Introduction

Literature demonstrates that, in certain circumstances, sociodigital networks (SDNS, hereinafter) are capable of facilitating the political participation of their users, overcoming both their effects of depoliticization, when primarily used as spaces for entertainment, and minimal engagement activism (^{Casteltrione, 2017}; ^{Gil de Zúñiga et al., 2012}; ^{Halpern et al., 2017}).

Due to the ease and speed that the Internet provides for collecting information, SDNS can generate a sense of security on the people’s own opinions and knowledge regarding topics of interest, which can translate into participation in political discussions. This happens particularly with users who have little political knowledge or are younger and tend not to consume traditional media (^{Gil de Zúñiga et al., 2016}). Evidence indicates that, as citizens spend more free time online, unintended consequences arise regarding political participation and discussion. Thus, everyday online conversations can eventually turn into a strategic form of communication and action (^{Gil de Zúñiga et al., 2016}).

However, one of the less explored aspects is the type of platform that has a greater effect in terms of its technological-communicative affordances, and whose algorithms enable certain types and degrees of depth in communicative exchanges.

A theoretically and normatively rich way to approach these problems is through the concept of deliberation. Understood as an informed and reasoned exchange of arguments among individuals about a topic that concerns them, and as a process that inherently implies and enhances participation (^{Gastil, 2008}), theorists of online deliberation argue that each digital platform contains certain structural characteristics that either foster or inhibit it, depending on the algorithms that govern them (^{Friess & Eilders, 2015}). This so-called discursive architecture enables a certain deliberative potential or degree of deliberativeness, which in turn influences user participation.

In recent years, SDNS have become popular for political purposes in Mexico, as they reduce campaign communication costs and allow political proposals to reach disengaged audiences. However, their relation to political participation is uncertain in the midst of a political environment that is resistant to it, which erects barriers through legal loopholes from State institutions and through the indifference of elected representatives towards their constituents in decision-making (^{Serrano, 2015}). It is also an environment characterized by a passive political culture; the majority of Mexicans do not participate in petitions, demonstrations, or political parties (^{Temkin & Ivich, 2013}). Voting remains the primary and almost sole form of participation exercised by citizens.

Nevertheless, the Internet boasts a significant political use: information search is the third most time-consuming activity for citizens (^{Instituto Nacional de Estadística y Geografía [INEGI], 2020}), while the most frequent forms of participation are reading or sharing political information on SDNS (^{Somuano & Nieto, 2017}). Case studies, particularly those related to digital activism, confirm an extraordinary political appropriation of SDNS by certain citizen groups, both during elections, for monitoring or defending liberties (^{Zires, 2014}) and in social movements (^{Slimovich & Lay Arellano, 2018}), although their conclusions are limited to specific populations.

In light of these promising glimpses and amidst the previously mentioned political and structural limitations, the objective of this article is to investigate the effect of social networks on political participation, but from a specific consideration of the deliberative potential of different platforms in this effect.

Political participation and social media

Extant empirical research on the possibility that SDNS’ raise levels of political participation (see ^{Boulianne, 2020}, for an analysis of 320 studies on this matter) assumes that the latter is important for democracy, while implicitly recognizing that pre-digital modes of participation have limitations that online networks can compensate for. Nevertheless, it is necessary to unravel these assumptions, as well as to analyze the Mexican case, where this relationship takes on a specific meaning in relation to the cultural and institutional obstacles that limit participation.

Participation is defined as the power and capacity of every person to act upon the world in which they live in order to improve the quality of public life (^{Valenzuela et al., 2009}). In the political realm, it can be developed both in the domain of institutional politics (political participation) and in the citizen sphere of rights, through non-institutional channels (citizen participation). The former implies citizen intervention to designate their leaders and influence the development or implementation of public policies. It is generally led by formal State institutions and instruments and includes conventional actions, such as voting, joining political parties, participating in demonstrations and rallies, among other activities, as well as direct democracy mechanisms that link citizens with public administration, such as legislative initiatives, referendums, or citizen consultations (^{Aguirre, 2013}).

On the other hand, civic participation refers to citizens’ influence in matters of public or political interest through methods that are external to political institutions. Here, the occupation of public space is more evident, with the aim of influencing the State to expand political spheres of action, assert existing rights, or demand the recognition of new rights. This sometimes involves unconventional, tougher, or coercive actions, such as boycotts, protests, blockades, or marches. Social movements represent the highest expression of this phenomenon (^{Aguirre, 2013}).

As for online political participation, it includes several of the aforementioned aspects, so it can be considered both civic and political. Defined as any political activity that distinguishes itself from offline participation and takes place in the Internet or social media (^{Gil de Zúñiga et al., 2016}), it is characterized as a fundamentally discursive activity. However, due to the connectivity characteristics of SDNS, which connect users with other users or content (^{van Dijck & Poell, 2013}), it can enable the organization of political action beyond the Internet, with consequences in the real world.

Various activities are encompassed under this concept, in an increasingly broad repertoire of both passive and active modes of use. Passive activities include involuntarily receiving e-mails, propaganda, and opinions. Active practices include searching for, forwarding, and disseminating political content, whether obtained from written or visual journalistic sources (photos or videos) or from third parties; commenting, expressing opinions, or discussing political issues on personal platforms such as blogs, as well as on other contacts’ platforms or ideologically aligned closed groups; donating money to groups or movements through digital means; writing e-mails to public representatives; or organizing protests, boycotts, and votes around various causes (^{Gil de Zúñiga et al., 2012}).

As a relatively young democracy, political participation in Mexico is contrasting. On one hand, the figures for voter turnout in federal elections demonstrate relative stability: while voter turnout was 58.2% in 2006, it was 63.1% in 2012, 63.4% in 2018 (^{Instituto Nacional Electoral [INE], 2018}), and 52.66% in 2021 (^{INE, 2021}). However, beyond voting, other forms of participation are rather weak and declining, as highlighted by ^{Nieto and Somuano (2020)}, who examined various indicators from 1994 to 2014. These indicators include petition signatures (31% to 18%, respectively), boycotts (6% to 2.5%), demonstrations (10% to 2%), strikes (6.6% to 5.5%), and occupation of public buildings (4.6% to 4.2%). These figures are significantly lower compared to developed countries and even other Latin American countries like Argentina, Brazil, and Chile. However, indicators of community civic participation, such as helping to solve community problems, attending school meetings, or participating in women’s groups, show similar frequencies to those of Latin American countries and even advanced democracies. This contrast suggests that in Mexico, rather than apathy, there is a sense of alienated or excluded civic participation from institutional channels and instances (^{Nieto & Somuano, 2020}).

In this context, citizen’s participation has activated in social media over the past two decades. As previously stated, information search, in general, ranks as the third most time-consuming activity for citizens (^{INEGI, 2020}). Additionally, other researchers found that the most common forms of political participation in the country, regardless of socioeconomic status, are reading or sharing political information on social media and collaborating in political party activities (^{Somuano & Nieto, 2017}). These findings, along with interpretive studies or studies focusing on specific populations (see below), provide indications of the potential of the Internet to foster citizen participation. The following section provides empirical support for this assertion while discussing preferred technological modalities to achieve it.

Political and deliberative uses of social media. Main assumptions

As originally conceptualized, online participation encompasses all Internet resources in service of political activity. However, a significant aspect of online participation that has experienced exponential growth in terms of information sharing and public engagement is Social Digital Networks (SDNS). This technology was specifically designed for participation after the era of Web 1.0, which was focused on the unlimited availability of information. Regarding a specifically political use of SDNS, literature highlights the broad availability and zero cost of political information on these platforms, which in turn reduces the cost for citizens to acquire, expand, and deepen their knowledge of public affairs in terms of time and effort. Consequently, this affordance could generate interest in political matters and provide the necessary selfconfidence for citizens to engage, either modestly by commenting or expressing opinions, or in a more articulated manner, such as through digital activism (^{Farrell, 2012}).

Furthermore, it is more likely that this information will be consumed since it is shared among previously known contacts and trusted individuals. This may result in inadvertent consumption, meaning that users may not actively search for or initially be interested in the information, but it can become relevant under these circumstances (Boulianne, 2015). This familiarity with network members, on the one hand, encourages continuous commenting and opinion exchange based on the information, while also gradually increasing the flow of information for the user as its network expands (^{Gil de Zúñiga et al., 2012}). In addition to these processes triggered by the widespread distribution of information, SDNS can provide a platform for debate and discussion, where the exchange of information, ideas, and comments crystallizes in a public space for deliberation (see below) (^{Lilleker & Vedel, 2013}).

In spite of those possibilities, SDNS are not exempt from politically unfavorable uses. They are populated, to a large extent, by passive users (lurkers) or those with low commitment who engage only peripherally -certain research suggests that 90% of the content is produced by 1% of users- (^{Jenkins et al., 2015}). Other studies confirm that the recreational uses of Facebook, which occupy a significant portion of users’ exposure time -especially among apolitical users- show a negative correlation with political and civic participation, effectively distancing them from politically relevant topics (^{Casteltrione, 2017}). Above all, there is evidence of the low quality of dialogue that occurs in these spaces, raising doubts about their real potential to stimulate democratic participation (^{Strandberg, 2008}). This phenomenon is evident, for example, in the comment sections of Mexican digital newspapers, which are often oriented towards falsehoods and invectives (^{Frankenberg, 2015}).

A cultural explanation of these findings would attribute polarization, conflict, or incivility to the political culture of users, understanding how culture or society shape the use of technology. In contrast, a perspective of “technological shaping”, but not determinism, emphasizes how a particular technology constrains but also enables certain actors and social practices (^{Schroeder, 2018}). This latter position has been developed by scholars of online deliberation, who propose the concept of discursive architecture. This is defined as a package of technological features that offer opportunities for expression and interaction, as well as possibilities and opportunities for online political action. These affordances either enable or limit the performance of democratic norms in the communicative realm and, consequently, the scope of public deliberation in the digital environment (^{Chekunova et al., 2016}; ^{Freelon, 2015}; ^{Halpern et al., 2017}; ^{Vaast et al., 2017}).

It is argued that this is a promising position to explain the technological antecedent conditions of online participation, not only because it has been empirically demonstrated that the experience of traditional deliberation in small groups steadily increases civic activities (^{Gastil et al., 2002}), but also because it allows postulating certain properties in platforms, or affordances, that enable more or less frequent and sophisticated participation. Defined as a collective practice that “carefully examines a problem and arrives at a well-reasoned solution after a period of respectful and inclusive consideration of diverse points of view” (^{Gastil, 2008, p. 9}), deliberation could be strongly mediated by the discursive architecture of socio-digital platforms, whose algorithms introduce differences in communication possibilities among participants.

For example, there are differences regarding the ability of each network to distribute relevant and complex information and foster deliberative interaction. Twitter, originally created as a “microblogging” platform, only allows 280 characters, limiting the breadth of arguments, while YouTube, on the other hand, favors audiovisual content, which apparently hinders verbal reasoning. Both of these disadvantages are absent in the Facebook network. Similarly, Facebook enhances interaction among users and the expansion of contact networks, even with non-close circles, as its algorithm favors the positioning of posts that generate conversation among users and expand their contact network. Conversely, this latter characteristic is absent in YouTube, which emphasizes personal content preferences over the activity of the contact network. In summary, due to the specificity of the technological-discursive characteristics of social platforms, it can be considered that each social media platform has a distinct deliberative potential and, therefore, would generate differentiated effects on citizens’ political participation.

On the other hand, the technological conditions for social platforms to enhance deliberation in networks are summarized in five attributes by ^{Friess and Eilders (2015)}. Firstly, the distinction between synchronous communication, which takes place simultaneously, and asynchronous communication, occurring at different times, is crucial. Synchronous communication, while attracting more participation, can lead to interactions that are incoherent, frivolous, or less inclusive. In contrast, asynchronous communication enables individuals to reflect on and substantiate their arguments, fostering more deliberate and justified exchanges. This approach prioritizes participants’ convenience and broader interactions, rather than fixating on a specific topic. Asynchronous communication more closely aligns with deliberative dimensions of rationality, civility, and inclusivity.

Secondly, there is anonymity versus user identification. The former can allow participants to express their opinions freely and authentically without fear of ridicule, although it could also encourage disrespectful behavior and reduce the quality of the discussion.

Thirdly, the role of platform moderation is paramount. This encompasses actions such as censorship, involving the removal of inappropriate comments, as well as facilitating dialogue by offering a synthesis of the discussion or spotlighting minority opinions. These practices adhere to the principles of deliberation, emphasizing rationality and inclusivity. However, the presence of a perceived censor moderator could decrease users’ motivation to participate.

Fourthly, communication empowerment pertains to the provision of tools that enhance dialogue, enabling users to acknowledge their influence on the discourse. This is exemplified by means of group chats, video conferences, direct responses to comments, and gaining insight into users’ opinions through surveys or the selection of pre-determined “emotion” buttons on social networks.

Fifthly, the division of labor holds significance, where the platform enables intricate subjects to be subdivided into smaller groups as chosen by users, thus facilitating discussions. Finally, mechanisms to provide information and motivate users to share it are important, as they increase knowledge and rationality in the discussion.

The theoretical assumptions and possible relationships previously described still lack empirical evidence in the Mexican case. Studies concerning the correlation between social media usage and participation primarily center around digital activism, with a particular focus on movements such as #YoSoy132 or Ayotzinapa (^{Meneses, 2015}; ^{Slimovich & Lay Arellano, 2018}). However, it’s important to note that broad conclusions should not be drawn exclusively from these cases.

Another set of quantitative studies focuses on university populations, which show a low political appropriation of technology, resulting in limited concrete actions and outcomes (^{Padilla, 2015}; ^{Ruelas, 2016}), as well as low levels of online participation (^{Torres et al., 2020}). Furthermore, the predominant forms of participation among this population are passive and largely related to political entertainment (^{de la Garza et al., 2019}). However, other research does find positive relationships between digital media consumption and online participation (^{de la Garza & Pineda, 2018}).

While some national studies confirm a positive, albeit modest, relationship between legacy media consumption and political participation (^{Díaz, 2017}; ^{Muñiz & Echeverría, 2020}; ^{Temkin & Ivich, 2013}), only the studies by ^{Echeverría & Meyer (2017)} and ^{Moreno & Mendizábal (2015)} demonstrate the lack of association between political participation and the use of social media, particularly among young populations, with a very modest offline mobilizing effect. Furthermore, we are not aware of a national-scale study that focuses on the relationship between the use of social media platforms and political participation, particularly considering the discursive architecture of each platform.

In light of the latter, we address this research gap through two research questions, which we empirically answer in the rest of the article:

Method

The initial phase of the research compared the social media platforms Facebook, Twitter, and YouTube, which are the most investigated in the literature, in order to identify the presence or absence of online deliberation characteristics and demonstrate their deliberative potential. Based on the criteria described earlier (^{Friess & Eilders, 2015}), a checklist was established to verify these attributes, and their presence (1) or absence (0) was recorded, resulting in an index of deliberative potential or deliberativeness for each platform (Table 1). This index allows us to observe which platform would hypothetically have a greater effect on participation, albeit only in an ordinal manner.

Subsequently, a survey-based methodological design was used to examine whether the utilization of these platforms influences citizens’ on and offline participation. For the survey implementation, the services of the company QuestionPro were contracted, as they curated a panel of participants at the national level. The fieldwork was conducted between April 10^th and 16^th, 2021, during the campaign period for the federal legislative elections.

The sample (N = 1 750) consisted of participants who were adults (≥ 18 years) and registered as voters in the Republic, across all of the 32 states. The sample consisted of 44.7% males (n = 782) and 55.3% females (n = 986), with ages ranging from 18 to 80 years (M = 40.40, SD = 14.36). Participants with different educational levels were included in the sample, with the majority having a college degree (n = 919, 52.5%) or high school education (n = 551, 31.5%). In this survey, socioeconomic levels were measured instead of income, using the following distribution: AB (n = 26, 1.5%), C+ (n = 315, 18%), C (n = 346, 19.8%), C- (n = 266, 15.2%), D+ (n = 707, 40.4%), and D (n = 90, 5.1%). Regarding the variables and items, the instrument developed by ^{Rodríguez et al. (2014)} was used. As independent variables, participants were requested to specify the frequency with which they utilized Facebook, Twitter, and YouTube to acquire information about the electoral campaign. This was measured on a scale ranging from “never” (1) to “very frequently” (5). As dependent variables, the constructs of offline political participation (6 items), offline civic participation (8 items), and online political participation (12 items) were measured, as described in Table 2. Control variables included gender, age, income level, education level, and consumption of traditional media (print, digital, radio, and television).

To estimate the statistical effects, a multiple hierarchical regression analysis was conducted to predict the effect of social media use on users’ online and offline participation. Three models were run, one for each type of participation, yielding satisfactory Durbin Watson tests for no autocorrelation for each model (2.012 for online participation, 1.961 for civic participation, and 1.972 for political participation).

Findings

First of all, it is noticeable that there is practically no difference in the average, on a scale of five, between offline civic participation (2.56) and online participation (2.36), while both differ considerably from offline political participation (1.84). Conversely, the impact of aggregated social media attention (SMA) on various forms of participation is relatively robust. After controlling for sociodemographic variables and traditional media attention, aggregated social media attention (SMA) exhibits a substantial influence on online participation (β = .259, p = .000), a moderate influence on political participation (β = .209, p = .000), and a modest influence on civic participation (β = .162, p = .000). These effects contribute to a respective increase in variance of 5%, 3%, and 2% across each model.

However, the breakdown of these elements by specific platform yields differentiated results that do not correspond to the expected order based on the degree of deliberativeness of each platform. We ran a model for each platform and type of participation, integrating sociodemographic and media control variables in each one (Table 3).

Regarding the effects of platform use on online participation, the model reveals an important influence of Twitter initially (β = .151, p = .000), followed by YouTube (β = .102, p < .000), while Facebook’s effect (β = .180, p = .842) is not statistically significant. It is worth mentioning that these results are somewhat lower than the effects of legacy media, as the variables of attention to news through print media (β = .204, p = .000) and digital media (β = .145, p = .000) were significant. The model as a whole account for a 4% increase in variance from the social media variables.

The relationships are similar in the model of civic participation. The effect of YouTube remains the most relevant (β = .082, p < .001), followed by Twitter (β = .037, p = .000), and Facebook, whose impact was not statistically significant. Again, these results are similar to those of legacy media, with print media at the forefront (β = .189, p = .000), followed by digital newspapers (β = .153, p = .000) and radio (β = .085, p < .001). In this case, the block of social media variables contributes 0.8% to the variance.

Regarding political participation, the results are, from an ordinal perspective, similar to the previous ones. YouTube has the highest impact (β = .083, p < .001), although this time it is less relevant than in other types of participation, and is even close to the impact of Facebook on online participation. Twitter follows in terms of impact (β = .046, p = .265), and Facebook comes next (β = .036, p = .173), although neither of them is statistically significant. With respect to traditional media, only print newspapers are significant (β = .287, p = .000). Social media variables account for 1% of the variance in the model.

On the other hand, there are participation practices that are more strongly influenced by social media platforms. In terms of political participation, the practice of “collaborating in a political cause” exhibits the highest increase, with significant effects attributed to both YouTube (β = 0.109, p = .000) and Twitter (β = 0.08, p < .001). The next practice, “mobilizing others to join a political cause”, is more notably influenced by YouTube (β = 0.100, p = .000) compared to Twitter (β = 0.078, p < .001).

On the other hand, there are participation practices that are more strongly influenced by social media platforms. When considering political participation, the practice that experiences the most substantial increase is “collaborating in a political cause”, with notable effects observed on both YouTube (β = 0.109, p = .000) and Twitter (β = 0.08, p < .001). Following this, the activity of “mobilizing others to join a political cause” is significantly influenced, with YouTube playing a more relevant role (β = 0.100, p = .000) than Twitter (β = 0.078, p < .001). Additionally, “participating in pressure or boycott actions due to disagreement in a political decision” demonstrates noteworthy effects on both YouTube (β = 0.086, p < .001) and Twitter (β = 0.073, p < .001). Facebook does not show significant effects in these three practices. However, the use of Facebook for political purposes is associated with “working for a candidate during campaigns” (β = 0.083, p<.001) and, above all, “attending political party meetings” (β = 0.17, p<.05).

At this level of impact, according to the coefficient size, only one civic participation practice stands out, which is “attending neighborhood meetings”, related to political engagement on YouTube (β = 0.098, p<.001) and Twitter (β = 0.084, p<.001). Two others show a moderate increase, specifically “making requests to authorities” (YouTube, β = 0.104, <.05; Twitter, β = 0.065, p<.000), and “attending meetings of a social group” (YouTube, β = 0.100, p = .000; Twitter, β = 0.064, p<.05). In the other six forms of civic participation, statistically significant but almost negligible effects are observed. It is worth noting that Facebook has no impact on these types of participation.

In the case of online participation, the practices influenced by social media platforms are widely differentiated among platforms. Of the effects of Facebook, only those related to “commenting on news or columns” (β = 0.087, p = .000) and “expressing a personal opinion on political matters” (β = 0.082, p<.05) are noteworthy. The rest of the expressions exhibit marginal effects. On the other hand, YouTube has a greater impact on other practices. The most relevant are “consulting political information” (β = 0.159, p< = .000) and “commenting on political news” (β = 0.147, p = .000). Other practices are smaller but still relevant: “mobilizing contacts around a social cause” (β = 0.126, p = .000), “attempting to solve neighborhood or community problems” (β = 0.118, p = .000), “expressing opinions on political matters” (β = 0.111, p = .000), and “being part of a social or community group” (β = 0.111, p = .000).

Twitter, on the other hand, has a more significant impact on certain practices, such as “signing (β = 0.212, p = .000) or creating (β = 0.166, p = .000) online petitions”, “expressing personal opinions on community issues (β = 0.189, p = .000) and political issues” (β = 0.181, p = .000), and “commenting on news about politics” (β = 0.159, p = .000). At the next level of impact, there are practices like “initiating or being part of a social (β = 0.134, p = .000) or political (β = 0.129, p p = .000) group”, “attempting to solve community problems” (β = 0.122, p = .000), “mobilizing contacts around political causes” (β = 0.115, p<.001), “consulting political information” (β = 0.113, p = .000), “supporting protests or mobilizations” (β = 0.12, p = .000), and “mobilizing contacts around a social cause” (β = 0.109 p = .000).

Discussion and conclusions

The aim of this study was to investigate the potential connection between the utilization of social media platforms and various forms of political participation within the Mexican electoral context. The research sought to differentiate the impact of different platforms, operating under the assumption that each platform possesses a distinct discursive architecture that influences deliberation and, subsequently, political participation.

Firstly, it is possible to observe that social media platforms as a whole, as well as each individual platform, have a positive but modest impact on different types of participation, particularly online participation, followed by political participation, and finally civic participation. This contradicts findings at the national level (^{Moreno & Mendizábal, 2015}) regarding the lack of association between social media use and participation, as well as the almost negligible offline mobilizing effect, as found in studies conducted among university populations (^{Padilla, 2015}; ^{Ruelas, 2016}; ^{Torres et al., 2020}). However, our results align with those of ^{Garza and Pineda (2018)} and ^{Rodríguez-Estrada et al. (2019)}.

Nevertheless, it is necessary to acknowledge that participation remains a lagging practice in Mexico. On the one hand, its frequency is moderate across the three measured types of participation, and on the other hand, online participation does not surpass traditional civic participation, which represents the latent type of community participation described earlier, despite being more accessible (^{Nieto & Somuano, 2020}). This may imply that the political appropriation of this technology remains constrained, in line with the passive cultural legacies and legal obstacles previously described. Furthermore, it’s important to emphasize that the impacts of social media platforms often mirror, and at times fall short of, those of traditional media, particularly print media, digital media, and radio. Despite the increasing usage and expansive interactive potential of social media, these relatively novel platforms appear to be somewhat less adept than “traditional” media in fostering participation.

On the other hand, there are substantial differences regarding the effects of each platform on the types of participation and specific practices. Facebook does not show significant effects on the three types of participation at an aggregate level and only has very small effects on two partisan political participation practices and two commenting practices. It appears, as we mentioned, to be a platform primarily used by passive users for entertainment purposes or by users managing party-related groups. On the one hand, Twitter has minimal influence on civic participation in the aggregate and weakly influences three specific practices of this type of participation (see Table 2). Its influence on political participation is non-existent at the aggregate level, and only in three practices does it have a barely marginal influence. However, the platform significantly stimulates practices such as expression, affiliation with causes or groups, and online commenting. In the aggregate, it has the greatest overall effect on online participation by a significant margin. Twitter appears to stimulate broad political activity within its own environment but very little outside of it.

Likewise, in the aggregate, YouTube demonstrates a moderate impact on civic participation and stands as the sole platform with an effect on political participation. This observation hints that the credibility of the latter, conducted via institutional avenues, might be compromised to an extent where these technologies are unable to enhance its level. (see Table 3). Political information consumption on YouTube also influences practices of association among neighbors and groups. The platform’s impact on online participation practices is also noteworthy, as two of them -political information consumption and commenting- show a strong effect, while four show a moderate effect.

In this regard, the data run counter to our assumption that platforms with more deliberative attributes would result in a more substantial impact on participation. Although Facebook is by far the platform with the greatest deliberative potential, its use does not increase participation. Unexpectedly, it is YouTube, with lower capacities for interaction and deliberation, that has the greatest effect on civic, political, and particularly online participation

In fact, the hierarchy of deliberative potential differs from its impact on real participation. Contrary to our categorization, the order of influence on participation is as follows: YouTube, Twitter (with half the coefficient value), and Facebook.

The most parsimonious explanation for these findings is that users do not utilize the platforms for deliberation purposes, so the potential of their discursive architecture is not exploited. If this is the case, the attribute that seems most relevant in the detected effects is the platforms’ capacity for abundant, free, and networked distribution of political information, rather than the features that enable interactions, discussions, and mobilization among users, which would translate into participation (^{Lilleker & Vedel, 2013}). For example, YouTube is a platform capable of distributing copious political information with high levels of engagement, thanks to its audiovisual nature, unlike Twitter, which is primarily text-based.

Thus, we could be facing a phenomenon of media effects on political participation that is very similar to what occurs in traditional media such as the press, radio, and television, according to international literature and previous national studies (^{Díaz, 2017}; ^{Muñiz & Echeverría, 2020}; ^{Temkin & Ivich, 2013}). Platform attributes such as active communication, synchronous communication, or thematic compartmentalization do not seem to be more relevant features than traditional media information when it comes to stimulating different forms of participation, at least in the cultural and institutional conditions of the country. However, we should not dismiss the effects of certain platforms on certain participation practices, where their public utility may lie.

This research has some limitations. As evident, it does not measure the deliberative activity of social media users, and in that sense, it cannot link usage data with effects but can only approximate the former. Secondly, the sample is somewhat skewed towards users with higher educational and socioeconomic levels, which somewhat nuances the preferred use of social media. Nevertheless, we believe that the findings can be relevant for an agenda focused, at a structural level, on how social media platforms can accelerate historical backlogs in political participation processes and, at a micro level, on the specific characteristics of the discursive architecture of the platforms that increase such participation. In a country like Mexico, with high levels of political disaffection and distrust, that agenda is both relevant and timely.