Social media could help improve policy outcomes by opening lines of communication between citizens and their governments and giving decision makers and their voters the chance to appreciate perspectives from all over the world.
I wanted to explore the idea that good tweeting could save lives, or just make citizen’s lives better. So last year I compared the 2017 Twiplomacy data, showing international leadership’s presence and interaction on Twitter, with World Bank indicators like the Gini coefficient and poverty headcount ratio. I found few countries with noticeably improving economic indicators following the establishment of government social media accounts. There were 5 exceptions, Ecuador being the first of them. This paper uses Ecuador as a case study to better understand the qualitative and quantitative relationship between social media and policy outcomes.
The first section reviews some policy indicators and the status of social media in Ecuador today. Sections II, III, and IV examine the data from two different angles: first statistically, through correlation and regression analysis compared to global data; then qualitatively, with an analysis of a sample of tweets from five Ecuadorian accounts, after a review of the domestic communication situation. Section V is the discussion and conclusion.
I. Brief on Ecuador
In my initial analysis of Latin American and Caribbean nations, Ecuador stood out: its accounts made it onto every “top” list while its poverty rate began to fall again after the first accounts were established in 2010 and 2011. On the other hand, declines in inequality changes were less consistent and timely. However, the data was interesting enough to warrant a closer look.
The Human Development Index is a more comprehensive measurement of a country’s development metrics with indictors including life expectancy, expected and mean years of schooling, and GNI per capita. In 2015, Ecuador ranked 89th in the world. Overall, these indicators show a populace with the literacy and means to use the communication and political tools at their disposal.
Globally, more people have access to internet and are using social media platforms every year. Statista estimates by 2019 2.77 billion people will be using online social networks. We Are Social, on the other hand, estimates the number of internet users worldwide is today at 4 billion, with social media users at close to 3.2 billion.
In 2016, Ecuador had 7 million internet users according to the BBC. We Are Social reports 13.5 million in January 2018 (80% of the population), with 11 million active social media users (66%), which grew 10% since last January. As of May 2017, 6 Twitter accounts represented institutions and personalities in Ecuadorian government, although nationally Facebook has long kept the title of most popular platform, with the exception of a brief period at the end of 2017 where Twitter was favored.
According to the 2017 Twiplomacy data, 92% of UN Member States are represented on Twitter, 88% on Facebook, 73% on Instagram, and 76% on YouTube (including foreign ministry and heads of state accounts).
II. Statistical Analysis
a. Poverty, Gini, and Accounts
While my original, relatively cursory comparison of Ecuador’s policy indicators and Twitter presence indicated a potential case study, establishing a firmer connection between policy and social media presence requires statistical analysis. Returning to the data at the global and Ecuadorian levels, I use correlation and regression analysis to determine their relationship and demonstrate that Twitter accounts had a more than random influence on the policy indicators.
For the global analysis, I first determined the most current year with the highest global volume of data points for poverty headcount ratio (at the $3.20 level) on the World Bank data site: 2014. I then made a cumulative list of active Twitter accounts, based on Twiplomacy’s 2017 data, each country had online on or before 2014.
Comparing the country account count to the poverty headcount ratio, then to the Gini Index, I found the account count to be negatively but weakly related to poverty ratio (correlation coefficient = -0.26), while there was an extremely weak and positive correlation between count and Gini (0.099). In short, there are signs that social media presence could be related to some, but not all, policy outcomes on a global level.
From there I conducted a regression analysis with account count as the independent variable and poverty and Gini as separate dependent variables. The null hypothesis for both is Β1 = 0 and the alternative hypothesis is Β1 ≠ 0, evaluated at the .10 level of significance. Key results are in the table below.
|Statistic||Count and Poverty||Count and Gini|
Based on these results, I can safely reject the count-poverty null hypothesis: the connection between account count and poverty is not random. However, the link between the number of social media accounts and inequality at the global level is again proven minor and, after this test, statistically insignificant.
I then looked at Ecuador’s data. I counted the total cumulative number of Twitter accounts each year and compared them to the poverty data going back to 1987, the first data poverty ratio data point. This data showed a stronger negative relationship between cumulative account count and both poverty (-0.76) and inequality (-0.79) than in the global sample.
I then again conducted a regression analysis with the same parameters:
|Statistic||Count and Poverty||Count and Gini|
Although with an admittedly much smaller sample size (n=18 compared to n=60 in the global sample), these results show an even stronger relationship between the cumulative number of accounts and both poverty and inequality in Ecuador. I can again safely reject the null hypotheses.
b. Voter Turnout and Accounts
In my earlier social media and policy literature review, I found multiple studies examined the impact of social media on voter turnout, with mixed results. To evaluate the possible effect in Ecuador and expand on my early analyses, I use the same methodology as above to determine the relationship between the cumulative count of Twitter accounts and parliamentary and presidential voter turnout in Ecuador. It should be noted, however, that n=10 and 9 respectively, and because of this small sample size the results should be examined with caution.
The correlation analysis showed relatively strong positive relationships between count and parliamentary turnout (correlation coefficient = .61) and presidential turnout (.69). With the regression, the relationships were confirmed statistically significant (see below). Voter turnout in Ecuador for both national level elections may be positively impacted by government’s participation on social media.
|Statistic||Count and Parliamentary||Count and Presidential|
There are, of course, numerous confounding variables that must affect the relationship between social media presence and policy outcomes, as well as voter turnout. Countries well represented on social media could be more likely to have consistent internet access, to be wealthy, and have lower poverty rates, regardless of how government and political personalities’ social media accounts are used. Additionally, poverty rates have generally been falling for decades and voter turnout has been rising since 2002, prior to the establishment of the accounts or the prevalence of social media. While the factors that have led to the decline and rise respectively could have been amplified by social media, the trends existed prior to the independent variable, weakening causal claims.
The analysis itself suffers from several weaknesses, including 1. small sample sizes, 2. the focus on a single social media platform (and not the most popular in the country), 3. the use of a single measurement of social media presence and activity, and 4. the use of so few example policy outcomes. For the statistically inclined, this demonstrates an obvious opportunity for further research.
In this paper, however, we will take a more qualitative route. Next, we will look more closely at how the accounts have reportedly been used in order to better flesh out their relationship to policy.
Ecuador makes an interesting case study in part because when imagining a country where social media could have played a role in reducing poverty, you might expect a country with a healthy appreciation of free speech. That has not been the case in Ecuador, a country rated “partly free,” and whose press was rated “not free,” by the Freedom House in 2017.
The Twiplomacy social media data was collected during the presidency of Rafael Correa, who left office in May 2017. During his presidency, Ecuador’s press freedom score rose (on a scale where 100 is worst), due at least in part to the president’s, and his administration’s, combative stance towards the press and political opposition.
Correa’s 2013 Communication Law was one of his administration’s main weapons in repressing the press and dissent. Its use for that purpose has been followed with alarm by NGOs and international organizations. Ley Orgánica de Comunicación exists “develop, protect and regulate…constitutionally-established communication rights,” but the language places responsibilities on media and gives rights to Ecuadorians without outlining specific requirements or scope, making it difficult for media to abide.
|Article 7||La información o contenidos considerados de entretenimiento, que sean difundidos a través de los medios de comunicación, adquieren la condición de información de relevancia pública, cuando en tales contenidos se viole el derecho a la honra de las personas u otros derechos constitucionalmente establecidos.||Information or content considered as entertainment, disseminated through the media, earn the status of information of public relevance, when such content violates the rights of people’s honor or other constitutionally established rights.|
|Article 21||El medio de comunicación será solidariamente responsable por las indemnizaciones y compensaciones de carácter civil a que haya lugar, por incumplir su obligación de realizar las rectificaciones o impedir a los afectados el ejercicio de los derechos de réplica y de respuesta ordenados por la Superintendencia de la Información y Comunicación, previo el debido proceso, y que han sido generadas por la difusión de todo tipo de contenido que lesione derechos humanos, la reputación, el honor, el buen nombre de las personas y la seguridad pública del Estado, de acuerdo a lo que establece la Constitución y la ley.||The media will be together responsible for the civil indemnities and compensations that may arise due to failing in their obligation to perform the rectifications or impede those affected their right of reply and of response ordered by the Superintendence of Information and Communication [SUPERCOM], prior to due process, and which have been generated by the dissemination of all types of content that might harm human rights, the reputation, the honor, the good name of the people and the public security of the State, in accordance with that which the Constitution and law establish.|
|Article 22||Todas las personas tienen derecho a que la información de relevancia pública que reciben a través de los medios de comunicación sea verificada, contrastada, precisa y contextualizada.||The people have the right that the relevant public information they receive through the media be verified, checked, precise, and contextualized.|
SUPERCOM, the enforcement body behind the law, has been variably severe. Government-critical social media activity of one group, Fundamedios, has been monitored and compiled as evidence for its dissolution by Ecuador’s Communication Ministry. Using private firms to root out critical YouTube videos and tweets, the government has filed complaints against dozens of accounts. Policing traditional media has resulted in dozens of fines, reprimands, and forced public apologies. Even the United Nations has spoken up against how the law has been applied.
Censoring, or creating the threat of external censorship of, written speech is not been the only way the Ecuadorian government has policed criticism and dissent (see https://www.hrw.org/news/2015/11/10/ecuador-crackdown-protesters). Correa has also used his personal Twitter account to rail against dishonest media, even since the end of his term.
La manipulación mediática hace más daño que la bomba atómica, porque destruye los cerebros (Noam Chomsky)..@MashiRafael @MashiRoberto @frank_dicee @Efrancoloor @PatricioMery @NandoRevolucion @SaraUbillus @AlexisSanchezEC @EmanLucho @Alfari57 @JorgeGlas @luchonGye @MaraJtiva pic.twitter.com/XTYJ6lBL6g
— Felú Oveja Negra (@FeliciaGarzon) April 30, 2018
Current president Lenín Moreno, Correa’s former vice president, has largely upheld the laws and practices of the previous administration. He has, however, spoken in support of free speech, saying in 2017, with respect to the Communication Law, that he believes “freedom of expression should stretch much more than what the Law includes” (“Yo creo que la libertad de expresión debe estirarse bastante más de lo que dice inclusive la Ley”) and opened the door to possible changes in the law. According to Fundamedios, in 2017 under Moreno attacks on the media fell by more than 50%. The differences between Moreno’s and Correa’s approaches are evident in their tweets.
IV. Account Analysis
I read and coded the tweets of Ecuador’s five active and verified accounts listed by Twiplomacy’s 2017 data set, including both Moreno’s and Correa’s. Their 50 most recent tweets as of April 30th, 2018 were coded for the policies, events, and other accounts they mentioned; direction of communication (ex. government to government); emotion and purpose; and basic interaction stats. The accounts examined in this section are @Lenin, the personal account of President Lenín Moreno; @MashiRafael, the personal account of former President Rafael Correa; @Presidencia_Ec, the institutional presidential account; @GuillumeLong, the personal account of the Minister of Foreign Affairs under President Correa; and @CancilleriaEC, the institutional foreign ministry account. Below is a side-by-side comparison of some of their basic account statistics according to the 2017 Twiplomacy data.
|Name||Creation||Total Tweets||Followers||Following||Tweets/day Since Creation||Interaction Rate|
|Lenín Moreno||29 OCT 2012||965||295,301||72||3||0.52%|
|Rafael Correa||29 OCT 2010||12,923||3,116,063||9||7||0.09%|
|Presidencia ECU||04 JAN 2010||46,060||1,414,041||35,228||32||0.01%|
|Guillaume Long||19 SEPT 2011||7,873||40,602||1,083||3||0.53%|
|Cancilleria Ec||30 OCT 2010||45,769||166,047||5,430||29||0.01%|
An important caveat to this data is that these tweets were not posted in the timespan included in the 2017 Twiplomacy data (from account creation until May 2017) and the change in administration since that time undoubtedly had a major impact on how the accounts are being used. Therefore, these observations cannot conclusively describe how these accounts interacted with the site, their network, and, by extension, policy, during that period. This analysis is intended only to describe the current tone and use of accounts, which can help shed some light on how they may have been used during that period.
a. The Current Presidency: @Lenin and @Presidencia_Ec
Each account is unique in style, attitude, and purpose. @Lenin features primarily original posts and content, rather than retweets, shares, or links to of external media. It is often positive and upbeat, mentioning “unidad” (unity) in 12% of his tweets and taking a celebratory tone in about 40%, holding up examples of Ecuadorian’s success or community events.
¡Nuevamente un compatriota nos llena de orgullo! Nuestra más sincera felicitación al ciclista carchense @RichardCarapazM, ganador de la edición 61 de la Vuelta a Asturias en España. pic.twitter.com/nLY2RmzgEz
— Lenín Moreno (@Lenin) April 29, 2018
Each day, a few posts document the president’s daily activities, from interactions with dignitaries to meetings with teachers. There are not direct mentions to political opposition, and the only mentions of the press are in #NosFaltaTres (“we are missing three”) posts. Overall, the account, although a personal one, sounds like the peppy, self-contained public face of an active politician and national leader.
The institutional presidential account primarily retweets @Lenin and @ComunicacionEC, the account of the communication secretary, with no commentary. Without editorializing, the accounts’ emotional tone is acknowledgement or reporting, and its dual purpose seems to be to inform and amplify the presidential message to its more than 1.5 million followers (compared to Lenin’s less than 700,000). It tweets on the same events as @Lenin, but with greater volume, sometimes live tweeting presidential quotes and pictures. This happened with three different events in the 50-tweet span, 70% of the account’s analyzed tweets.
The account shares presidential decrees and includes links 22% of the time, making the page a resource for those following presidential actions. This behavior may play role on the efficacy of tweets in policy change- official accounts share law changes, lend transparency to government action, and make leadership’s agenda more accessible. The live tweets are an excellent example of agenda transparency.
b. The Former Administration: @MashiRafael and @GuillumeLong
In contrast with the current administration’s accounts, former president Correa’s tweets are more about his opinion than transparency or community building. This may be, in part, a luxury of being out of political office. 38% of his tweets contain statements against political opponents or policy (66% of tweets of this nature in the whole data set). 37% support a political figure. None of the current administration’s accounts make these kinds of statements. Also, his tweets make up more than half of the anti-press posts in the dataset, 16% of his own 50.
Half the time Correa’s tweets are frustrated or angry, sharing his or the original tweeter’s disapproval of certain actions or policies. Most often his tweets are simply sharing his opinion (40%) while 28% inform, with news or additional information. He primarily in retweets instead of posting original content (84%). His retweets and comments remarking on foreign or domestic leaders and issues complement similar mentions from other political figures.
Nuestra solidaridad con el hermano @MashiRafael, víctima de una persecución política, que como en Brasil y Argentina, usa golpes judiciales y procesos penales para defenestrar a líderes socialistas que cometieron el delito de trabajar en beneficio de los más pobres y excluidos. pic.twitter.com/FcPUfNgxDo
— Evo Morales Ayma (@evoespueblo) April 11, 2018
Like Correa, Guillume Long is no longer in a government role and he was a representative of the previous presidency. Accordingly, his tweets are supportive of Correa, his policies, and his legacy, but his tweets’ purpose, presentation, and emotion have likely changed since he was Foreign Minister. Some of his tweets are critical of the media or political opponents, and explicitly support Correa and/or defend against “attacks” against him. He is second only to Correa in his tweets against political opponents (20%) and against the press (20%) and outpaces him in support of a political figure (38%).
Comunicado | En honor a la veracidad de los hechos y al adecuado manejo de las fuentes periodísticas, señores de la revista digital @4pelagatos4 a continuación les presentamos nuestra respuesta oficial en relación a los dos (2) artículos publicados en su revista digital. pic.twitter.com/yX1SJrQssD
— INPC Ecuador (@INPCEcuador) April 6, 2018
Understandably, Long disproportionately comments on foreign affairs compared to the other accounts (ex. his tweets on the Colombian peace process or refugees make up 73% of the total tweets on the subject) and also more frequently mentions NGOs, press, academics, and politicians.
c. A Questionable Contribution: @CancilleriaEC
How could a foreign ministry improve domestic policy outcomes? An effective foreign ministry, and its social media, could smooth immigration processes and make the country more enticing for potential expatriates, leading to positive immigration and its possibly positive effects (ex. Kerr and Kerr, 2011). It could negotiate domestically beneficial trade outcomes, which could improve policy indicators, or encourage the development of processes and institutions to do the same.
While it is hard to say if the Cancilleria account is intended to support such objectives, it is trying to draw people to Ecuador. 14% of its 50 tweets promote travel to or around Ecuador and feature the beautiful locations in the country.
Much of the account’s other material is retweets of other government or politician accounts, including some of the communication ministry’s presidential live tweets. CancilleriaEC itself livetweeted an interview with María Fernanda Espinosa, current Minister of Foreign Affairs. Like the presidential account, these tweets are probably used to amplify her message and give community-building insight into her agenda.
It is not immediately clear what role these accounts play in improving policy outcomes. Although they could be inspiring domestic debate and promoting transparency, they do not generally encourage action of their readers (7%) or seek out citizens’ counsel. They instead disseminate information: their function is to inform more than half the time, a rate similar to that found by studies in the literature review.
Overwhelmingly, tweets from these accounts are directed at citizens (95%) and/or other government bodies or personalities (57%). As shown in the chart below, the accounts retweet and respond to each other and other politicians. The network created by those political thinkers may contribute to policy changes.
|Account||Average likes||Average retweets||Average comments||Primary type||Most RT’d account||Most mentioned account||Most mentioned policy subject||Days|
|Lenin||704||765||305||Original (96%)||ASotomayorEcu (50%)
|MashiRafael||550||424||64||RT (84%)||ErwinJarrin (9%)||MashiRafael (29%)||CPCCS decision (30%)||5|
|Presidencia_Ec||253||384||40||RT (86%)||ComunicacionEC (88%)||Lenin (95%)||Crime (45%)||5|
|GuillumeLong||319||322||90||RT (53%)||EvoesPueblo (11%)
|LulapeloBrasil (10%)||Economic (42%)
Colombian peace/refugees (42%)
|CancilleriaEC||137||229||40||RT (60%)||ComunicacionEC (60%)||Lenin (43%)||Health (39%)||3|
The most popular policy topics are crime, including narcotrafficking and border control, and the economy, including labor and trade issues. As might be expected, the most often mentioned events are government functions, like press conferences, which are included in 47% in tweets that mentioned events.
The data set brings up questions on the role of former political leadership in shaping the tone of national debate and potentially policy outcomes, as well as the role of externally-facing accounts on the same issues.
V. Discussion, Areas to Explore, and Conclusion
Statistical analysis indicates a relationship between governments’ presence on social media and at least some policy outcomes, but that relationship is complicated. The same factors leading to an increased presence on social media could contribute to reducing poverty, for example. Based on a review of tweets from Ecuadorian accounts, a causal relationship may be rooted in the communication networks between institutions and political leaders, in the role of those accounts in informing citizens and maintaining transparency, and/or in the ability for citizens to better understand political agendas.
Given the sometimes antagonistic environment towards opposition both legally and on social media in Ecuador, its believable that Ecuador’s positive policy indicators are disconnected from the theorized benefits of social media. Suppressive policies and administrations have endangered effective citizen to government communication.
At the same time, political participation, in the form of voter turnout, is up and correlated with the number of government Twitter accounts. Citizens are still engaging with their government in the loudest way possible and that may be due in part to politicians’ accessibility. However, there is little “get out the vote”-type language in the examined tweets. Once again, the impact of social media on the outcome is unclear qualitatively.
Additionally, Ecuador experienced some historical policy changes in recent years, namely a new constitution in 2007, constitutional amendments in 2015, and the Bono de Desarollo Humano cash transfer program, which started in 2003. These changes may have had an impact, with or without social media’s influence, on the country’s policy outcomes. The relationship between social media and these policies will need to be explored in a more substantial historical analysis of social media across multiple platforms and more accounts.
The true power of social media in policy change may be driven by the characteristics of the network rather than of any account or group of accounts. This study has shown that verified political Twitter accounts communicate with and about each other, with foreign political and civic figures, and the public, both their own citizens and foreign readers. Their network(s), plus the role of unverified political and activist citizens’ accounts, needs to be explored more deeply.
The research presented here is limited in several ways, most importantly: 1. that the sample sizes are small for statistical analysis and 2. that the sample is small and not from the optimal timeframe for the social media analysis. The latter could be easily rectified by an interested researcher with the budget to spend on access to historical Tweets, and I very much invite them to do so.
There are a few other interesting gaps which will need to be addressed with further research. First, although Spanish is commonly spoken in Ecuador, many other indigenous languages are used to communicate, even if they, or their speakers, are not well represented online. Cultural, racial, and ethnic differences impede accessibility and complicate policy. Further research should look specifically into the impact of Ecuador’s diversity in its social media landscape as well as general government responsiveness to relevant issues for those communities.
As noted before, only Twitter was included in this study. Other researchers with access to the data and tools would do much to help illuminate the connection between social media and policy by studying other platforms. A step further would be to study messaging apps in this context, especially Whatsapp.
Regardless of the gaps and opportunities, this study has taken a step at establishing and evaluating the connection between social media and policy. It has shown that there is a relationship between social media presences (at least on one platform) with at least one policy outcome. In Ecuador, I found a strong connection, but not clear causes. I do not know yet know if a tweet could save a life. But I have found a connection between governments on social media and impact on their citizens.
 This difference may be due in part to We Are Social using a wide variety of sources with wildly different estimates (see the comment for Ecuador below.) They also use data directly from social media site’s reports of monthly users, which not only are more often reported than official statistics but have the incentive to estimate a larger number of users. Access for information on Statista’s sources is considered “premium” and is therefore behind a pay wall.
 This is the highest estimate, coming from World Internet Stats. The lowest estimate presented by We Are Social is Internet Live Stats with 7.06 million.
 For example, in 1987 there were 0 accounts and a 41.2 poverty rate.
 Account establishment years indicated by vertical lines.
 Voter turnout data from the International Institute for Democracy and Electoral Assistance: https://www.idea.int/data-tools/question-countries-view/521/98/ctr, https://www.idea.int/data-tools/question-countries-view/522/98/ctr
 As the MFAEcuador account is unverified, it was excluded in the original regional analysis and will be in this section as well, although it and all other unverified accounts were included in the global and Ecuadorian correlation and regression analyses in order to provide as large a dataset as possible.
 Three journalists were reporting on the rising violence along the Columbian border due to post-peace destabilization when they went missing and were later reported dead.
 The basic account information chart is based on the 2017 Twiplomacy data, collected before the new president had taken office and gained a more substantial following.
 Covered by the 50 tweets (indicates rate, where a higher number of days indicates fewer tweets per day)
 A judge suspended a CPCCS (Council of Citizen Participation and Social Control) resolution and Judicial Council contest to appoint judges.