings are reinforced by the results of 2018 elections in Pakistan, where, the incumbent political party (PMLN) lost a sizeable vote bank and opposition party (PTI) doubled its vote bank in 2018 as compared to previous election.
The offline political participation was found strongly related to online political participation confirming the results of previous studies.In Pakistan, the citizens who involve themselves in traditional offline politics also participated more in online political activities.This result confirms the normalizer's view that those people tend to participate in online politics who are already involved in offline politics.
Political disposition is considered as a citizen's self-actualization or the belief that they possess political skills.Political disposition was found to be significantly and positively associated with online political participation as the citizens who consider themselves more expert on political matters are more likely to engage in online politics.This result confirms the findings of Dimitrova, et al that the citizens who demonstrate more political wisdom are likely to participate more in online politics .At the same time, those respondents who consider that they possess political abilities did not change their voting preferences as the political disposition had insignificant relationship with voter realignments during the election.The citizens who perceive themselves better equipped with political skills remain engaged with their political beliefs.In other words, their support to the party was not influenced by external factors.Our model explains 77% variance in online political participation and 40% of variance in voter realignments.The explanatory power of our model is robust as compared to other studies in the field.

Conclusions and Future research directions
The study investigates about the use of social media by a public sector organization with authoritarian past, that, how it influences the country's politics using SMAs.The results supported our hypothesis that the citizen's connectedness with SMAs aligned with military organization positively and significantly influences their political choices.The study brings important insights about the use of social media by an authoritarian legacy embedded within the state structures.The current understanding of Pakistani military's use of SM and its political impact is inadequate despite the military in Pakistan uses a vast network of SM operations to influence citizen's political behaviours.The conceptual framework of our study is applicable to politicized militaries as well as authoritarian regime parties and the far right political parties that use defensive (image building) and offensive (image spoiling) tactics to influence public spheres.We emphasise the application of our model for entities possessing authoritarian genes, because, the image spoiling strategies against opponents using both overt and covert mechanism-as we explained in this paper-features among non-democratic forces and far right political groups.
This study provided a counter-democratic view of SM, in that, how a security sector organization linked with country's authoritarian past can use it to constrain democracy.There is an optimistic perspective also present about the use of social media and democratization research.The political parties also attempt to influence public through their SMAs.There are other groups of SMAs associated with civil society and the advocacy groups for the freedom of speech, expression and human rights, which contribute to the advancement of democracy in a country.We call for future studies about these democratic-friendly SMAs and investigate that how they are affecting political participation and voting outcomes in transitional democracies with

Figure 1 :
Figure 1: Conceptual Model of the study

Figure 2 :
Figure 2: Voting pattern by Age group

Figure 4 :
Figure 4: SEM output of maximum likelihood estimation

Table 1 :
2x2 Taxonomy of OIM strategy through SMAs by military organization in Pakistan

Table 2 :
Comparison of Baseline Model (5 factors) with other models

Table 3 :
Descriptive statistics show important findings about citizens' voting pattern in 2018 general elections in Pakistan.Past studies indicate that the young people get easily influenced by the SM as compared to the citizens of older age brackets.Figure2indicates that the younger age group (21-30 years) tilted in favour of Pakistan Tehreek-e-Insaf (PTI) (55%) in their voting pattern as compared to Pakistan Muslim League-Nawaz (PMLN) (33%).Majority of the older age group (31-40 years) showed voting preferences in favour of PMLN.Figure3shows that in all the categories of educational qualifications, PTI voters had comparatively higher education levels.Demographic and Voting Party Profile of Respondents

Table 4 :
The values for AVE, factor loadings, Cronbach's Alpha (α) and CR should be above 0.5, 0.6, 0.7, and 0.7 respectively.Values for MSV should be lower than AVE for all the variables.Values for VIF should be <3.All factor loadings are significant (at P<0.05 levels) Measurement model evaluation We used Fornell and Larcker's traditional metrics (Table5) for testing discriminant validity.For all the constructs, the square roots of AVE (given in bold in diagonals of table 5) were compared with their correlations with other constructs.This test indicated that there are no discriminant problems in our structural model.Note: The values given in bold in diagonals are the square roots of AVEs, while off-diagonal values are their correlations with other constructs.

Table 6 :
displays SEM output for maximum likelihood estimations, and the R square values for two dependent variables, i.e.Online Political Participation (0.77) and Voter Realignments (0.44).Based on the study of Hair, et al , value of R 2 in our study are in the range of moderate to strong.Our model explains 77 percent variation in online political participation and 44 percent variation in voting realignments based on the hypothesised relationships.SE stands for Standard error; t stands for Critical ratio.Significance values should be smaller than 0.05 and the critical ratios should be larger than ± 1.96.Summary of findings Note: Inner Model= Path Coefficients (standardized regression weights), Outer Model= Factor Loadings, above constructs=R 2 The descriptive statistics indicate that 54% respondents of 21-30 years age group voted for PTI during 2018 elections as compared to PMLN (34%) and PPP (9%).Majority of the voters who voted for PTI were young, therefore, they were more likely to interact with digital media, as indicated by the past studies.Voters who remained more connected with military aligned SMAs were found more likely to participate in online politics and change their voting alignments.The results indicate that the military's presence on social media had positive and significant relationship with citizens' voting realignments during 2018 elections.This voting realignment influenced by military aligned SMAs swelled voter ranks of PTI during 2018 elections.PTI vote bank swelled from 7.6 million in 2013 elections to 16.8 million in 2018 elections; demonstrates that PTI was the only party that benefited this voting realignment.Other major political parties (PMLN and PPP) either lost or maintained their previous vote bank.