Keywords: Political behavior, participation and culture

Democratic systems require the participation of citizens, particularly the manifestation of preferences through the vote in elections, since the legitimacy of the authorities rests on the ability to represent majority sectors. The sustained contraction of voter registration represents a problem that calls both politicians and scholars (González et al, 2005; Torcal, 2003).
From the perspective of social networks analysis, numerous studies published in the last two decades have sought to understand the relationship the level of political disagreement to which individuals are exposed in their conversation`s networks and their electoral participation (Mutz, 2002; Huckfeldt Johnson and Sprague, 2003; McClurg, 2003; Eveland, 2009; Bello, 2012).
Political dialogue between individuals with divergent viewpoints regarding politics is considered a basic component of democracy, insofar as it makes it possible to understand the existence and validity of multiple positions around political issues, as well as to legitimize those that are imposed in public life (Mill, 1859; Arendt, 1968; Mutz, 2003; Huckfeldt and Sprague, 2003).
However, some scholars argue that the democratic benefits gained from communication between diverse points of view would sacrifice participation, something that Mutz (2002) called the "dark side" of disagreement. That is, as suggested by Lazarsfeld, Berelson and Gaudet (1968), and then Eliasoph (1998) and Mutz (2002, 2006), heterogeneous social networks would be associated with a lower probability of participating in elections. This could be due to a greater complexity in the decision process given the cross-pressures, as well as the fear of offending some member of the social network.
Other studies report less clear results. McClurg (2006) says that the relation between participation and heterogeneous conversation disappears when the characteristics of a neighborhood related to one's own ideas are incorporated. Leighley (1990), Rubenson (2005) and Nir (2011) find mixed and even contrary results in the relationship between both variables. Bello (2012) proposes that disagreement in mixed or heterogeneous networks does not have a statistically significant impact on participation and only the "universal opposition" generates a significant decrease in participation.
The study use data of the Chilean version of the Comparative National Elections Project (CNEP), a representative survey conducted in the last presidential election in 2017. The 1,600 people surveyed answered a battery of questions associated with their informal political conversation networks, among other subjects. This country represents a good case of study, since its electoral participation has experienced a paradigmatic downward trend since the return to democracy. In the 1988 plebiscite, 89.1% of those over 18 voted, but in the presidential election of 2017 only 46.67% (Servel, 2017).
Preliminary results show that political conversations between homogeneous positions is associated with higher levels of electoral participation when these conversations take place with relatives and neighbors, but the relation is not significant when political discussion is associated with friends and co-workers. Additionally, and consistent with the literature, there is a positive and statistically significant relationship between the frequency of conversations about the election and having participated on the election.