Social media campaigns and political agenda-setting

Material Information

Social media campaigns and political agenda-setting the case of Mexico
Gomez Cruces, Juan S. ( author )
Physical Description:
1 electronic file (76 pages) : ;


Subjects / Keywords:
Social media -- Political aspects ( lcsh )
Politics and government -- Mexico -- 2000- ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


he aim of this thesis is to understand how social media campaigns are affecting political agenda-setting in a new democracy such as Mexico. The thesis takes into consideration the role of traditional media and electoral events as part of elements with prospects of affecting political agenda-setting and social media campaigns.
The thesis relies on combined methods to understand the dynamics of the variables above mentioned. A qualitative analysis performs an in-depth study of five paradigmatic social media campaigns. In the quantitative part of the thesis, timelines, multivariate and bivariate regressions are used to observe correlations between the variables.
The results of the thesis provide some evidence regarding a correlation between social media campaigns size and levels of political agenda-setting affectation. Results also show that for this kind of campaigns, social media has a strong influence over traditional media. Finally, the study also provides evidence suggesting a correlation on how left leaning traditional media and opposition parties are more prone to be influenced by social media campaigns
Thesis (M.A.)--University of Colorado Denver
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Juan S. Gomez Cruces.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
on10071 ( NOTIS )
1007153376 ( OCLC )


This item has the following downloads:

Full Text
B. A., Instituto Tecnologico y de Estudios de Monterrey, Estado de Mexico, 2006
A thesis submitted to the Political Science Program of the University of Colorado, Denver in partial fulfillment Of the requirements for the degree of Master of Arts in Political Science College of Liberal Arts and Sciences 2017

This thesis for Master of Arts in Political Science degree by Juan S. Gomez Cruces Has been approved for the Political Science Program by
Christoph Stefes, Chair Amy Hasinoff Michael Berry
Date: May 13, 2017

Gomez Cruces, Juan S.
Social Media Campaigns and Political Agenda-Setting: The Case of Mexico Thesis directed by Associate Professor Christoph Stefes
The aim of this thesis is to understand how social media campaigns are affecting political agenda-setting in a new democracy such as Mexico. The thesis take into consideration the role of traditional media and electoral events as part of elements with prospects of affecting political agenda-setting and social media campaigns.
The thesis relies on combined methods to understand the dynamics of the variables above mentioned. A qualitative analysis performs an in-depth study of five paradigmatic social media campaigns. In the quantitative part of the thesis, timelines, multivariate and bivariate regressions are used to observe correlations between the variables.
The results of the thesis provide some evidence regarding a correlation between social media campaigns size and levels of political agenda-setting affectation. Results also show that for this kind of campaigns, social media has strong influence over traditional media. Finally, the study also provide evidence suggesting a correlation on how left leaning traditional media and opposition parties are more prone to be influenced by social media campaigns.
The form and content of this abstract are approved. I recommend its publication.
Approved: Christoph Stefes

A mis padres y mi hermano Miguelpor creer en miy apoyarme siempre.
A mi amada Lorena, quien fija mi agenda con amor y felicidad.

INTRODUCTION.................................................................. 1
Korenfeld case, social accountability through social media platforms.25
#MiPrimerAcoso, social media and public policy.......................26
#SalvemosTajamar, public decision reverted...........................28
#Ley3de3, social media platforms pushing legislation.................30
#JusticiaparaRuben, call to action...................................32
Medina Mora case, apparent no political response.....................33
III. DATA AM) METHODS.....................................................36
IV. RESULTS..............................................................46
V. DISCUSSION...........................................................54
CONCLUDING REMARKS.........................................................58
NOTES..................................................................... 61

1. Mexican political parties located in the political spectrum........................22
2. Mexican political actors in Twitter..................................................23
3. Description of variables in the analysis.............................................41
4. Political agenda-setting affectation.................................................49
5. Interactions.........................................................................49

1. Twitter interactions among Mexican political actors.............................24
2. Social media size and political response as regards to 2015 election............ 52
3. Social media campaigns behavior in 30 days......................................53

Increased use of new technologies has modified several aspects of human life. In the political realm, the adoption of social media platforms has empowered new forms of political participation (Castells 2012; Diamond and Plattner 2012; Shirky 2010), but also new forms of authoritarian control and manipulation (Deibert and Rohozinski 2012; MacKinnon 2012; Morozov 2011). In 2010, the events of the Arab Spring promoted the idea that social media platforms were able to mobilize large segments of the population to remove long-lasting dictators (Aman and Jayroe 2013; Lotan et al. 2011). This notion of empowerment through new technologies also drove movements such as Los Indignados in Spain (Barbera et al. 2015; Castells 2012) and Occupy Wall St. in the U.S. (Barbera et al. 2015; Tremayne 2014; Castells 2012).
While the role of social media platforms during specific social movements has been widely studied, how these technologies may affect the political agenda-setting in a more regular basis remains understudied in political science. Thus, this study attempts to understand the impact that social media platforms may have over political agendas. According to Green-Pedersen and Krogstrup (2008), political agenda refers to the agenda of key political decision makers in contrast to the concept public agenda which refers to the issues the public and mass media emphasize or find important (2008). The study looks to answer, how does social media campaigns affect political agenda-setting in regular basis? Does higher activity on social media platforms results in more influence over political agenda-setting and traditional media? To answer these questions, the study takes into consideration other factors that may or may not affect political agenda-setting. This is the case of factors that have been studied as intervening variables in political agenda-setting: traditional media (Walgrave et al 2008; Tresch et al. 2013),

ideological orientation, electoral calculations (John et al 2014), and constituents preferences (Green-Pedersen and Walgrave 2014). These factors add new questions that must be addressed by this study. Is traditional media required for a social media campaign success? Are social media platforms influencing traditional media, or vice versa? Are both faces of the same coin? Is ideology driving political response to certain social media campaigns? If so, is government or opposition more responsive to social media campaigns? Are elections boosting social media campaigns? Moreover, are they promoting political responses?
This study relies on a quantitative analysis of 45 social media campaigns targeting Federal authorities in Mexico. To reduce the social media anarchic order, such cases were selected under three rules. First, social media campaigns must have an online petition with more than 4,000 signatures in the website. Second, the online petitions must be issued between June 2014 and June 2016. Third, the topic must reach 100 tweets in any single day within the 30 days previous or after the petition was issued. To control for traditional media, the study relies on front and editorial pages of two major national newspapers in Mexico, Reforma and La Jornada. The former with an editorial bias favoring the right while the latter with an editorial bias favoring the left (Gutierrez and Cuevas 2012). The assumption is that front and editorial pages are good indicators of the relevance that certain topic has on the traditional media agenda. Electoral calculations are another factor that might affect political agenda. This, an upcoming election could create a more politicized environment, boosting social media campaigns. To the same token, elections could increase the incentives of political parties to address certain issues. Thus, electoral calculations are controlled by setting the timeline of the cases a year before and a year later after a major electoral event, such as the Federal Congress election of June 2015 f Following the concept of Petrocik (1996) on political ownership, the study assume that being

opposition or government could bias political response. Political responses are also the mean to measure changes on the dependent variable, political agenda-setting affectation. The study takes into consideration eight kinds of political responses: points of agreement issued, points of agreement approved, bills issued, bills approved, public policies implemented, public servants resignations, public decisions reversed, and press releases published. The assumption here is that these kinds of responses require different levels of political involvement. Thus, every response obtained receives a value corresponding to the level of political involvement required. In the end, political affectation equals the total value of political responses. According to previous works, citizens preferences are a factor affecting political agenda-setting. Such preferences are expressed through online or offline mobilizations, and ultimately through elections. While offline mobilizations have proved to be effective as a mean for political change, they are rare events. Therefore, this study will focus on online mobilizations as the variable affecting or not the political agenda-setting.
To test the studys assumptions it relies on a mixed method research strategy. A qualitative analysis of paradigmatic cases explores in more detail the uprising of social media campaigns, their development and outcomes. For the quantitative analysis, this study relies on multivariate and bivariate regressions, as well as on timeline series. Even though most cases share similarities; the ways in which social media campaigns emerge and evolve vary from one to other. Furthermore, political responses also vary in forms and outcomes, resulting in different degrees of political agenda-setting affectation. Hence, in order to complement the quantitative analysis above described, a qualitative approach is provided. Such qualitative analysis describes the context and dynamics of social media and Mexican political actors. It also details five

paradigmatic social media campaigns in order to shed light on their emergence and evolution, as well as on the different ways and levels of political agenda-setting affectation.
As mentioned before, to analyze the political agenda-setting and the particular of role social media platforms have in that process, this study examines national cases in Mexico. The Mexican state existed for 70 years as a one-party regime (Smith 2005), which ended in 2000 with a free and fair presidential election2. Mexico meets the criteria required to assess the impact of social media over the political agenda-setting. The country has relatively free, fair and periodic elections as well as a multi-party system; even though some media outlets are allegedly controlled by the governmental institutions (La Jornada 2015a) there is likewise an independent mass media; and by 2016, Twitters user base in Mexico was of about 21.3 million (Emarketer 2016), the largest within Latin America, following Brazil.
This study is organized as follows. In the following chapter, a theoretical framework is provided to address the existing scholarly research on the topic; as well as the theory and hypotheses proposed by this study. Thereafter, chapter two analyzes some of the dynamics of social media campaigns and political agenda-setting in Mexico; here some paradigmatic cases are described to depict the theory proposed. The third chapter includes the data gathered; the methodological framework and the statistical analysis. The fourth chapter presents the results obtained throughout both analyses performed and the last chapter includes a discussion of the study findings regarding previous works. Finally, concluding remarks summarize the theoretical and methodological contributions of this study and provide some advice for further research.

Social media has increasingly captured the interest of political scientists in the past decade. Even though the Arab-Spring seems to be the catalyst of this interest (Aman and Jayroe 2013; Lotan et al. 2011), the role of social media during the 2008 U.S. Presidential Election presents one of the first case studies among scholars (Diamond and Plattner 2012). Although political science scholarship regarding social media is becoming more extensive, most of it is focused on rare events, such as elections and social movements. How social media may or may not affect political agenda-setting on a more regular basis has been an understudied aspect of this subject. Since the literature on social media and political agenda-setting remains underdeveloped, the next lines provide a review of the scholarly on both subjects.
In a seminal article published in the December 1962 edition of the American Political Science Review3, Peter Bachrach and Morton S. Baratz discussed how groups and individuals might determine routine or key issues that enter the political arena. Even though Bachrach and Baratzs (1962) article was a major contribution to the study of political agendas, as mentioned by Green-Pedersen and Walgrave (2014), it did not result on the emergence of political agendasetting literature. Nonetheless, in recent years scholarship on political agenda-setting has seen an increase. Within this new wave of recent studies4 on political agenda-setting, it was not possible to find literature addressing the role of social media on the political agenda-setting process. Scholars on political agenda-setting are mainly focused on the influence of factors such as mass media (Thensen 2013; Walgrave et al 2008), political systems, and political actors (John et al. 2014; Green-Pederson and Mortensen 2010; Petrocik 1996) on political agenda-setting.

In recent years, some agenda scholars have developed relevant empirical research on the dynamics between media, political actors, and systems and the political agenda setting process. Some of these scholars have contributed to understanding better how certain topics may or may not draw enough attention to be included on the political agenda. Gathering data on Danish parliamentary activities, Green-Pederson and Mortensen (2010) found evidence suggesting that party-system agenda is a relevant factor to determine the overall political agenda. Thus, they conclude that opposition parties have more opportunities to focus continually on issues that are advantageous to them, whereas government parties are compelled to respond to issues brought up on the party system agenda (Green-Pederson and Mortensen 2010:273). This conclusion builds upon the broader notion of issue ownership developed by Petrocik (1996) which states that parties own specific issues that are closer to their ideological orientation. In a study examining the United Kingdom, John and his colleagues (2014), analyzed speeches and acts of parliament to determine the importance of party politics over the agenda-setting. These scholars found that parties are focused on two tasks: they try to maintain their attention on emerging issues while remaining owners of certain issues (John et al. 2014).
Recent studies that highlight the role of mass media over political agenda-setting process are worth mentioning. Walgrave and his colleagues contributed to the literature with relevant findings on the influence of mass media over political agenda-setting; the exceptional influence of print media over audiovisual media; and the differences between sensational issues, prominent issues, and governmental issues (2008). Relying on data sets from a period from 1993 to 2000 in Belgium, these scholars design quantitative models to measure media-policy interactions. In doing so, they were able to determine that media effects are greater over parliament than over government branches. Furthermore, they show that print media has more

effect in terms of political agenda-setting than audio-visual media; and they likewise found that sensational issues are more prone to be included on political agenda versus other types of issues (Walgrave et al 2008). Due to the link between political dynamics and mass media over political-agenda setting, Gunnar Thesen shed light in understanding this process and why some issues are included while others are not. Thesen (2013) utilized data on more than 2,000 stories on Danish radio during the period between 2003 and 2004. According to the findings, opposition parties are more likely to be responsive when issues have a negative tone, whereas government tend to be more responsive to a positive tone (Thesen 2013). Moreover, both government and opposition are highly responsive to news containing blame attributions (Thesen 2013). Finally, regarding responsiveness based on issue ownership, Thesen (2013) suggests that the reaction effect is only moderate for both government and opposition parties. While these studies provide relevant theoretical contributions to understand the processes by which political agendas are affected, none takes into consideration the role that social media might potentially play. The following section reviews studies that analyze the central role of social media within the political arena.
Since the arrival of the Web 2.0, when online applications created a more interactive virtual environment (Steinberg 2015), the use of social media platforms, such as Facebook and Twitter, has become common among those with access to these technologies. During the 2008 U.S. presidential elections, Barack Obamas campaign successfully created a wide network of support throughout social media platforms (Diamond and Plattner 2012). Since then, almost any electoral campaign in consolidated and new democracies rely on social media platforms (Steinberg 2015) as a tool to disseminate discourse, call to action, and to recruit supporters. By 2010, the world witnessed another side of this coin when citizens of North African and Middle

Eastern states became organized via Facebook and Twitter in order to take to the streets in massive protests against authoritarian regimes (Aman and Jayroe 2013; Lotan et al. 2011). The so-called Arab Spring and Obamas 2008 campaign are now only two of several examples in which new technologies, and in particular social media platforms, have played a role in political processes.
The relationship between social media and politics has been analyzed by scholars from different disciplines: ranging from computational science (Bond and Messing 2015) to psychology, including communication science (Barbera et al. 2015; Bond and Messing, 2015; Barbera and Rivero, 2014; Breuer and Groshek, 2014; Groshek and Bachman, 2014; Fenton and Barassi, 2011; Shirky, 2011) as well as journalism, and political science (Barbera et al. 2015; Scherman, Arriagada, and Valenzuela 2015; Breuer and Groshek 2014; Oser, Hooghe and Marien 2013; Bekkers et al. 2011). Given the novelty of social media, existing research has been mostly published in the last ten years. The existing scholarly research on social media and politics can be divided in three categories: 1) social media and the political individual; 2) [rospects and threats for democratization; 3) influence and political agenda-setting.
In recent years, several studies have directed their attention to the relationship between social media and its users, particularly those using these technologies for political purposes. Profiling social media users has been an important focus for scholars interested in this topic. In doing so, they construct a clearer image of who is using social media platforms for political ends and how they are using these platforms. Studies following this trend have examined both demographic and ideological characteristics (Bond and Messing 2015; Barbera and Rivero 2014), whereas others have analyzed the transferal of online activities into offline actions (Barbera et al. 2015; Scherman, Arriagada, and Valenzuela, 2015; Breuer and Groshek, 2014).

Finally, Oser et al, provide research on demographics, ideological characteristics and online/offline transferal (2013).
By using a database of nearly 70 million tweets, Barbera and Rivero analyzed the social media behavior of citizens during the 2011 Spanish legislative elections and the 2012 U.S. presidential elections (2014). Among their findings, there is evidence indicating that Twitter users leading online political discussions tend to be in the extremes of the political spectrum (Barbera and Rivero 2014). Taking advantage of large datasets, Robert Bond and Solomon Messing (2015) use Facebook political endorsements to determine the demographic and ideological profile of U.S. citizens involved in online political discussions. Relying on politicians fan pages, amount of likes, friends in the network, among other variables, they were able to construct a characterization of social media users and their political views. Among other findings, they suggest that older people are more conservative than younger; women are more liberal than men; and college graduates tend to be as liberal as non-college graduates until their early 30s, then, college educated tend to be more conservative than before but less than their counterparts without college education. Also notable is their finding regarding the ideological closeness of friends within the network. They explain this through three possible causes: exposure to a shared environment, people choosing friends based on ideology, friends changing their friends views (Bond and Messing 2015).
One study on the differences between online and offline activists and mobilization use a database of telephone interviews to classify online participation (Oser, Hooghe and Marien 2013). They identified four different participation actions: sending emails, signing a petition, making online donations, and joining an online group supporting a cause. They classified offline participation into six different actions: signing a petition, contacting a government official in

person, by letter, or by phone, attending a rally, speech or protest, volunteering for a political party or candidate, and offline donation (Oser, Hooghe and Marien 2013). They devised four groups of individuals: disengaged, contacters, offline activists, and online activists. The authors noted that online activists do not seem to be substituting online activities for offline ones but rather are incorporating acts of participation along with offline activities into their individual-level participation repertoire (Oser, Hooghe and Marien 2013; 96).
In contrast to Oser et al, Breuer and Groshak (2014) examined the correlation between offline and online participants in the Ficha Limpa Campaign in Brazil to identify the links and gaps between actual activists and the so-called slacktivists which have been defined as low commitment participants constrained to social media political participation (Barbera et al 2015; Morozov 2011; Gladwell 2010). Using data from a survey on the participation in the mentioned campaign, they found a significant, but negative correlation between low-effort online activities and offline participation (Breuer and Groshek 2014). This was particularly clear regarding their variable measuring liking campaign content (Breuer and Groshek 2014). They also found no correlation between other low-effort online activities (i.e. sharing content, commenting) and offline participation (Breuer and Groshek 2014), which for them is a clear sign of a high level of slacktivism across online campaigns. However, they found a correlation between high-effort online activities, such as uploading pictures and video, writing for blogs or other online outlets, and participation in offline activities related with Ficha Limpa campaign. Although from a more general approach, Scherman et al. (2015) found a positive and significant correlation between Facebook and Twitter accounts holders and their participation in offline protests related to education and the environment in Chile.

In order to measure the impact of so-called slacktivists, Pablo Barbera et al used millions of tweets, to calculate the total size of online campaigns and the final impact on offline protests (2015). Their statistical analysis allowed them to conclude that those users tweeting about the protests even miles away from them actually created a positive soundboard effect to disseminate information.
These studies demonstrate an open debate on how individuals take advantage of new technologies as a form of political participation. Nonetheless, it seems clear that even those low-effort activities might have a positive impact on certain political processes.
The Arab Spring is frequently cited as an example of the prospect of online activity being a catalyst for democratization, the debate concerning the real impact of social media is still ongoing. The dichotomy between skeptics and optimists seems even more polarized. On skeptical side of this debate, Fenton and Barassi developed a qualitative study on the possible effects of social media on political participation (2011). Using the case of the Cuba Solidarity Campaign in the United Kingdom and conducting semi-structured interviews with members of the organization in their headquarters in London they conclude that the use of social media platforms in certain collective environments might result in the erosion of collective bonds due to the individuation characteristics of such technologies (Fenton and Barassi 2011). It is plausible that the goals and strategies of a given social organization or movement need a different framing than the one currently offered by the available social media platforms. However, it is also worth mentioning that Fenton and Barassi (2011) conducted their analysis before the events of the Arab Spring and other social movements that have shown how social media platforms create collective bonds.

In a pioneering article on the effects of social media on politics, Shirky (2011) highlights the relevance of these platforms in democratization processes in authoritarian contexts, as long as it moves forward freedom of speech and organization. Even though Shirky (2011) emphasized the role that the U.S. foreign policy must play to empower citizens in authoritarian regimes, its main contribution was to be one of the first scholars in highlighting the liberation potential of new technologies.
Scholarly research on social media and politics has been prolific in studying rare events such as regime change, elections, and social movements. Relying on data from 2011 environmental activists and student mobilizations in Chile, Scherman et al. show a correlation between Facebook and Twitter users and participants on the aforementioned street protests (2015). Contradicting the Sherman et al findings, Groshek and Bachman (2014) analyze the use of new technologies within the context of democratic transitions on Latin America; they conclude that such technologies had no influence over democratic advances in the region. However, data utilized by Groshek and Bachman (2014) correlates democratization with new technologies from 1946 to 2009. As thus, the findings are at least problematic given that most of the regime changes in Latin America occurred during the Twentieth century, while the expansion on the use of new technologies, particularly social media platforms, occurred at the end first decade of Twenty-First century.
Utilizing Green-Pedersen and Krogstrups definition of political agenda, two studies provide examples on the use of social media platforms as means to incorporate topics onto the agenda. In both cases, social media had an instrumental role to play in broader political participation. These examples come from the studies conducted by Bekkers et al. (2011) on the secondary student mobilization against the 1040-hour norm in the Netherlands, and Breuer and

Groshek (2014) on the Ficha Limpa campaign in Brazil. Through an empirical analysis, Bekkers et al. (2011) show how by using the MSN instant message service and Youtube, secondary students all over the Netherlands achieved significant levels of mobilizations against the 1040-hour norm. This norm pretended to increase the mandatory hours in Dutch secondary schools, which immediately sparked protests. These protests started through MSN instant messenger and Youtube, and eventually secondary students called to street protests (Bekkers et al. 2011).
The opposition to the norm grew all over the country and the topic became national and eventually reached the political agenda leading to its ultimate rejection (Bekkers et al. 2011). In a more recent but similar study, Breuer and Groshek (2014) used social media campaigns levels during Ficha Limpa to suggest that Facebook and Twitter were essential to successfully positioning the campaign onto the Brazilian political agenda by lowering the costs of organization for the sponsor NGO and by fostering mobilization as well. Even though these two cases show the potential of social media to affect political agendas, both represent studies that focus on a single event instead of analyzing daily politics.
Closer to the approach proposed by this study, in the book Interpreting Hashtags Politics, Stephen Jeffares (2014) analyzes how social media may be used by political actors to debate and move forward policy ideas. Jeffares (2014) examined Twitter activity related to an official hashtag created for the 2012 Police and Crime Commissioner Election in the United Kingdom. He measured the number of tweets published during October and November using the hashtag #MyPCC. While Jeffares conclusions suggest that social media capabilities remain underexplored by political actors, he suggests that a better use of these technologies may help to move forward policy ideas. It is worth noting the tracking work implemented by Jeffares (2014),

finding that during the campaign, the #MyPCC hashtag had a fluctuation between 150 and 600 per day until the election day when the peak reached 1,700 tweets. This finding shed light on the behavior of social media campaigns.
As seen before, this study is located in between two groups of studies currently empirically and theoretically disconnected. Nevertheless, both groups offer findings and approaches that are fundamental to understand how social media campaigns influence political agendas. Thus far, scholars on political agendas have shown how traditional media could affect political agenda-setting (Thensen 2013; Walgrave et al 2008). Considering such evidence is possible to establish certain considerations on how social media is now also affecting political agenda-setting. For instance, Thesen (2013) have found evidence on how news are generating key political decision-makers public statements, whereas Green-Pedersen and Mortensen have measured political response in the form of parliamentary activities. Hence, the assumption in this study is that social media platforms are affecting political agenda-setting in the form of statements or actions executed by key political decision-makers. Furthermore, for this study both dimensions of political responses have different degrees of influence. While Twitter and Facebook statements are becoming increasingly common in current politics, these forms of statement have less political commitment than parliamentary statements or official press releases. To the same token, key political decision-makers actions require different degree of commitment. Sponsoring a bill could represent certain degree of political agenda-setting affectation, but making it passes is a whole other level. Also from political agenda studies, the concept of issue ownership coined by Petrocik (1996) with empirical evidence in Green-Pederson and Mortensen (2010) is relevant to understand who is responding to social media campaigns. Since, such campaigns are frequently targeting national governments, it is plausible

that opposition is politically reacting more to social media than executive branches and lawmakers from the party in government.
The influence of electoral calculations over political agendas (John et al 2014) is another theoretical approach in political agenda studies that is relevant when observing possible effects of social media campaigns over political agenda-setting. Hence, political responses to social media campaigns might be also biased by such calculations. As such, social media campaigns closer to major electoral events might receive stronger political responses than campaigns further away. Online campaigns right before elections represent a risk for politicians of losing votes, creating incentives to be more responsive. On the other hand, campaigns away from elections might be more easily neglected since the risk of losing votes is not present. Also related with elections, political social media studies provide evidence of the behavior of users during electoral campaigns. Barbera and Rivero (2014) have shown that during elections, social media activity tends to be polarized. Despite Barbera and Rivero (2014) focused on issues directly associated with electoral campaigns, social media might take advantage of these kind of highly politicized environments to increase its activity on political issues not necessarily connected with the electoral agenda.
Social media and traditional media coexist offering information flows to the public; even more interestingly, they seem to be constantly influencing one to each other. From the literature on the Arab Spring, Lotan and colleagues (2011) provide evidence on how during those events, such influence was stronger from social media to traditional media than inversely. Since social media platforms are increasingly initiating campaigns with political ends, they are becoming a main source of information for traditional media. Nonetheless, the support of traditional media to

issues initiated or promoted through social media seems to be in the end essential to affect political agendas.
Another interest for this work is to establish parameters for the study of political social media campaigns. Previous works provide evidence on how online campaigns may behave over time. In Jeffares (2014), a U.K. government campaign had between 150 to 600 tweets per day. While is difficult to set a starting point for social media campaigns, Jeffares study suggests that 150 in a day might be the lower point for a campaign to be so. Since there is no more evidence in literature on how social media campaigns behave over time (beyond rare events), lowering the bar to 100 tweets per day seems adequate to identify these campaigns. Moreover, an assumption in this study regarding social media campaigns behavior is that a shocking effect is driving them. Although without a specific reference to it, Jeffares (2014) work provide evidence of this effect in an online campaign. Jeffares shows that a social media campaign had peaks associated with specific dates. However further evidence is needed to observe such effect in more than one case. The assumption in this study is that under such shocking effect, online campaigns reach their peak a few days after they started; then campaigns tend to dilute.
Similar to Jeffares work, this study proposes that social media can affect the political agenda. Contrary to Jeffares approach, this study suggests that social media users influence the political agenda, instead of government officials trying to move forward their agendas through these platforms. As mentioned before, this study suggests that social media campaigns are affecting political agendas in a similar fashion than traditional media do. Thus, the first hypothesis proposed is as follows:

Hypothesis 1 (HI): An increase in social media campaigns size, generates more political responses then it affects the political agenda-setting.
As mentioned above, various scholars have pointed out the relevance that traditional media has over the agenda-setting process (Thesen 2013; Walgrave et al. 2008). Furthermore, studies analyzing the relationship between social media on rare events have garnered attention to the role that traditional media has or does not have in these cases. For instance, Pablo Barbera and his colleagues (2015) have found that social media campaigns was useful in disseminating information which otherwise would plausibly be censored or ignored by traditional media during certain protests. By the same token, a study conducted on the Tunisian and Egyptian Revolutions, Lotan and colleagues (2011) focused on the information flows and the determinant role of Twitter in influencing traditional media outlets. Finally, Scherman and his colleagues (2015) suggest that during the development of two big mobilizations in Chile, Facebook was one of the main sources of political information among those attending to street protests.
As mentioned before, the literature available on this topic suggests new questions for this study: does social media always influences traditional media as Lotan and colleagues (2011) found occurred during the Arab Spring? Does traditional media influence social media? Are politicians equally responsive to social media as they are for traditional media (Thesen 2013; Walgrave et al. 2008)? While, the dynamic between social media and traditional media might not be established yet due to the growing influence of the former, it is worth trying to provide some answers to such a relationship. In doing so, this study builds on Lotan and colleagues (2011) work, suggesting that issues with strong support on social media are more prone to be covered by traditional media, even under business as usual conditions. Hence, the second hypothesis:

Hypothesis 2 (H2): If social media campaigns on a given political issue increases, then traditional media will plausibly cover the issue more extensively.
Furthermore, this study suggests that those issues initiated through social media platforms and covered by traditional media will have a greater affect over the political agenda. The theoretical proposal here is that there are certain political topics with initial impact on social media that end up being covered by traditional media. For such topics to affect political agenda, the boost of traditional media is necessary. Thus, the third hypothesis proposes:
Hypothesis 3 (H3): If social media campaigns are covered by traditional media, they are more likely to influence political agenda-setting.
Hypothesis 3b (H3b)\ If not covered by traditional media, social media campaigns are less likely to influence political agenda setting.
Scholars on political agenda-setting have shown that issue ownership might determine whether a topic affects the political agenda. While issues initiated through social media platforms might not align with a specific ideological orientation, they are a response to political actors actions or a demand to take action. Thus, building on previous literature, this study suggests that opposition parties maintain attention on emerging issues (John et al. 2014), particularly on those that represent critiques to government. In the case of Mexico, the governing party, the PRI and its allies are ideologically located on the center-right of the political ideology spectrum. Therefore, it is

expected to see more responsiveness from opposition parties located somewhere else on the political spectrum. Therefore, a fourth hypothesis is proposed:
Hypothesis 4 (H4)\ If an issue initiated through social media platforms affects the political agenda-setting, then higher levels of political responsiveness from opposition parties are expected.
Finally, this study explores two more aspects of the impact of social media on day-to-day political events. The first aspect has to do with the role that electoral environment plays on both, the social media campaigns size and on the political agenda-affectation. It is being proposed in this study that closeness to a major electoral event has relevant consequences over the level of political media conversation. Electoral times tend to politicize citizens that otherwise would remain apart from political debates. Thus, a fifth hypothesis proposed is as follows:
Hypothesis 5 (H5) : Other things being equal, if a major electoral event is close, then social media political conversation will increase.
Furthermore, major electoral events may have effects over political actors calculation. Therefore, upcoming elections are expected to produce higher political response to issues initiated through social media platforms. The hypothesis for this goes:
Hypothesis 6 (H6)\ Other things being equal, if a major electoral event is close, then political response to social media political conversation will be higher.

Finally, there is one last aspect considered for this study. Given multi-issue and long-term studies on social media and politics are still unavailable, is difficult to tell when topics start or when they reach their peak. Beyond the political consequences of understanding the dynamics of social media political conversation, this will contribute to future research. Considering the evidence gathered by Jeffares (2014) for the case of the PCC election in the UK where tweets fluctuated from 150 to 600 per day, it is plausible that more than 100 tweets on a day (without considering retweets) would be a minimum threshold5 to consider certain topic as important. Moreover, giving that the campaigns analyzed in this study emerged from users and not from a governmental agency, as is the case in Jeffares (2014) study; it is also plausible that within the first days after the reaching such threshold the issue reaches its peak as well. Two final hypotheses are proposed:
Hypothesis 7 (H7): If a social media campaigns reaches 100 tweets in a day, then such
issue will reach its peak within the next ten days.
In sum, the main theoretical aim of this study is to understand the role of social media in day-to-day politics to affect political agenda-setting. Given that previous studies have contributed to understanding the power of traditional media in affecting the political agenda-setting, it is essential to understand how social media and traditional media are interacting. The political environment will be taken into consideration in order to control for possible effects over social media content production and political actors responses. A final objective is to understand the behavior of social media campaigns in order to achieve basic knowledge for further research on the dynamics between social media and daily politics.

Mexico is a relatively new democracy. After 70 years of one-party rule, in 2000 the country transitioned to democracy. Contrary to other Latin American countries, Mexico never experienced a hard and traumatic breaking point. Thus, most of the previous political institutions remain intact, including corruption6, opacity, and influence peddling (International Transparency 2016). Yet, some democratic advances have taken place in the country since 2000, allowing citizens access to some accountability mechanisms. For instance, the 2002 Federal Transparency and Access to Governmental Public Information Act (DOF 2002) enabled citizens to make inquiries to the government on budget expenditure, governmental programs efficiency, public servants salaries, and organizational structures, among others. At the same token, after citizens demands for a more open political system7, in 2012 a constitutional reform (DOF 2012) enabled citizens to issue bills8 before the Mexican Congress. Hence, Mexicos democratic transition seems still inconclusive, but it has made some advances.
In terms of institutional design, Mexico is a presidential democracy with a multi-party system. After the transition in 2000, the right-wing National Action Party (PAN) captured the executive branch and remained in power for 12 years before the center-right Institutional Revolution Party9 (PRI) retook the presidency. In 2006, a coalition of left-wing parties were close to winning the presidential election, but in the end the PANs nominee won amid nationwide protests accusing electoral fraud and demanding a full vote recount. As of today, the country has 10 registered political parties; Table 1 shows them according to their location on the political spectrum. According to their political power, these 10 parties can be divided into two groups.

There are at least four major parties: MORENA, PRD, PRI, and PAN, all of them are able to win major elections by themselves depending on geographic and socioeconomic conditions. PT, MC, PVEM, PAN, and PES belong to the group of smaller parties. The latter group of parties has marginal positions in federal and state congresses, and in some small municipalities. Nevertheless, major parties need them for legislative and electoral alliances.
Table 1. Mexican political parties located in the political spectrum
Left Center Left Center Right Right
(National (Workers (Citizens (Democratic (Institutional (Green (New (National (Social
Regeneration Party) Movement) Revolution Revolution Ecologist Alliance Action Encounter)
Movement) Party) Party) Party of Mexico) Party) Party)
Source: National Electoral Institute (INE) data.
Mexican political parties and other political actors are deeply involved in social media platforms, particularly Twitter. Social media debates between politicians are not rare in Mexican politics and the federal government constantly uses Twitter as its preferred outlet to reach decisions relevant to the public (Ahmed 2017). To have a clearer image on how Mexican political actors are adopting Twitter, Table 2 depicts the number of tweets and followers of political parties and three relevant politicians in the country. The table also establishes whether or not Twitter has already verified their accounts. As shown on Table 2, four opposition parties (PAN, PRD, MC and MORENA) have tweeted more than 34,000 times, being more active than the government party (PRI), with less than 26,000 tweets. Nonetheless, the government party has more followers than any other party with 370,000 tweets, more than double the second most followed party, the PAN. Also interesting is the fact that out of all nine registered parties, eight have official accounts and four of them are verified by Twitter. Regarding the other actors, the Mexican

president has more than 6 million followers, even though he has tweeted less than five thousand times.
Table 2. Mexican political actors in Twitter
Actor Account Tweets Followers Verified
PAN @ AccionN acional 34.6K 181K Yes
PRI @PRI_Nacional 25.9K 370K Yes
PRD @PRDMexico 72.2K 97.4K No
PVEM @partidoverdemex 10.IK 34.7K Yes
MC @MovCiudadanoMX 34.3K 126K Yes
PANAL @NuevaAlianza 17.3K 19.6K No
MORENA @Redes_MORENA 35.3K 10.8K No
PES @PESoficialPPN 6.3K 27.5K No
President @EPN 4.7K 6.19M Yes
Foreign Minister @LVidegaray 8K 994K Yes
Interior Minister @osoriochong 5K 1.1M Yes
Note: Unverified accounts included here are those listed in the political parties official websites. Source: Twitter data on February 3, 2017.
These actors use their Twitter accounts to interact between each other and frequently with social organizations, think tanks, media outlets and citizens. Figure 1 shows the interactions of Mexican political actors between June 2014 and June 201610. During this period, these actors exchanged 1,841 tweets. Mexican president, Enrique Pena Nieto (@EPN) was the main receiver of these tweets with 1,267 (68%), although he only sent one tweet to these particular actors. The two most active accounts showed on Figure 1 are the Center of Research for Development11 (@CIDAC) a local think tank with 519 tweets (28%) and PRI (PRI Nacional), the governing party, with 339 tweets (18%).

Figure 1. Twitter interactions among Mexican political actors
Source: Authors own work based on Twitter data, February 3, 201712 with Circos Table Viewer.
Amid this vast network of digital interactions, social media campaigns emerge aiming to affect the political agenda-setting. Such affectation occurs in different forms; sometimes corruption is exposed leading to resignations, others lawmakers issue bills or other legislative instruments to address the issue. The following cases are part of the sample included on the model developed on this study. The aim is to provide a more detailed analysis of paradigmatic cases; some of them are good characterizations of trends and practices within social media platforms. The ultimate

objective here is to better understand how social media is triggering campaigns that eventually might be able to affect political agenda-setting.
Korenfeld case, social accountability through social media platforms
In April 1, 2015, social media users witnessed and shared a series of pictures of the then Mexican National Water Commissioner, David Korenfeld. In these pictures, Korenfeld and family members were captured boarding a National Water Commission helicopter for personal purposes. The topic generated around of 21,112 original tweets13 in five days (Twitter 2016) and an online petition asking for his resignation gathered 20,899 signatures ( 2016). In April 2, 2015, national newspapers covered the Korenfeld case on their front (La Jornada 2015b) and editorial (Bartolome 2015) pages. That same day, the Ministry of Public Administration set a term of five days for Korenfeld to hold a hearing and explain the situation (Secretaria de la Funcion Publica 2015a). Five days after the pictures were published on social media platforms, two points of agreement14 on the topic were presented in the Mexican Congress, one of them asking for Korenfelds resignation (SIL 2015). Amid an intense social media campaign against him, in April 9, Korenfeld resigned (Cervantes 2015). Finally, in April 30, the Ministry of Public Administration concluded its investigation and sanctioned Korenfeld with a fine of around $42,000 U.S. dollars (Secretaria de la Funcion Publica 2015b).
The Korenfeld case shows the potential of social media platforms in keeping public servants accountable. New technologies allow citizens to record videos or take photos of acts of corruption and other wrongdoing of public servants. Moreover, social media platforms are becoming an ideal mean to spread and denounce these kinds of behavior. Hence, when used appropriately, new technologies and particularly social media platforms can become tools for

what Smulovitz and Peruzzotti (2006) have defined as social accountability. In this kind of accountability, citizens monitor political authorities, expose wrongdoing, and activate institutional mechanisms of control, or what Guillermo ODonnell (1999) defines as horizontal accountability. In the Korenfeld case, a smartphone was used as a means to monitor a political authority; social media platforms worked as outlet to expose wrongdoing; finally, the points of agreement issued in the Mexican Congress and the Ministry of Public Administrations requisition for a hearing are examples of horizontal accountability activated by citizens actions.
The Korenfeld case represents an ideal where a citizen was able to take photos of a wrongdoing and spread them afterwards on social platforms. Here, social media influenced mass media and affected political agenda-setting. Finally, the most relevant affectation was the resignation of the National Water Commissioner. It is true that wrongdoings are still committed when nobody is watching and likely remain unknown; however, it is also true that new technologies allow more citizens to watch and when possible publicly expose wrongdoings.
#MiPrimerAcoso, social media and public policy
In April 23, 2016, a Mexican feminist organization launched on Twitter the hashtag #MiPrimerAcoso (#MyFirstSexualHarassment would be a fair translation). Originally, the objective was to join a street protest denouncing increasing levels of violence against women in Mexico (Paullier 2016). In only two hours, the hashtag received attention enough to become a trending topic (Paullier 2016). Under this hashtag thousands of Mexican women (and from other Latin American countries) shared heartbreaking stories in 140 characters or less. The topic reached around 15,000 tweets in that day, totaling 24,510 in five days (Twitter 2016) since the launch. Furthermore, analysis of the Twitter campaign revealed that 62% of the stories shared

were not only sexual harassment but also sexual abuse and that 8 was the most frequently reported age of first sexual harassment (Mujeres sin Violencia 2016a).
Days after the Twitter campaign and the street protest, traditional media covered the topic highlighting both, offline and online actions (La Jornada 2016a; Reforma 2016a). Weeks after the street protests, international media outlets such as the BBC, CNN en Espanol and El Pais conducted interviews with the creators of the #MiPrimerAcoso hashtag and some other special coverage (Abu Shihab 2016; Paullier, 2016; Reina 2016).
In the following weeks, the Mexican Congress issued three points of agreement and one bill on sexual harassment (SIL 2017a); all of the points of agreement were approved, while the bill remains under review. More relevant are the contents of these points of agreement. The first one was a call to all state legislatures to include sexual harassment on local legislation (SIL 2017a). The second point of agreement was a call to the Ministry of the Interior and the Ministry of Public Education to implement policies in higher education to prevent sexual harassment (SIL 2017a). The last point of agreement was a call to governors to implement public programs to prevent and sanction cases of sexual harassment. The executive branch response was developed throughout two strategies. On one hand, taking advantage of the already existing public campaign Mujeres sin Violencia (Women without Violence), it increased the outreach on child sexual harassment and abuse, particularly focused on early detection of possible cases of harassment (Mujeres sin Violencia 2016b). It is worth mentioning that Mujeres sin Violencia used data from MiPrimerAcoso campaign to nourish its own campaign (Mujeres sin Violencia 2016a). On the other hand, weeks after #MiPrimerAcoso, the Ministry of Public Education and the Human Rights National Commission agreed to take actions against sexual harassment and

abuse in elementary schools (SEP 2016); these actions include training for teachers and parents, and new protocols to prevent and react to these cases.
The case of #MiPrimerAcoso is interesting insofar it shows the prospects of social media to affect actual public policy. Here, the hashtag became a point of reference even for policymakers, and public campaigns such as Mujeres sin Violencia used the hashtag to engage more women in the prevention of sexual harassment. Moreover, federal agencies such as the Ministry of Public Education not only included the topic on its agenda but also developed specific public programs to address the issue.
In terms of academic research, this case also exposes some of the complexities in studying the links between social media and politics. Since #MiPrimerAcoso emerged to support a street protest, it is hard to tell if the success was due to the online or the offline action or both. Nonetheless, the fact that a public campaign such as Mujeres sin Violencia incorporated the hashtag UMiPrimerAcoso is symptomatic of the profound impact that the online campaign had on public decision-makers.
#Salvetnos Tajamar, public decision reverted
As suggested by Walgrave and colleagues (2008), sensational issues often have a better chance to affect political agendas, and environmental issues are highly sensational. In January 2016, pictures of dead wildlife began to flow into Facebook and Twitter. Such pictures were taken from Tajamar, a mangrove swamp located nearby Cancun in the Mexican state of Quintana Roo. According to governmental information, in 2005, a building permit was issued allowing the construction of a residential and commercial complex in the area (SEMARNAT 2016); however,

environmental activists argued that the Ministry of Environment and Nature Resources issued several irregular permits during that period of time (Azuela 2016).
After days of a social media campaign, sharing photos and other information related to the Tajamar case, on January 21, Twitter and Facebook users called for massive and coordinated online activity using the hashtag #SalvemosTajamar (#SaveTajamar) targeting political actors involved in what they said was an ecocide. That day, more than 10,000 tweets were published using exact or similar format as this:
@gurudlapolitica 21 Jan 2016
No to the ecocide in Tajamar (ajPROFEPA Mx @SEMARNAT_mx (ajbetoborge @PaulCarrillo2
@FonaturMX#SalvemosTajamar15 (Twitter 2016)
This wave of tweets targeted three Federal agencies, two of them related to environmental issues, the Federal Attorney for Environmental Protection (@PROFEPA), the Ministry of Environment and Nature Resources (@SEMARNAT), and a third one in charge tourism topics, the National Trust Fund for Tourism Development (@FonaturMX). Two more actors were also targeted, Roberto Borge, former Quintana Roo governor (@betoborge) and Paul Carrillo, former major of Benito Juarez, Quinta Roo (@PaulCarrillo2). Following this wave of tweets, sixteen points of agreement were issued in the Mexican Congress (SIL 2017b) and some small demonstrations took place in Mexico City and Cancun (Tasneen 2016). Despite federal authorities trying to defend the project when protests arose, when opposition increased even more and traditional media started to cover the issue, they changed their tone. In January 26, PROFEPA announced new inspections in the area (PROFEPA 2016) and days later, Fonatur published a press release detailing the fulfillment of a rescue program of wildlife in Tajamar (Fonatur 2016). Finally, a legal battle started in which several judges granted suspensions orders to stop the construction in the area (Mendez 2016).

The case of Tajamar illustrates three phenomena: 1) intra-social media organization; 2) the impact of sensational issues over the political agenda-setting; and 3) the power of social media to change already taken public decisions. Regarding the intra-social media organization, Tajamar offers an example of the dynamics within social media campaigns. Similar to offline actions, social media have leaders and develop strategies to potentiate its impacts; contrary to street protests and other forms of offline participation, leaders in social media are more diffuse. Often, strategies such as the tweet wave in the case of Tajamar are communicated in a horizontal fashion by sharing and retweeting. Thus, is difficult (if not impossible) to track who started a given strategy. While it is not the aim of this study to measure the impact of sensational issues over the political agenda, it is worth to highlight some aspects of this phenomenon. As demonstrated by the case of Tajamar, environment is part of these sensational issues. Citizens can be more easily moved by these kind of topics (Walgrave et al. 2008), resulting in a deeper degree of involvement and participation. Furthermore, Tajamar is an example of social media platforms being key in reversing unpopular public decisions, regardless of the economic and political implications that these reversals may have.
#Ley3de3, social media platforms pushing legislation
In March 2016, a group of citizens and civil society organizations sponsored a bill before the Mexican Congress regarding the responsibilities of public servants (SIL 2017c). Among other things, such bill aimed to adopt specific requirement in order to reduce political corruption. This bill was sponsored under the category of citizens initiative, therefore, according to the Mexican constitution; it had to have at least 0.13% of registered voters (or 106,963) signatures supporting the bill (D.O.F. 2016). In the end, 634,143 signatures were obtained supporting the

bill (IMCO 2016). The proposal became quickly known as #Ley3de3 (#Law3of3) since it included three main requirements for individuals in office: 1) statement of assets, 2) declaration of interests, and 3) tax returns (Ley3de3 2016).
To gain popular support and put pressure on lawmakers and other actors involved in the process, the sponsors launched an aggressive social media campaign. In April 2016, they started an online petition on the site asking for the approval of the #Ley3de3; this petition collected more than 130,000 signatures (Dresser 2016). A video on YouTube explaining the main features of the initiative had more than 160,000 views (Youtube 2016) and on Twitter, the hashtag #Ley3de3 reached almost 20,000 tweets during April and May 2016. The traditional media covered the topic on front pages (La Jornada 2016b) and editorial pages (Pardinas 2016). Furthermore, the issue received coverage by international outlets such as the BBC and the Wall Street Journal. In the Mexican Congress, debate around the bill was intense and even extraordinary sessions were needed to discuss and approve it.
Similar to the Ficha Limpa campaign in Brazil, #Ley3de3 gathered three elements to push legislation from civil society. First, both Ficha Limpa and #Ley3de3 took advantage of the issue of political corruption, which has plagued most Latin American societies almost since their independence. (Smith 2005). Second, civil society organizations involved in these cases took care of developing the legislative piece and assure popular support in order to increase the political cost of rejection. Third, the strategy on social media platforms was carefully rolled out to keep the issue on the political agenda until its final approval. Contrary to cases described above, the topic did not emerged on social media platforms, but from a utilitarian perspective they were essential to achieve the objectives.

#JusticiaparaRuben, call to action
After allegedly being persecuted by Javier Duarte, former Veracruz Governor, on June 9, 2015 Ruben Espinosa, local photographer and journalist, autoexiled to Mexico City. Less than two months after his arrival in Mexico City, Ruben Espinosa was tortured and killed along with four women (Quintero et al. 2015). By the afternoon of August 1, 2015, Twitter accounts published tweets asking for help in finding Ruben Espinosa:
@DliaRiveros 1 Aug 2015
Where are you Rubencillo???? #Journalist #Disappeared #RubenEspinosa #Veracruz #Mexico
@Latamite 1 Aug 2015
After fleeing #Veracruz, he sought safety in #Mexico City now #RubenEspinosa is currently missing
(Twitter 2016)
A few hours later, online news websites started to inform readers about the assassination of the photographer (Sin Embargo 2015). This news shocked the Mexican social media community and in the following five days, the hashtags #RubenEspinosa and #JusticiaparaRuben (#JusticeforRuben) reached 33,702 tweets. A petition in asking for an adequate investigation of this case collected almost 90,000 signatures. Traditional media covered the homicides on their front pages (La Jornada 2015c; Reforma 2015b). Social media reactions and traditional media coverage pushed legislators from almost all the parties in Congress to issue and approved 12 points of agreement demanding action by the federal and Mexico City governments to clarify the homicides of these five individuals (SIL 2017d).
In a rare case for any police department in Mexico, only five days after the homicides, the Mexico City Police Department captured its first suspect. However, despite previous threats that led Ruben Espinosa to autoexile in Mexico City, these were never taken serious during the investigation in determining a possible motive for the homicides. Instead, the Mexico City Attorney General persecuted three former local police officers for aggravated feminicide,

aggravated homicide, and aggravated robbery (Pantoja 2017). In January 2017, one of the police officers was sentenced to 315 years in prison. While the case was officially solved, some journalism organizations remain skeptical about the way the investigation was conducted and the outcomes obtained (Pantoja 2017).
The case of Ruben Espinosa shows how anger and ire can drive social media to demand action on certain situations. Furthermore, contrary to the case of #Ley3de3, this is a case of spontaneous online protest. Here the initial promoters are not easily tracked and the strategies emerged in a more natural way, in contrast with the carefully planned actions of #Ley3de3. The case of Ruben Espinosa falls into the category of sensational issues, which as proposed by Walgrave and colleagues (2008) tend to be more easily included in political agenda-setting. Even though the outcome of the investigation did not thoroughly satisfy the initial demands on social media platforms, they were successful in calling for a quick response from the local and federal authorities.
Medina Mora case, apparent no political response
The cases mentioned above depict examples of social media contributing, to different extents, in affecting Mexican political agenda-setting. The case of Medina Mora is paradigmatic insofar a high level of activity on social media platforms was unable to cause a clear effect on the political agenda. The next paragraph describes the case and attempts to provide an explanation as to the apparent lack of political response in this case.
Until March 10, 2015, Eduardo Medina Mora was one of the most controversial politicians in Mexico. Among other positions, he worked as Attorney General during the administration of Felipe Calderon Hinojosa, the former President of Mexico, who among other

things decided in 2007 to start a large-scale war on drugs that militarized many parts of the Mexican territory. Amid criticism of his performance as Attorney General, Medina Mora was sent to London to act as the Mexican Ambassador. By 2013, the new President of Mexico, Enrique Pena Nieto, appointed Medina Mora as Mexican Ambassador in Washington D.C., needless to say, a strategic post for the Mexican foreign policy. After two years as Ambassador in the U.S., Enrique Pena Nieto nominated him for a vacant seat at the National Supreme Court of Justice.
For sectors within Mexican civil society, this nomination was unbearable because the previous ties of Medina Mora with current and past governments. They decided to launch an online campaign against Medina Mora. The campaign included a petition in that collected above of 60,000 (Madrazo 2015) signatures. On Twitter, the topic generated around 14,000 tweets from mid-February to mid-March 2015 (Twitter 2016). Although traditional media covered the topic, they focused more on the lack of qualifications of Medina Mora to become judge than on the online opposition movement (Mendez 2015).
The Legislative branch never proposed a point of agreement asking the President to withdraw or change his nomination and there are at least two explanations for this apparent lack of response. First, nominations require approval by the Senate and it is plausible that opposition parties preferred to use stages within the legislative process to argue against Medina Moras nomination. For instance, in March 3, 2015, during a Senate committee meeting, two center-left senators dropped the petition with 20,000 signatures to establish the popular opposition against Medina Mora (Animal Politico 2015). A second explanation has to do with the very institutional design of the Mexican division of power. Nominating judges to the Supreme Court is an exclusive presidential faculty, thus a point of agreement asking for a

withdrawal or change would be seen as an intromission that would never be approved by the Congress.
The case of Medina Mora illustrates the complexity in tracking political agenda setting affectation. It also shows that certain political decisions are more difficult to revert than others are. Even though the online campaign was very successful in terms of citizens participation, a presidential decision supported by enough Senate votes seems to be difficult if not impossible to revert by tweeting and signing online petitions.
The five cases characterize the different modes in which social media platforms trigger campaigns. In some cases, a campaign is triggered by anger, disapproval, or indignation; in some others, a campaign is carefully planned and triggered by presenting arguments to social media users. Furthermore, these cases show the different degrees of planning in online campaigns. While some campaigns emerged almost out of the blue and remained under a spontaneous logic, other campaigns were planned by leaders who organized the campaign step-by-step. Finally, the cases above also illustrate different levels of political agenda-setting affectation. Some of them are examples of very successful online campaigns with affectations that led to changes in legislation or for public servants resignations. Others had mixed results, sometimes succeeding in positioning a topic on the political agenda-setting, but sometimes with outcomes against their initial goals. In the next sections, this study analyzes more cases from a quantitative perspective to determine how and to what degree social media shapes political agenda-setting processes.

The quantitative analysis in the study includes data for 45 cases of online campaigns between June 2014 and June 2016. The aim is to measure possible correlations, interactions, and behaviors between social media, elections, traditional media, and political agenda-setting. This chapter describes the methodology followed during the case selection process, as well as the adopted measurements for the variables in the analysis. Finally, the chapter describes the selected methods used to perform the analysis.
Anarchy and chaos could describe the way in which social media initiate, promote, discard or forget topics. A trending topic today might be irrelevant tomorrow, or remain relevant for five days or a month. Hence, social media research requires clear boundaries during case selection process.
For this study, the starting point was to preselect online petitions from the website Since the goal was to have a sample of online campaigns targeting Federal authorities, by utilizing it was possible to establish a specific time line for petitions targeting specific political actors. Three rules were followed in deciding which petitions to include in a preliminary sample. First, the time line was determined by petitions initiated between June 2014 and June 2016; second, in order to include a minimal threshold of relevance, petitions must have had more than 4,000 signatures; third, they must have targeted one or more Federal authority or authorities, regardless of whether it was an individual or an institution. After performing this research, 98 petitions complied with the three rules and were included in a preliminary sample.

A second step followed was to determine whether petitions contained within the preliminary sample had a significant size on Twitter. In doing so, keywords were assigned to every petition in order to perform a search using the Twitter advanced search engine16. For some cases, these new searches allowed identification of new keywords, often in the form of hashtags. To categorize a topic as relevant on Twitter, this new search followed two conditions. First, in order to establish a minimal threshold of relevance, any combination of keywords or hashtags (original and newly identified) must have generated at least 100 tweets in a single day. Second, that single day with 100 tweets or more, must have been within 30 days before or 30 days after the petition was initiated. These conditions allow for more confidence when establishing a link between petition and online campaign on Twitter. After conducting this process, the final sample was narrowed to 45 cases of online campaigns on Twitter.
The analysis includes four different main variables, social media campaign size, traditional media coverage, closeness to major electoral event, and political agenda-setting affectation.
Some of these variables are broken into smaller pieces of data to observe specific effects. For instance, traditional media is broken into coverage prior to and after a social media campaign, and prior political agenda-setting affectation. In addition, traditional media is sometimes presented according to the ideological leaning of a specific newspaper. To the same token, political agenda-setting affectation is broken into government and opposition responses. Some of the variables in the study interact with each other in different forms. This situation makes necessary to understand some variables as both an independent and dependent variables. The

next lines provide a detailed description of the variables included, how they are operationalized and their measurement.
Political agenda-setting affectation can be an elusive concept to measure. If, as proposed by Green-Pedersen and Krogstrup (2008), political agenda refers to the agenda of key political decision makers, then the question is how to know the topics relevant for such actors. While some authors have relied on governmental attention and opposition responses to news (Thesen 2013), others have preferred using parliamentary activities (Green-Pedersen and Mortensen 2010). Building on these two approaches, is possible to suggest that political agenda-setting is determined by statements and actions of key political decision makers.
Nowadays, political statements on traditional media or social media are common currency. Hence, an answer to media inquiries or political statements tweeted or posted on Facebook, represent a low involvement form of political statement that is disregarded in the study as measure for political agenda-setting. Thus, to capture the declarative side of political agenda-setting, two political responses are taken into consideration: 1) issuing points of agreement on the side of the Legislative branch and 2) publishing press releases by Executive branch agencies. On the other hand, to measure political decision-makers actions the study relies on five kinds of responses: 1) approving points of agreement and bills and 2) issuing bills on the Legislative branch. Whereas, for the Executive branch this study identifies, 3) public policy implementation, 4) public servant resignations and/or appointments, and 5) public decisions reversed.
To measure Legislative branch responses, the study relies on data from Sistema de Informacidn Legislativa (Legislative Information System, or SIL). The SIL is an official, open and online tool that allows access to data on Mexican Congress activities. Even though, SIL is a

Federal government tool, its data is reliable. During the study, searches in SIL were performed to obtain data on every case in the sample. These searches used keywords related with every case under pre-established range of dates. The range of dates started on the day when the campaign reached 100 tweets and went 60 days after that. The assumption is that any effect on political agenda-setting related to social media platforms must be observed within the next 60 days after the online campaign began. The study bases this assumption on evidence found in Jeffares (2014) work on the #MyPCC in U.K. There it was possible to observe fluctuations of tweets following a shocking effect. Throughout searches in SIL, it was possible to identify the points of agreement or bills issued and approved that were related with the cases in the sample.
To measure Executive branch activities, this study relies on the Federal government agency websites. In a similar fashion and for the same reasons outlined with SIL, in the case of government agencies websites17 keywords were used to identify responses associated with the 45 cases in the sample. In addition, searches were limited to a range of 60 days from the first day with 100 tweets. The aim was to identify press releases, public policy implementation, public servant resignations and/or appointments, and public decisions reversed.
Since all these legislative and executive responses require different degrees of involvement, the study establishes a scale from 1 to 5 according to the different level of political agenda-setting affectation. Thus, low involvement activities such as issuing points of agreements or publishing press releases receive one point; issuing bills receives one and a half points since they are less declarative than points of agreement, but without approval do not produce any tangible action. Points of agreement approvals receive two points because they require a legislative consensus within Congress. While points of agreement are calls to action, they are not binding instruments, as with bills. Hence, approving bills receives five points for the

political and legal implications they might have and the political involvement they require. Finally, public policy implementation, public servant resignations/appointments, and public decisions reversed all receive five points due to them requiring high political involvement with tangible implications. In sum, political agenda-setting affectation is measured as a continuous variable; insofar as political responses might be infinite, theoretically speaking.
A single issue could receive multiple sources and forms of political response. For instance, some of the online campaigns mentioned on the previous chapter received responses from both branches; others received more than one response from a single branch. Hence, the measure for political agenda-setting affectation is represented by a continuous variable. Furthermore, in addition to the total measure of affectation, the study differentiates between responses from the opposition and from the Federal government. To account for the latter, the study identifies responses from the Executive branch, plus those from the government party (PRI) and its electoral allies (PVEM and PANAL) in the Congress. The opposition response is measured by Congressional responses that are different from those of the PRI and its allies. Table 3, shows a summary of the data obtained for all the variables in the study. Regarding the aggregated political-agenda setting affectation, the maximum for a single case was 40.5 and the mean for all 45 cases was 5.03. Opposition responses had a mean of 2.98 and a maximum of 28.5, while government responses had a mean of 2.06 and a maximum of 16.

Table 3. Description of variables in the analysis
Variable name N Mean SD Min. Max.
Social media size 45 7,079 12,114 191 49,759
Day 1 to 5 45 3,353 6,747 120 33,702
Day 6 to 10 45 1,199 3,167 3 20,889
Day 11 to 15 45 855 1,925 1 8,471
Day 16 to 20 45 623 1,457 0 6,599
Day 21 to 25 45 568 1,745 0 10,920
Day 26 to 30 45 468 1,495 0 9298
Media coverage before social media campaign (aggregated) 45 0.28 0.98 0 5
Left leaning 45 0.11 0.43 0 2
Right leaning 45 0.20 0.65 0 3
Media coverage after social media campaign (aggregated) 45 1.64 2.31 0 6
Left leaning 45 0.82 1.27 0 3
Right leaning 45 0.82 1.18 0 3
Media coverage before political response (aggregated) 45 1.35 2.07 0 6
Left leaning 45 0.67 1.19 0 3
Right leaning 45 0.69 1.07 0 3
Closeness to major electoral event (social media campaign) 45 90 167 -355 323
Closeness to major electoral event (political response) 21 90 165 -352 341
Political agenda-setting affectation (aggregated) 45 5.03 10.01 0 40.5
Opposition responses 45 2.98 6.48 0 28.5
Government responses 45 2.06 4.41 0 16
To capture social media campaign size, the study relies on the last part of the method described in the selection case process. By running searches for the identified keywords and hashtags on the Twitter search engine, it was possible to determine the number of tweets published in a period of 30 days after a given topic reached 100 tweets. While this process was almost automatized, it required human intervention to observe that tweets were actually related to the topic. Since all these topics used keywords or hashtags in Spanish, sometimes unrelated issues from Spain or other parts of Latin America appeared within the search results. In such cases, the search was narrowed by incorporating elements to the searching keywords combination in order to obtain only tweets related with the campaign, regardless of the geographic zone from which users tweeted18. Social media size is measured as a continuous variable with a starting point at 100 tweets. To observe the behavior of these campaigns, Twitter

activity was measured in periods of five days. Table 3 shows data for the total size of social media campaigns and for the six periods of five days. Data gathered shows the largest online campaign reached 49,759 tweets while the smallest only reached 191, and the mean for the 45 cases was of 7,069. The six periods of six days make it possible to observe a decreasing mean over time. Thus, while on the first five days the mean is 3,353 tweets, by the end of the period it is only 468.
Identifying links and interactions between traditional media in regards to political agenda-setting and traditional media is essential for the theoretical purposes of the study. Thus, to measure traditional media coverage variable, the study relies on front page and editorial pages of two major national newspapers. The assumption is that this kind of news has a better chance to affect political agenda-setting and to influence social media campaigns than less relevant news such as those in inside pages. To reduce biases as much as possible, the selected newspapers were La Jornada and Reformat the former often considered as center-left leaning and the latter as center-right leaning. While both newspapers allow online access to their printed versions, Reforma provides a search engine that makes easier to find topics with keywords and under particular filters such as publication page or date range. For the case of La Jornada, it was necessary conduct an individual review of front pages and editorial pages under the criteria required. To identify possible interactions between traditional media, social media campaigns and political responses, the timeline for newspaper searches went from 15 days prior to the beginning of the online campaign to 30 days after. Hence, traditional media coverage was disaggregated into three different variables. First, traditional media coverage before a social media campaign captures all the coverage of the issue prior to the beginning of the online campaign. Second, traditional media coverage after a social media campaign, refers to all the

coverage of the issue after the beginning of the online campaign. Third, traditional media coverage before a political response, includes all mention of a topic prior to any political response regardless the beginning of the online campaign. Furthermore, all three variables were disaggregated into left leaning and rights leaning media to capture possible ideological trends.
Traditional media coverage before and after social media are measured as discrete variables with possible values from 0 to 6; whereas traditional media coverage before political response is also measured as discrete variable but with possible values from 0 to 12. Measurement was divided into two periods: 15 days prior online campaign beginning and 30 days after campaign beginning. Measurement was also divided between front pages and editorial pages. Thus, values were assigned only to the first mention of a given topic on every period and every newspaper location. For instance, a single topic could obtain points only four times for every newspaper. Regarding the scale proposed, mention on the front-page received 2 points; whereas mentions were identified on editorial page only received 1 point. Table 3 shows that traditional media coverage before social media has a mean of 0.28, considerably lower to the 1.64 of traditional media coverage after social media. Also interesting is that traditional media coverage before political response only reach a maximum of 6, even though 12 was the maximum possible. For disaggregated variables, data shows similar values for left and right leaning media, except for the coverage prior social media where right leaning media have a slightly higher mean.
Accounting for influence of electoral interests is relevant to explain affectation over political agenda-setting and social media activity. Therefore, the last variable in the study is closeness to a major electoral event. As suggested by previous studies, electoral calculations might have a role in political agenda-setting affectation (John et al 2014). To measure that effect,

this study relies on the June 2015 federal election in Mexico as a temporal reference. The election corresponded to the renewal of the entire Mexican House of Representatives, and occurred at the middle of the presidential term. Since Mexico has no presidential reelection, Congress renewal elections (always in the middle of the six-year presidential term) are often seen as a referendum to approve or disapprove presidential performance thus far. To take advantage of this event, the cases selected have as temporal criteria a range from June 2014 (one year before the election) and June 2016 (one year after the election). Thus, this variable is measure as an interval variable going form -365 days to 365days, having June 2015 as meaningful zero. However, for the purposes of this analysis, the absolute value of the variable is the one considered. Thus, if a campaign or a political response happened 40 days prior the election, its value is measured as -40 but for regression analyses such value would be 40. This variable is measured twice, one time for social media campaigns beginning and one for the first political response date. According to table 3, data gathered shows that the mean for social media campaigns and for political response is 90 days for both cases.
To observe possible affectations, interactions, and behaviors between the aforementioned variables, this study relies on a series of regressions and times series plots. Three multivariate regressions are used to explain political agenda-affectation. A fourth multivariate regression captures interactions between social media campaigns, traditional media coverage, and major electoral event. Three bivariate regressions were used to show interactions between social media and traditional media. In order to observe the behavior of social media campaigns along 30 days,

a time series plot was performed. Finally, to observe the behaviors of social media and political responses in relation to electoral events, another time series plot is provided.
Multivariate regressions related to political agenda-setting affectation include political response as dependent variable, while accounting for three independent variables: social media campaign size, traditional media coverage (prior political response), and closeness to the 2015 election. To test different degrees of affectation between opposition and government parties, three different models were built. Model one has as its dependent variable only the value of opposition response, and as independent variables, social media campaign size, traditional media coverage prior political response and closeness to 2015 election. Model two has the same characteristics as model one, except for the dependent variable, which is replaced by government response. For model three, the dependent variable is the aggregated value of the political response. To observe whether media leaning has influence over political agenda-setting, all models include disaggregated values for such variable.
To test interactions between variables, the study relies on multivariate and bivariate regressions. The first interaction studied is between traditional media coverage prior social media campaign and 2015 election as independent variables, and social media campaign size as dependent variable. A second interaction is between social media campaign size as the independent variable and traditional media coverage after social media campaign as the dependent variable. Here bivariate regressions were performed for disaggregated variables, left leaning media and right leaning media.
To observe the behavior of social media campaign sizes over the 30 days of analysis, the study provides a time series plot of the mean for all the 45 cases. Time series plot includes data for every five days (see Table 3). Finally, to complement the analysis regarding the influence of

2015 election, a time series plot depicts the values of social media size for the 45 cases along the two years of the study and for political response of the 21 cases (only those with at least one point of political response).
This chapter presents the results derived from the methods employed above. First, it presents the results obtained for the political agenda-setting affectation. Second, results on interactions between traditional media coverage, social media size, and electoral atmosphere are presented; and the last section addresses the variation of social media campaigns size over time.
As mentioned before, political agenda-setting affectation was tested throughout a multivariate regression. Table 4 shows the three models proposed according to the type of actor responding: government response, opposition response, and the aggregated response. It also account for the disaggregated and aggregated influence of traditional media, and the role of 2015 election over political responses. Regarding the relationship between social media campaigns and political agenda-setting affectation, these results support the first proposed hypothesis in this study. Evidence found in this study suggests a correlation between larger social media campaign size and increase in the degree of affectation expressed in political agenda-setting. An explanation for this finding is that political actors are increasingly aware of social media platforms. Hence, when social media campaigns become larger, political actors feel more pressure and a need to respond. To complement these findings, paradigmatic cases analyzed in this study provide evidence on how social media platforms had affected political agenda-setting in different ways. Such cases show how political actors responded to social media campaigns;

sometimes they tried to avoid responsibilities but when pressure became unbearable it was possible to see resignations, new laws adopted or public decisions reversed. In some cases, social media platforms were used as a tool for social accountability, for instance denouncing wrongdoing of a particular public servant or unlawful permits damaging the environment.
When comparing opposition and government models, it is possible to observe that social media size has a higher coefficient for opposition responses than for government responses. Moreover, in both cases, this correlation is statistically significant. Thus, the results indicate a slightly higher degree of responsiveness from opposition to online campaigns in comparison with governmental responsiveness. This finding supports hypothesis 4 of this study, insofar as, social media size seems to have a stronger affectation on opposition parties response than for government response. This could be explained through the very nature of social media demands. Since such demands are often directed to actors in the Executive branch, opposition parties have more incentive to advocate on the side of social media users.
Regarding the role of traditional media coverage and contrary to the expectations expressed in hypothesis 3, the variable seems not to boost political response to social media campaigns. The results obtained in this study show no statistical significance in the three models presented. For the cases of total response captured and the opposition response, traditional media coverage seems to have a small negative effect. Media ideological leaning showed no statistically evidence of affectation over political agenda-setting either. These results do not indicate that social media platforms have in general terms more influence over political agendasetting affectation. Nonetheless, for issues initiated or highly promoted on social media, it seems that the level of activity on such platforms offers a better explanation for political response than traditional media.

Regarding the major electoral event variable and its relationship with political agendasetting affectation, the study could not find a statistically significant correlation. These results contradict the proposal of hypothesis 6, since closeness to election was expected to be a variable explaining change on political agenda-setting. However, figure 3 offers a picture of the distribution of cases according to their political responses in regards to the 2015 election. Relying on this figure is possible to observe that two of the cases with higher political response

Table 4. Political agenda-setting affectation
Variables Government response Opposition response Aggregated
Social media size Media coverage 0.82 (0.11)*** 0.82 (0.11)*** 0.97 (0.10)*** 0.97 (0.10)*** 0.99 (0.07)*** 0.99 (0.07)***
Left leaning media 0.11 (0.12) -0.15 (0.11) -0.04 (0.08)
Right leaning media -0.08 (0.12) -0.03 (0.11) -0.05 (0.09)
Aggregated 0.03 (0.11) -0.17 (0.10) -0.09 (0.07)
Major electoral event 0.11 (0.08) 0.10 (0.09) 0.02 (0.08) 0.03 (0.07) 0.06 (0.06) 0.06 (0.05)
N 45 45 45 45 45 45
Notes: Standard errors reported in parentheses. Traditional media coverage prior political responses. ***p< 0.01
Table 5. Interactions
Variables Social media size Media coverageA Left leaning media Media coverageA Right leaning media Media coverageA Aggregated
Media coverage1. Left leaning media -0.05 (0.22)
Right leaning media 0.10(0.22)
Aggregated 0.03 (0.15)
Major electoral event -0.11 (0.16) -0.09 (0.15)
Social media size 0.70 (0.10)*** 0.66 (0.11)*** 0.72 (0.10)***
N 45 45 45 45 45
Notes: Standard errors reported in parentheses, t Traditional media prior social media campaign beginning. ATraditional media after social media campaign
beginning. ***p<0.01

are close to the day of the election. To that same token, several smaller cases are also close to the 2015 election. An explanation to these results can be found in the effects of electoral campaigns over political responsiveness. Even though under an electoral atmosphere political actors could be more eager to respond to publics demands, it is plausible that topics unrelated to electoral issues or with less electoral gains would be neglected.
As for interactions models, the results obtained are mixed. The interaction between traditional media coverage before a social media campaign and the 2015 election, with respect to their affectation on social media campaigns, were not statistically significant. The lack of influence of traditional media over social media could be explained by the nature of the case selection. Most cases in the analysis gained visibility first on social media platforms rather than on traditional media. In other words, the specific conditions surrounding the cases in this study might explain why traditional media was not a driver to boost social media size. Regarding the disaggregated variables for traditional media prior to a social media campaign, although left leaning media showed a negative coefficient and right leaning media a positive one, none of them were statistically significant
For the case of major electoral event variable, table 5 shows slightly negative coefficients, close to zero which might indicate that closeness to election is better for social media campaign size; however, none of them was statistically significant. These results reject the proposals for hypothesis 5; since the 2015 election was expected to, at least partially, explain social media campaign size increases. Nevertheless, figure 3 shows the distribution of cases and their corresponding social media sizes in regards to the 2015 election. It is worth highlighting that two of the cases with the largest social media sizes are close to election day. Moreover, several smaller cases are within the three months before and after the election. An explanation as

to the results in the model could be found in the effects of electoral campaigns over social media platforms. Even though electoral periods tend to politicize public opinion, it is also plausible that under such atmosphere, social media campaigns have to compete harder for attention against topics directly connected with electoral issues. Hence, even when citizens are trying to initiate online campaigns, the electoral agenda might be capturing more social media attention. Yet more evidence and further research would be needed to confirm or reject such hypothesis.
The study also account for the interaction between social media size and traditional media coverage. Here, the results show statistically significant evidence for social media size influencing traditional media coverage. These results give support to hypothesis 2, by showing a correlation between social media size and traditional media coverage. For the case of disaggregated variables, the analysis suggests a slightly higher coefficient for left leaning media than for right leaning media; both analyses were statistically significant. Such a result could suggest a higher degree of responsiveness from left leaning media towards social media campaigns. Considering that PRI and its allies are closer to a center-right ideology, is understandable that left leaning media covers online campaigns demanding action of the government. It is worth mentioning that the interaction observed is for cases intrinsically related to issues initiated or highly promoted on social media platforms. It is not being stated that social media platforms are determining traditional media coverage. The analysis of paradigmatic cases shows a strong influence of social media over traditional media. For some of these cases, even international media outlets covered the campaigns or performed in-depth interviews with organizers.

Figure 2. Social media size and political response as regards to 2015 election
45 O
4-U ^ Q X
33 Qfl X
jO O s
o on o X
^yj X 1 g X O A X
TO o o i n V
X 10 # 5 \> < o X o
v 3 O o On V A O Ooxx *o o -365 -325 -285 -245 -205 -165 -125 -85 -45 -5 35 75 115 155 195 235 275 315 355
35000 £ Id 30000 ~
25000 a 43
20000 S
15000 cw
Closeness to 2015 election (days)
XPolitical response OSocial media size
Regarding the behavior of social media campaigns over time, the study found that the first five days are crucial for the overall campaign size. Figure 3 shows the mean for all 45 cases during 30 days. According to data obtained, a peak is reached in the first five days, followed by a strong drop on the next five. Then, the decline seems to continue in a more steady fashion. These findings support hypothesis 7 by showing that social media campaigns reach their peak in the following 10 days after the date when 100 tweets were published. These results suggest that social media campaigns are for the most part a short-lived phenomenon with a shocking effect. Such effect is also visible in paradigmatic cases; in some of them social media campaigns grew to their peak point in one or two days, generating political responses almost in a couple of days after they reached such point.

Figure 3. Social media campaigns behavior in 30 days
3500 3000 2500
£ 2000
2 1500
^ 1000 500 0
0-5 days 6-10 days 11-15 days 16-20 days 21-25 days 25-30 days
Sources: Authors work with Twitter data gathered from June to November 2016.
Evidence provided in this study seems to support the notion of social media platforms acting as a new element in explaining political agenda-setting affectation for cases under certain circumstances. Furthermore, for the cases analyzed in the study, social media seems to be determining traditional media coverage and not the other way around. Elections cannot be discarded as explanatory variable for the success of certain issues promoted in social media platforms; nonetheless, further analyses are necessary to observe or entirely discard such interaction. Parallel to the quantitative results obtained here, qualitative analysis also found high influence of social media over political responses and traditional media. Finally, social media behavior seems to follow a shocking effect. The first days are crucial for campaigns to show their political muscle. Afterwards, campaigns lose strength but remained active for the following

weeks. Paradigmatic cases analyzed also show the shocking effect of social media campaigns, having cases in which in a single day thousands of tweets were published and key political decision makers had no option but to act as fast as possible.
The findings in this study support previous works suggesting a role of social media on political issues (particularly, social movements and elections). It also gives support to the concept of issue ownership as a key element for political parties when setting their agendas. Regarding the interactions between social media and traditional media, the findings support previous works suggesting that social media is increasingly affecting traditional media coverage by generating information flows that are transformed into mainstream news. However, as detailed, these findings should be taken with a measure of caution. The expected boosting effect of electoral events over social media size and political responsiveness was difficult to observe, but cannot be discarded. For the most part, this study bridges earlier work on social media as an emerging channel for political participation with recent political agenda-setting studies. Social media platforms are becoming an outlet for thousands of citizens looking to express ideas or demands on an everyday basis. Furthermore, political actors are taking advantage of social media platforms to interact with citizens, organizations or among each other. Therefore, social media exchanges increasingly affect political agenda-setting sometimes with long-term political and legal implications. Although these findings are optimistic regarding the use of social media as a political tool, the qualitative analysis performed in this study also offers paradigmatic cases where social media were effective in creating a soundboard for certain topics but with no political consequences.

Turning to social media as an emerging channel for political participation, this study supports the empirical evidence of single case studies (Breuer and Groshek 2014; Jeffares 2014; Bekkers et al. 2011) where such technologies were effective in mobilizing citizens to affect the political agenda-setting. In contrast to these studies, this study relied on a long-term and multiple case approach. This approach allows us to observe social media campaigns affecting political agenda-setting to different extents and forms, and under different electoral circumstances. Recent political agenda-setting studies neglect the role of social media campaigns as an emerging element affecting agendas (Green-Pedersen and Walgrave 2014; John et al. 2014; Thesen 2013; Tresch et al. 2013; Green-Pederson and Mortensen 2010; Walgrave et al 2008). However, some of their theoretical proposals and empirical evidence are applicable to the case of social media. For instance, this study has shown that opposition responses to social media campaigns are stronger than government responses. This might be explained because of issue ownership, a concept developed by Petrocik (1996) with previous empirical evidence in Green-Pederson and Mortensen (2010). Moreover, recent political agenda-setting studies suggest that sensational issues are more prone to affect political agenda-setting (Walgrave et al. 2008). In the qualitative analysis of this study it is possible to identify cases supporting this notion in issues initiated or promoted from social media platforms. The qualitative analysis presented here shows that crime and environmental issues are indeed sensational issues triggering social media campaigns and make them grow. Nonetheless, specific quantitative analyses on sensational issues on social media are required to observe stronger correlations between both variables. This study also contributes with a differentiated scale of political response. While this is an intuitive scale, it could serve as baseline for research measuring political agenda-setting affectation. By considering different degrees of political agenda-setting affectation, the study builds on previous

works suggesting the existence of two kinds of political responses, declarations (Thesen 2013) and actions (Green-Pedersen and Mortensen 2010). Moreover, the qualitative analysis allows observation of cases with different degrees of affectation. It is possible then, to observe that social media platforms sometimes work as a tool to strengthen what Smulovitz and Peruzzotti (2006) termed social accountability mechanisms.
This study also supports previous empirical work on trade-offs that occur between social media and traditional media, and how the former often affects the latter. According to the findings of this study, there is a strong correlation suggesting that social media campaigns are influencing traditional media outlets coverage. In contrast, this study suggests that traditional media are not influencing the size of social media campaigns. While these findings support previous work on the relation between both variables (Lotan et al. 2011), it is essential to bear in mind that cases in this study are inherently associated with campaigns that emerged from social media. As such, while social media is increasingly setting issues on traditional media, this is not a rule but a plausibly exception for topics with particular traits. A relevant contribution of this study to the literature is that left leaning media seems to be slightly more prone to react to social media campaigns than right leaning media. Since this study is located in Mexico, other country case studies might be necessary to observe whether this particular finding may apply to other contexts. Regarding the relationship between traditional media and political agenda-setting affectation, this study suggests that the former is not helping social media platforms in affecting the latter. Despite the findings in this study, they do not contradict previous political agendasetting studies suggesting that traditional media is an element affecting political agenda-setting (Thesen 2013). What this study suggests is that for these particular cases of social media campaigns, traditional media is not the variable explaining political-agenda setting affectation.

Both qualitative and quantitative analyses in this study were inconclusive regarding the influence of electoral events over social media campaigns size and political responsiveness.
Only two cases close to the electoral event received a high amount of online support and political response. However, the study also offers empirical evidence of small online campaigns proliferating around the electoral event. This might indicate that a politicized citizenry is initiating online campaigns with low online support and political response. An explanation for the lack of support and political response could be that online campaigns have to fight for attention with topics directly related with an electoral event. Since there are no previous works on how social media participation (other than electoral) might be affected by electoral events, these findings could be a baseline for future research.
Establishing a starting point for online campaigns could be essential for social media research in any field, but is particularly important for political science. The starting point of an online campaign, looking to affect political agendas, determines the political atmosphere where such topic is situated. Furthermore, it could determine the prospects of online support and political response. Hence, by observing the behavior of social media campaigns, this study found evidence to consider the first 10019 tweets in a single day as a starting point for these kinds of campaigns. Empirical evidence in this study showed that in the immediate days after the starting point, most of the online campaigns reached their peak. This evidence is similar to Jeffares (2014) study on a public campaign in United Kingdom. Online campaigns seem to follow a shocking logic, where first two weeks have stronger power than the following weeks when their power seems to dilute.
In sum, this study has linked previous works on social media platforms as an emerging tool for political participation and those on political agenda-setting. Contrary to previous works

on social media platforms, this study explores their impact from a multiple case perspective, relying on a combination of quantitative and qualitative analyses. Thereby, the most relevant finding in this study is that social media platforms can be counted as an element affecting the political agenda-setting on a more permanent basis. This contributes to previous works focused on rare events, which have already shown the political power of social media platforms under rare circumstances. For further research, this study has provided some findings of note, such as the proclivity of opposition toward online campaigns; as wells as the role of elections in promoting social media campaigns and at the same time plausibly discouraging online support and political responses to such campaigns. Another contribution to the field is empirical support with regard to the behavior of online campaigns, which may allow future research by providing a reference point. Finally, from a tangential point of view, the qualitative analysis allows to observe cases where social media is a tool to strengthen social accountability mechanisms.
The study conducted is an attempt to bridge two groups of studies currently disconnected: political agenda-setting and political social media. Within studies on political social media, this work contributes with a multiple case study on non-rare events, contrasting with current trends of single case studies. Within political agenda-setting studies, this study contributes by proposing social media platforms as an emerging element to take into consideration for understanding affectation on political agendas. In terms of methodology, the study is innovative not only because of the use of multiple cases, but by relying on search engines for political science research, such as and Twitter for case selection, data gathering, and measurement. It also proposes two methodological baselines for further research: as is the case of the starting

point of social media campaigns and a scale to capture declarative and active responses affecting political agendas.
Research on social media faces challenges that should be considered. First, due to the anarchic nature of social media and the relative novelty of the phenomenon, case selection is complex. Depending on the methodology for case selection, researchers could find only a few cases or an unmanageable amount of them. Here the challenge is to find a middle ground and reach a sample as homogenous as possible. Second, long-term analyses in social media are difficult due to the constant changing of online practices. For instance, even though Twitter began in 2006, it was not popular until around 2009; even then, smartphones were not as popular as they are in 2017. Third, platforms are constantly changing; sometimes they change their interfaces complicating access to reliable and comparable information along over time. Also, it is common that users change from one platform to another, even leading to the eventual disuse of some of them (i.e. MySpace or Hi5).
Regardless the abovementioned challenges, social media research in political science must increase in volume and depth. The link between social media, traditional media, and political agenda-setting is an unexplored area with interesting angles. This study examined only a few aspects of such link. Further research could address similar problems presented here but perhaps for a consolidated democracy or semi-democratic regimes, or even under authoritarian regimes where social media is not banned or highly censored. Also relevant is to understand how social media campaigns coexist with electoral campaigns during elections. In other words, how much politicians can afford neglecting online campaigns not directly related to electoral issues.
As suggested in this study, social media platforms are becoming increasingly influential over traditional media and key political decision makers. Sometimes these technologies work as

a tool for social accountability. This situation poses new opportunities and challenges for democratic governance. Even though this study focused on the positive side of social media, is important to bear in mind their limits and threats. On the one hand, social media does not substitute other forms of political participation and social mobilization. This study has shown that sometimes, even large online mobilizations are not enough to reverse political decisions. With regard to potential threats, social media has opened the door to spreading so-called fake news, which could harm institutional arrays, even in consolidated democracies. Another potential threat comes from regimes that are tempted to take advantage of new technologies to spy their constituents, particularly those in the opposition. The use of internet bots (software performing tasks over the internet to substitute human activities) and censorship are also important phenomena to weight when we think about the power of social media. Although these challenges were not prominent in this study, all of them were taken into consideration into the research design in order to minimize their negative effects.
Beyond the challenges above described, there is a fundamental problem with social media as a political participation tool, as well as, object of study for political science scholars. This problem has to do with restricted access to new technologies. According to a UN (2016) report on new technology access, 80% of the population in developed countries have access to internet, while this figure is only 40% for developing countries and 15% in less developed countries. In the case of Mexico, the gap mirrors the worldwide data: In Chiapas, the poorest state, only 17% of households have an access to internet, while in Mexico City the access is 63% (INEGI2015). These figures suggest that income is determinant of internet access. Then, only those with the means to afford internet access can be empowered through new technologies. Hence, if social media platforms can be used as a viable tool for social accountability and

political participation, as this and other studies suggest, then free and democratic access to these technologies is necessary.
1 Since Mexico does not have presidential reelection, the mid-term Federal Congress election is seen as a referendum to approve or disapprove presidents performance up to this point.
2 Some Mexican scholars have argued that the 2000 election does not precisely represent a democratic transition since most of the political structures of the one-party regime remains to this day. (Merino, Mauricio. 2003. Mexico: La transicion votada. America Latina Hoy, 33, 2003, pp. 63- 72, Ediciones Universidad de Salamanca.)
3 According to the Cambridge University Press database, Two Faces of Power is among the first five most cited articles of all times for the American Political Science Review.
4 For purposes of this article, the literature reviewed on political agenda-setting includes studies from 2008 to 2016 with the only objective of finding references to the role of social media in such process.
5 The number of Twitter users vary from one country to the other. Therefore, considering that Mexico has less users then a lower size of social media campaigns on these topics is expected.
6 Mexico ranks 123 out of 176 countries according to Transparency Internationals corruption perceptions index 2016. Mexico share such position with Azerbaijan, Djibouti, Honduras, Laos, Moldova, Paraguay, and Sierra Leone.
7 During the 2009 Federal Election in Mexico, an online movement calling for null vote achieved to increase for such political option and in the end helped to open the Mexican electoral system.
8 The requirements are still difficult to meet, but some bills have been already issued and even approved.
9 The PRI ruled the country for 70 years after the Mexican Revolution.
111 This date range has to do with the sample selected for this study (see Section 3).
11 For data collection two local think tanks were included, the International Affairs Mexican Council (COMEXI) and the Center of Research on Development (CIDAC). Both think tanks were among the most important in Mexico according to a study performed by the University of Pennsylvania, (see 2014 Global Go To Think Tank Index Report http//
12 For the complete dataset see Annex 1.
13 This study is based on original tweets since they represent a more demanding action than retweeting.
14 Point of agreement is defined as A document in which a position and a proposal is exposed regarding a given law, social, political, and/or economic conflict. Similar to a bill, it establishes the motivations, [and] the proposal specifications. (Camara de Diputados 2011) Points of agreement are similar to U.S. legislative resolutions, however, some procedural elements are binding details are different. All Mexican points of agreement require approval of both chambers. While points of agreement are not binding, they do require a formal answer from the actor or actors inquired by the document. Thus, since points of agreement are a sui generis type of legislation, for the purposes of this study such a term will be used instead of adopting resolution.
15 This is the English version of a tweet originally published in Spanish, (see
16 While there are some tools and marketing services allowing access to historic tweets, Twitter advanced engine search seems to be a reliable source for this kind of data. (See
17 There is no governmental search engine, thus these searches were performed through Google advanced search. To have certainty on the results, the search was limited to those sites with the termination which corresponds to all the Mexican government websites.
18 While this study analyses the influence of social media on the Mexican political agenda-setting, constraining tweets to only those from Mexico would neglect the fundamental principle of blurred boundaries in the use of new technologies.
19 100 tweets is an amount applicable to the case of Mexico and other countries with similar characteristics. For research under other contexts, might be needed other calculations considering factors such as population size, type of political regime, and access to new technologies, just to mention a few.

Abu Shihab, Lalla. 2016. "Todas las mujeres hemos sido acosadas sexualmente alguna vez" CNN en Espanol, October 18 http ://cnnespanol, eres-hemos-sido-acosadas-sexualmente-alguna-vez/
Ahmed, Azam. 2017. Mexicos President cancels meeting with Trump over Wall New York Times, January 26 (February 15,2017)
Aman, Mohammed M., and Tina J. Jayroe. 2013. "ICT, social media, and the Arab transition to democracy: from venting to acting." Digest of Middle East Studies 22, no. 2 (2013): 317-347.
Animal Politico. 2015. Entregan 20 mil firmas contra candidatura de Medina Mora a la Corte Animal Politico, March 5 (February 2, 2017)
Azuela, Maite. 2016. SEMARNAT en Tajamar: destruir antes que proteger. El Universal, January 25 (January 25, 2017)
Bachrach, Peter, and Morton Baratz. 1962. Two Faces of Power. American Political Science Review 56: 947-52.
Barbera, Pablo and Gonzalo Rivero. 2014. "Understanding the Political Representativeness of Twitter Users". Social Science Computer Review 2015, Vol. 33 (6) 712-729.
Barbera, Pablo, Ning Wang, Richard Bonneau, John T. Jost, Jonathan Nagler, Joshua Tucker, and Sandra Gonzalez-Bailon. 2015. "The Critical Periphery in the Growth of Social 10, no. 11 (2015): e0143611. Protests." PloS one
Bartolome, F. 2015. Templo Mayor Reforma, April 2 (January 20, 2017)
Bekkers, Victor, Henry Beunders, Arthur Edwards, and Rebecca Moody. 2011 New Media, Micromobilization, and Political Agenda Setting: Crossover Effects in Political Mobilization and Media Usage. The Information Society.
Bond, Robert and Solomon Messing. 2015. Quantifying Social Medias Political Space: Estimating Ideology from Publicly Revealed Preferences on Facebook. American Political Science Review, Vol. 109, No.l, February 2015.
Breuer, Anita and Jacob Groshek. 2014 Slacktivism or efficiency-increased activism?: Online political participation and the Brazilian Ficha Limpa anti-corruption campaign in Breuer, Anita and Yanina Welp (eds) 2014. Digital technologies for democratic governance in Latin America. New York: Routledge.

Camara de Diputados. 2011. Cuadernos de Apoyo: Terminologla Legislativa. (January 13, 2017)
Castells, Manuel. 2012. Networks of Outrage and Hope: Social Movements in the Internet Age. Polity.
Cervantes, Evlyn. 2015. Renuncia Korenfeld a Conagua Reforma, April 9 (January 20, 2017)
Deibert, Ronald and Rafal Rohozinski. 2012. Liberation vs. Control: The Future of Cyberspace in Diamond, Larry, and Marc F. Plattner (eds). 2012. Liberation technology: Social media and the struggle for democracy. JHU Press, 2012.
Diamond, Larry, and Marc F. Plattner (eds). 2012. Liberation technology: Social media and the struggle for democracy. JHU Press, 2012.
D.O.F. 2002. Ley Federal de Transparencia y Acceso a la Informacion Publica Gubemamental. 1/06/2002 (December 27, 2017)
D.O.F. 2012. Decreto por el que se reforman y adicionan diversas disposiciones de la Constitucion Politica de los Estados Unidos Mexicanos, en materia politica. (December 27, 2017)
D.O.F. 2016. Constitucion Politica de los Estados Unidos Mexicanos (January 12, 2017)
Dresser, Denise. 2016. Exigimos #PeriodoExtraordinarioYA para aprobar el Si sterna Nacional Anticorrupcion #Ley3de3 (January 20, 2017)
Emarketer. 2016. Twitter's User Base in Latin America Continues to Grow May 6 (December 23, 2017)
Fenton, Natalie, and Veronica Barassi. 2011 "Alternative media and social networking sites:
The politics of individuation and political participation." The Communication Review 14, no. 3: 179-196.
Fonatur. 2016. FONATUR cumplira programa de rescate en malecon Tajamar en estricto apego a derecho February 5 (January 25, 2017)
Green-Pedersen, Christoffer, and Jesper Krogstrup. 2008. Immigration as a Political
Issue in Denmark and Sweden: How Party Competition Shapes Political Agendas. European Journal of Political Research 47 (5): 610-34.

Green-Pedersen, Christoffer, and Peter B. Mortensen. 2010. Who Sets the Agenda
and Who Responds to It in the Danish Parliament? European Journal of Political Research 49 (2): 257-81.
Green-Pedersen, Christoffer and Stefaan Walgrave (eds). 2014. Agenda setting, policies, and political systems: A comparative approach, The University of Chicago Press.
Groshek, Jacob and Ingrid Bachman. 2014. A Latin Spring? Examining Digital Diffusion and Youth Bulges in Modeling Political Change in Latin America, in Breuer, Anita and Yanina Welp (eds) 2014. Digital technologies for democratic governance in Latin America. New York: Routledge.
Gutierrez, Silvia and Yazmin Cuevas. (2012) Representaciones sociales de Enrique Pena Nieto, candidato a la presidencia de Mexico 2012-2018, en la prensa escrita (Social representations of Enrique Pena Nieto, 2012-2018 Mexican Presidential candidate) Cultura y Representaciones Sociales, Revista electronica en Ciencias Sociales, ano 7, numero 13, pp. 63-95.
IMCO. 2016. Segunda entrega de firmas de la Ley 3de3 al Senado de la Republica (December 10, 2016)
INEGI. 2015. Encuesta Nacional sobre Disponibilidad y Uso de Tecnologias de la Informacion en los Hogares (ENDUTIH) lados/hedo220.xlsx (March 10, 2017)
International Transparency. 2016. Corruption Perception Index" hRegionalTables.xlsx (January 29, 2017)
Jeffares, Stephen. 2014. Interpreting Hashtag Politics Policy Ideas in an Era of Social Media, Palgrave Macmillan UK.
John, Peter, Shaun Bevam, and Will Jennings. 2014. Party Politics and the Policy Agenda: The Case of the United Kingdom in Green-Pedersen, Christoffer and Stefaan Walgrave (eds). 2014. Agenda setting, policies, and political systems: A comparative approach, The University of Chicago Press.
La Jornada. 2015a. Aristegui, MVS y la libertad de expresion March 17 (December 27, 2017)
La Jornada. 2015b. Korenfeld: Error inexcusable, usar nave de Conagua con mi familia. April 2 (December 27, 2017)
La Jornada. 2015c. Huye de Veracruz por amenazas y lo asesinan en el DF August 2 (February 10, 2017)
La Jornada. 2016a. Violencia de genero: normalizacion ominosa April 26 (January 16, 2017)

La Jornada. 2016b. La ley 3de3 podria desatar caceria de brujas: PRI April 22 (February 2, 2017)
Lancendorfer, Karen M. & Lee, Byoungkwan. 2010. Who influences whom? The agendabuilding relationship between political candidates and the media in the 2002 Michigan governor's race. Journal of Political Marketing, 9(3), 186-206.
Ley3de3. 2016. Iniciativa ciudadana de Ley 3de3 (February 2, 2017)
Lotan, Gilad, Erhardt Graeff, Mike Ananny, Devin Gaffney, and Ian Pearce. 2011. "The Arab Spring, the revolutions were tweeted: Information flows during the 2011 Tunisian and Egyptian revolutions." International journal of communication 5 (2011): 31.
Lozano, Genaro. 2016. Primavera de mujeres Reforma, April 26 (February 12, 2017)
McGann, James G. 2015. "2014 Global Go To Think Tank Index Report". TTCSP Global Go To Think Tank Index Reports. 8 University of Pennsylvania. (January 29, 2017)
MacKinnon, Rebecca. 2012. "China's 'Netwoerked Authoritarianism'" in Diamond, Larry, and Marc F. Plattner (eds). 2012. Liberation technology: Social media and the struggle for democracy. JHU Press, 2012.
Madrazo Lajous, Alejandro. 2015 jNo hagan Ministro de la Suprema Corte a Medina Mora! (February 2, 2017)
Mendez, Ernesto. 2016. Nueva victoria ciudadana en el caso Tajamar. Excelsior, March 3 (January 25, 2017)
Mendez, Alfredo. 2015 Expertos cuestionan la postulacion de Medina Mora a la SCJN: no es jurista La Jornada, February 20 (February 2, 2017)
Merino, Mauricio. 2003. Mexico: La transicion votada. America Latina Hoy, 33, 2003, pp. 63-72, Ediciones Universidad de Salamanca.
Mortensen, Peter. 2009. Political Attention and Public Spending in the United States.
Policy Studies Journal 37 (3): 435-55.
Morozov, Evgeny. 2012. The net delusion: The dark side of Internet freedom. Public Affairs, 2012. Harvard.
Mujeres sin Violencia. 2016a. ^Sabias que el primer acoso de las mujeres sucede en promedio a los 8 anos? Conoce todos los datos de #MiPrimerAcoso

mujeres-sucede-en-promedio-a-los-8-anos-conoce-todos-los-datos-que-se-recopilaron-del-hashtag-miprimeracoso (January 16, 2017)
Mujeres sin Violencia. 2016b. Yo te creo: el primer paso para la deteccion del abuso sexual infantil. April 29 (January 16, 2017)
Nicolas, Tasneen H. 2016. Protestan ante Semarnat por ecocidio El Universal, January 26. (December 8, 2016)
Odonell, Guillermo. 1999. Horizontal Accountability in New Democracies in Schedler,
Andreas. 1999. The self-restraining state: power and accountability in new democracies. Lynne Rienner Publishers.
Oser, Jennifer, Marc Hooghe, and Sofie Marien. 2013. "Is online participation distinct from offline participation? A latent class analysis of participation types and their stratification." Political Research Quarterly 66.1: 91-101.
Pantoja, Sara. 2017 Sentencias a 315 anos a expolicia por caso Narvarte Proceso, January 19 (January 29, 2017)
Pardinas, Juan E. 2016. El desayuno de mis palabras Reforma, April 3 (February 3, 2017)
Paullier, Juan. 2015. #MiPrimerAcoso, la creadora del hashtag que sacudio internet y la importancia de que las mujeres no callen. BBC, April 25 _violencia_mujeresjp (January 20, 2017)
Petrocik, John R. 1996). Issue-ownership in presidential elections with a 1980 case study. American Journal of Political Science 40(3): 825-850.
PROFEPA. 2016. Verificara PROFEPA si remocion forestal en proyecto tajamar, cumplio condicionantes de SEMARNAT January 26
Quintero, Josefina, Poy, Laura and Gomez, Eirinet. 2015 Asesinan en el DF a fotoperiodista que huyo de amenazas en Veracruz La Jornada, August 2 (February 20, 2017)
Reina, Elena 2016. #MiPrimerAcoso: testimonios contra el drama de la mujer en Mexico El Pais, June 15 ml (January 20, 2017)
Secretaria de la Funcion Publica. 2015a. David Korenfeld tendra hasta cinco dias, a partir del lunes, para ejercer su derecho de audiencia: Virgilio Andrade April 2 (January 4, 2017)

Secretaria de la Funcion Publica. 2015b. Resolution Expediente: 0016/2015 April 30 (January 4, 2017)
SEMARNAT. 2016. Informa SEMARNAT sobre proyecto Malecon Cancun Tajamar. January 20 (January 25, 2017) (January 25, 2017)
Scherman, Andres, Arturo Arriagada, and Sebastian Valenzuela. 2015. "Student and
environmental protests in Chile: The role of social media." Politics 35, no. 2 (2015): 151-171.
Shirky, Clay. 2011. "The political power of social media: technology, the public sphere, and social change." Foreign Affairs.
SIL. 2015. Proposition con punto de acuerdo por el que se exhorta al presidente de la republica a realizar las acciones conducentes para solicitar la renuncia del titular de la Comision National del Agua April 8 1428508400.pdf (January 20, 2017)
SIL. 2017a. Resultados de la busqueda de asuntos legislativos Acoso Sexual (January 16, 2017)
SIL. 2016b. Resultados de la busqueda de asuntos legislativos Tajamar (January 25, 2017)
SIL. 2016c. Resultados de la busqueda de asuntos legislativos 3de3 (February 2, 2017)
SIL. 2017d. Resultados de la busqueda de asuntos legislativos Ruben Espinosa (February 10, 2017)
Smith, Peter H. 2005. Democracy in Latin America: Political Change in Comparative Perspective, Oxford University Press.
Steinberg, Alan. 2015. Exploring Web 2.0 Political Engagement: Is New Technology Reducing the Biases of Political Participation?" Electoral Studies 39 May 2015.
Thesen, Gunnar. 2013. When Good News Is Scarce and Bad News Is Good: Government Responsibilities and Opposition Possibilities in Political Agenda-Setting.
European Journal of Political Research 52 (3): 364-89.

Tremayne, Mark. 2014. "Anatomy of protest in the digital era: A network analysis of Twitter and Occupy Wall Street." Social movement studies 13, no. 1 (2014): 110-126.
Tresch, Anke, Pascal Sciarini, and Frederic Varone. 2013. The Relationship between
Media and Political Agendas: Variations across Decision-Making Phases. West European
Politics 36 (5): 897-918.
U.N. Broadband Commission for Sustainable Development. 2016. The State of Broadband: Broadband catalyzing sustainable development (March 10, 2017)
Walgrave, Stefan & Van Aelst, Peter. 2006. The contingency of the mass media's political
agenda setting power: Toward a preliminary theory. Journal of Communication, 56(1), 88-109.

CIDAC Center of Research for Development
FONATUR National Trust Fund for Tourism Development
MC Citizens Movement
MORENA National Regeneration Movement
PAN National Action Party
PANAL New Alliance Party
PES Social Encounter
PRD Democratic Revolution Party
PRI Institutional Revolution Party
PROFEPA Federal Attorney for Environmental Protection
PT Workers Party
PVEM Green Ecologist Party of Mexico
SEMARNAT Ministry of Environment and Nature Resources

Full Text