Social Media as a Battleground for Electoral Violence among Netizens in Africa: The Case of Kenyan Youth

1 Department of Population, Reproductive Health and Community Resource Management, Kenyatta University, P. O. Box 43844 – 00100, Nairobi, Kenya; ORCID: https://orcid.org/0000-0001-71733-0498. 2 Department of Educational Management and Curriculum Studies; Department of Enterprise Computing, Mount Kenya University, P. O. Box 342 – 01000, Nairobi, Kenya; ORCID: https://orcid.org/0000-0001-8531-6212. 3 Department of Population, Reproductive Health and Community Resource Management, Kenyatta University, P. O. Box 43844 – 00100, Nairobi, Kenya; ORCID: https://orcid.org/0000-0002-9778-8839. 4 Department of Population, Reproductive Health and Community Resource Management, Kenyatta University, P. O. Box 43844 – 00100, Nairobi, Kenya; ORCID: https://orcid.org/0000-0002-2449-0564. *Author for Correspondence email: kristinenjuguna@gmail.com.


INTRODUCTION
The world is witnessing an explosion in the use of social media, especially among the youth amidst rampant political aggression. However, very little has been studied about the interactive relationship between social media and electoral violence among the youth. In Kenya, as in many sub-Sahara African countries, traditional media such as newspapers and radio do not appeal to the youth because they are viewed as opinionated and old fashioned. This has resulted to youth opting for social media use as an alternative for dialogue, as they have learnt that they can use this channel to shape discourse surrounding (dis)empowerment (Iwilade, 2013). The growing awareness of new media, as posited by Zuckerman (2007), however, could provide a platform for discussions among the public and information sharing. He contends that media tends to work best not in highly repressive nations but in moderately repressive ones. To this end, the assumption that there is high demand for social media in Kenya in the face of traditional media could be true and is expected to grow as long as there is excessive control on traditional print media.
The rising unprecedented political infractions, disturbances and electoral violence in Africa with the youth playing a significant role in determining the political discourse of their countries front the emerging trends (Ukeje & Iwilade, 2012). According to Burchard (2016), sixty percent of African countries between 1990 and 2015 have experienced electoral violence attributed to incidental and structural factors. Kenya is among countries like Burundi, Cambodia, Ethiopia, Guyana, Haiti, Sri Lanka and Zimbabwe in which the electoral process has often become a catalyst to destructive social conflicts.
Historically, pre-election violence in the form of political rival fights and destruction of campaign materials has been rampant. The Election Day often characterized by rival fights in polling stations due to perceived fraudulent results and a manifestation of Post-Election protests have been prevalent in Kenya's elections. This may be heightened by social media interactions as much as the same platform has been used to preach peace. These general patterns of perpetration and victimization informed the need for this research. The study, therefore, broadly investigated social media use and electoral violence among the youth in Kenya using two objectives that were to assess the use of social media platforms among the youth and to investigate the relationship between social media use and electoral violence among the youth.

LITERATURE REVIEW
The study was guided by the Dependency theory proposed by Sandra Ball-Rokeach and Melvin DeFleur in 1976. The theory combines the aspects of psycho analytics and social system theory to emphasize the role of the audience in a communication process. According to this theory, there is an internal link between media, audience and the larger social system. The audience learning from the real life is limited, so they can use media to get more information to fulfil their needs. Extensive use of media generates dependence relationship in the audience. Moreover, media can create a dependence relationship with target audiences to achieve their goals by using its media power. The youth participating in the electoral process create overdependence on the available social media platforms to achieve the desire of mobilizing fellow voters. However, the theory is criticized with its limitation in showing less focus on effects.
The Dependency Theory was complemented by Social Responsibility Theory drawn from "the Commission of the Freedom of Press" in the United States in 1949. The theory first appeared in the mid-20 th century when most of the developing countries and third world nations began to connect media to its social role (Ghosh, 2019). The theory states that there is freedom of media with self-regulation. The theory is pegged on the assumptions of the public being free to voice its opinion, recognition of community opinion, private ownership of media and media taking the social responsibility of the communications. The theory has been much used by media entities in creating professionalism by checking on the accuracy and truthfulness of information. The theory is, however, criticized through its avoidance of conflicts through acceptance of the public opinion. This relates to the problem under study on whether the acceptance of the demands of the youth during the electioneering period would end the electoral violence in urban suburbs in Nairobi.
Several existing works of literature show that there have been genuine peaceful means denoted by democratic electoral processes (Zuckerman, 2007;Fischer, 2002;O'Hare & Moss, 2014). However, the situations in developing countries have been dominated by political power, characterized by widespread electoral violence. For instance, the Sudanese youth, despite having limited access to the democratic electoral process, gained government attention to the effect of forcing the state to enter into the greater use of social media (Kadoda & Hale, 2015). In regard to this, in 2013, the Egyptian government had to shut down access to the Internet by its population to quell down civil 'democratic uprising' among its citizens (Yang, 2013). However, this literature fails to link the use of social media and political stability in different countries.
The literature also explains the criticism of platforms like Uwiano, PEACETXT and others that had the effect of pacifying the violence rather than providing lasting solutions to the grievances that were considered to be factors causing the people to fight (CIPEV, 2008). The gap in knowledge on the controversies surrounding the use of social media in sparking election violence is in the backdrop of the Sustainable Development Goals (SDGs) set by the United Nations. SDG number 16 calls for promotion of peaceful and inclusive societies for sustainable development, provide access to justice for all and build effective, accountable and inclusive institutions at all levels (United Nations, 2019). Thus, electoral violence in the context of increased use of social media undermines the target to significantly reduce all forms of violence and related death rates among the youth and especially in the city suburbs including Mathare slum in Nairobi.

STUDY METHODOLOGY
Guided by a descriptive correlation research design, the study objectives were addressed using quantitative approaches and primary as well as secondary data on social media use and electoral violence from journals and past research was collected. On specific approaches, the study investigated the nature and pattern of electoral violence among the youth. The study took place in Mathare Constituency, Nairobi County, Kenya. Mathare Constituency was one of the hotspots of 2007/2008 and recent 2017 post-election violence in Kenya. The 2019 National Bureau of Statistics (Kenya) census put Mathare at an approximate population of 206,564 of which 62% of them are aged between 18 and 35 years. Since this study had targeted the youth, the target population of the study was 128,069. The sampling was done using Krejcie and Morgan (1970) sampling table. According to the sampling table, when the target population is over 100000, the sample size should be 384, hence, the study had a sample size of 384.
In particular, it is a unique Sub-County in the Country with the highest population density of 68,941 persons per square kilometre, way far from the second populated Sub-County of Kamukunji with the population density of 25,455 persons per square kilometre. The constituency is known for its inadequacies in the provision of basic public services including water, sanitation, garbage collection, roads and footpaths, drainage, electricity and public lighting. Housing units are predominantly semi-permanent. Social amenities are inadequate, with facilities such as schools and hospitals unable to cope with the population demands. Key respondents to the survey were Kenyan youth electorate. Data was collected through the use of questionnaires, key informant interviews and focus group discussions. Quantitative data was collected from structured questionnaires and analyzed using Statistical Package of Social Sciences (SPSS) version 24.0. Descriptive statistics including frequencies such as percentages and measures of central tendency were generated. Inferential statistics significantly involved correlation and regression analyses. The presentation of findings was through the use of tables and charts.

Demographics
The results are based on the data collected from respondents who were youths composed of 86.9% males and 13.1% females. Other demographics were as presented in Figures 1-4 as follows.  Figure 1 shows that the highest percentage of youths who participated in the study was between 22 to 25 years representing 34.3% of the respondents. Those who were between 30 to 35 years were 26.8%, those between 26 and 29 years were 25.8% whereas those between 18 and 21 years were 13.1%. Figure 2 shows that among the respondents, 49% were single, 47% being married, 3.5% cohabiting and 0.5% divorced. Figure 3 shows that the majority of the youths (55.1%) had attained secondary education followed by 26.8% who had primary education. 16.2% had attained tertiary or university level education whereas 2% had no formal education. Figure 4 shows that a high proportion of respondents at 52.5 were self-employed, 17.7% were unemployed, while only 5.6% were in formal employment. In addition, 8.1% were casual workers, 6.6% worked in the informal sector, 3% were housewives or husbands and 1.5% were still in school.

Figure 4: Demographics of the Respondents
Moreover, the average number of years that the respondents had stayed in Mathare was about 15 years. However, many of the respondents had been residents of Mathare for 20 years while the median time was 14 years. The respondents who had stayed for the longest time in Mathare had been there for 35 years, an indication of their residency in the area since childhood, while the respondents who had stayed for the least time had been there for less than one year. The average income for the respondents was at Kshs. 13, 462.30 with the highest income being recorded at Kshs. 120,000 and the lowest earned nothing. Most of the respondents reported earning an average income of Kshs. 10,000 per month. Overwhelming 90.4% of the youths indicated that they were registered voters with the authority to demonstrate their democratic rights during elections. Only 9.6% were not registered as voters. Out of those registered as voters, 99.4% had participated in elections while a mere 0.6% of the registered respondents had never voted.

Use of Social Media Platforms among the Youth
Through the questionnaires, the respondents were asked to indicate whether they were active users of In another indicator, the respondents were asked to show their frequency of using social media during elections. The findings showed that during the period of election, most of the respondents indicated daily use of WhatsApp (63.1%) and Facebook (63.6%). It was reported that these were commonly used due to the high number of followers and subscribers. Moreover, a higher proportion of Twitter users (46%) and Google chat (38.4%) reported the use of these social media platforms daily during the election period. Conversely, one-third of Google chat respondents reported having never used Google chat during elections period as was the case of 26.8% of respondents for Twitter, and 5.6% for Facebook and WhatsApp users.
The respondents were also asked to indicate who they mostly interacted with on social media during electioneering periods. The study found that most of the youths engage with fellow youths (68%) on social media during an election period. This was followed by 27.9% who reported to commonly interacting with politicians, 7.6% with agents of leaders and 2% with friends among others in social media during an election period. According to Kamwaria et al. (2015), political leaders have taken to social media to spread propaganda as well as carry out ethnic mobilization of voters, sometimes spewing hate speech and incitement. The increased spread of hateful remarks through different media, including social media, influenced the occurrence of the post-election violence in 2007/08 (CIPEV, 2008).
Out of these groups of people, it is the interaction with politicians on social media that is likely to stimulate the youth to engage in aggressive politics and/or electoral violence as reported by 56.6% of the respondents. Moreover, 27.3% reported that it is the interaction with fellow youths that is likely to stimulate them. Agents of leaders (10.1%), bloggers (3%), and friends (1.5%), were also likely to stimulate youths to engage in aggressive politics and/or electoral violence.

Relationship between Social Media Use and Electoral Violence among Youth
The collected data that associate social media and electoral violence was analyzed using different statistical models as follows.

Descriptive Statistics of Social Media and Participation in Electoral Violence
Generally, there was an agreement that social media has limited or no influence on opinion related to matters elections. This was evidenced by the high proportions (percentage/number?) of respondents who reported that the platforms have never influenced them on electoral issues. At least 6 out of 10 youths indicated that the platforms did not influence their opinion on elections. Putting it into perspective, those who indicated that the social media has had no influence cited Facebook (60.6%), WhatsApp (64.1%), Twitter (78.8%), YouTube (77.8%), and Instagram (84.4%). However, there were small proportions of respondents whose thoughts were influenced by social media on matters related to elections. Cumulatively, 11.6% (reporting often and every time) of the Facebook users were often influenced by information from the platform on election matters. On the same note, 10.6% indicated that WhatsApp influenced their thought on matters elections. This was the case with 7.0% of Twitter users, 6.5% of YouTube users and 4.5% of the Instagram users as shown in Table 1.

Correlation Analyses of Social Media and Electoral Violence
Further analysis revealed that there were correlations between youth participation in electoral violence in Mathare and the frequent use of social media platforms. The following are the results from correlation analysis pitting the use of various social media platforms against the participation in electoral violence. .000 .000 .000 .000 .000 **. Correlation is significant at the 0.01 level (2-tailed).

Regression Analysis and Hypotheses Testing
The study was on the relationship between the use of social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) and electoral violence among the Kenyan youth in Mathare Constituency, Nairobi County. To ascertain the relationship between the use of social media and electoral violence among the Kenyan youth in Mathare Constituency, the study used correlation and regression analysis. The aggregate mean score of electoral violence among the Kenyan youth was regressed against the aggregate mean scores of individual social media. The pertinent outcome is as shown in Table 3. From the Model summary table, the outcomes revealed that social media had an important and positive effect on electoral violence among Kenyan youth in Mathare (R = .812). Social media had a strong explanatory strength on electoral violence among the youths in Mathare variations since they were found to accounts for 65.9 percent of electoral violence among the youths in Mathare change (R 2 = .659). The ANOVA outcomes of social media against electoral violence among the Kenyan youths are presented in Table 4. The ANOVA results revealed that social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) had an overall effect on electoral violence among the Kenyan youths in Mathare and the effect was statistically significant (p-value = .001). Study outcomes revealed that Social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) had an overall important influence on electoral violence since the overall p-value was .001 which was less than .05. The ANOVA outcomes showed that the F-ratio was greater than one (F = 4.392) indicating that the prediction capability of the model contribution of Social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) to electoral violence was significant. This has the implication that the model can meaningfully predict change in electoral violence among Kenyan youth in Mathare. Individual, ANOVA outcomes exhibited that influence of Social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) on electoral violence among Kenyan youth in Mathare was significant in that p-value was < .05 (p-value = .001). The coefficients outcomes for Social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) against electoral violence among Kenyan youth in Mathare are displayed in Table 5.

CONCLUSION
The study found out that social media has an important and positive effect on electoral violence among the Kenyan youths in Mathare (R = .812). Applications including Facebook, WhatsApp, Twitter, YouTube and Instagram had a strong explanatory strength on electoral violence among the Kenyan youths in Mathare variations. These applications were found to account for 65.9 percent of electoral violence among Kenyan youths in Mathare change (R 2 = .659). This means that social media accounts for 65.9 percent of electoral violence among Kenyan youth in Mathare Constituency, Nairobi County while the remaining 34.1 percent is accounted for by other factors in the study. The study outcomes also revealed that social media (Facebook, WhatsApp, Twitter, YouTube and Instagram) had an important and positive effect on electoral violence among Kenyan youth in Mathare (p-value < .05). The findings conform to the postulation of the Dependency Theory which shows the role of the voting youth in a communication process where the internal link between media, audience and large social system is depicted. Social media usage is attributed to getting more information to fulfil needs. In relation to the Social Responsibility Theory, the social media usage confirms the postulation to freedom of media with self-regulation on assumptions that the public is free to say their opinion; there is recognition of community opinion, private ownership of media and media taking the social responsibility of the communications.
The study, therefore, concludes that there is a relationship between social media and electoral violence among the Kenyan youth in Mathare. The findings conform to Becker and Copeland's (2016) revelation that there was a significant interaction between connective social media use and political interest among LGBT Americans. However, this is contrary to the findings of Goldsmith (2014) who found that there is potential for peaceful elections in many African countries including Kenya regardless of other existing factors. For Harish and Toha (2017), electoral violence varies in different ways based on the target victims, violence timing and specific locality. Pre-election violence is also likely to be within separatist areas while spreads faster after elections in ethnocommunal areas irrespective of the socialization that takes place within the population.
Based on the study outcomes that exposed that social media had an important and positive effect on electoral violence among the Kenyan youth in Mathare, this study recommends that the government should put more emphasis on handling social media during the election period since this study found out that social media (specifically Facebook, WhatsApp, Twitter, YouTube and Instagram) had a strong relationship and effect on electoral violence among the Kenyan youths in Mathare. Enforcement of regulatory laws and promotion of productive engagement as a way of handling online hatred should be encouraged by the government. This recommendation is also noted by Alkiviadou (2018).