INTRODUCTION¹

A solid democracy requires steady institutions as well as a firm political system which may allow the alternation of power among the different political options by holding free and just elections. However, although such structures are needed to manage a system wholly democratic, they are not enough. Oftentimes a citizenry committed in a political and civic degree is considered a condition for the good functioning of democracy, it should be informed about the events which take place in the political sphere of the country (^{Buendía and Somuano, 2003}; ^{Delli-Carpini, 2004}; ^{Delli-Carpini and Keeter, 1996}). Citizens will be able to correctly express their opinions and appropriately articulate their political preferences inasmuch as they are informed (^{Elenbaas et al., 2014}).

The motivation and interest to acquire information about politics and the ability to process and accumulate it are key elements to structure the citizen's political knowledge. (^{Luskin, 1990}; ^{Zaller, 1992}). The consequences of political knowledge stored in citizens not only impact the existence of a citizenry of quality, but also the health of democracy upheld by society (^{de Vreese and Boomgaarden, 2006}). This encouraged different disciplines to study the generation process of political knowledge and its repercussions for the system. Usually in these works the concepts of political knowledge, political awareness and political sophistication are found as interchangeable or at least very related with each other (^{Popa, 2015}; ^{Rhee and Cappella, 1997}; ^{Zaller, 1992}). The latter, political sophistication, will be taken as a reference in this study, as it encompasses the other two.

Having sophisticated individuals in the political field is crucial for the good development of democracy, since it leads to the efficient exertion of rights which pertain to them as citizens. This is so because they have a more complete and exact knowledge, which is based on a wider and better organized and incorporated structures of mental knowledge on politics (^{Luskin, 1987} and ¹⁹⁹⁰; ^{Rhee and Cappella, 1997}). This enables them to make use of more and better cognitive resources in order to, for instance, vote thoughtfully or take part in the different political fields which democracy offers them (^{Buendía and Somuano, 2003}; ^{Hollander, 2014}; ^{Prior, 2005}). An ideal democracy should have a great number of highly sophisticated citizens, however, in reality this is not easy to occur. (^{Dassonneville, 2012}). Normally, less sophisticated citizens conform an important part of society, and it is not even surprising that the average citizen shows little interest for the events in the political field and so does not have a special motivation to look for information about politics.

Now, even if citizens were motivated to access information, (^{Dassonneville, 2012}; ^{Luskin, 1990}), it is not always easy to obtain it. In fact, political knowledge is generally acquired in a vicarious way through media (^{Elenbaas et al., 2014}; ^{Rhee and Cappella, 1997}), where in a great amount political campaigns and debates take place. Media, therefore, assume a significant role since they are a technique of transference and translation of political reality to citizens (^{Muñiz, 2012}). They constitute to a large extent the main source of information for most citizens (^{Cho et al., 2009}; ^{Rojas, 2006}; ^{Woolley et al., 2010}), being their information coverage on politics the currency with which democracies function (^{Gerth and Siegert, 2012}). Therefore, since they allow to establish a bound between the opinions of citizens on political subjects and their electoral decision, they contribute to governance.

Probably, one of the political moments in which media acquire a fundamental role as a mechanism of information is during electoral campaigns. In fact, it has been seen that during electoral campaigns citizens broaden their knowledge on public affairs (^{Craig et al., 2005}; ^{Hansen and Pedersen, 2014}), which contributes to make a knowledgeable decision when voting.

It is possible that these stages of political life in countries conform a propitious situation for the increase of civic political sophistication, not precisely the one linked with the degree of factual political knowledge (^{Weisberg and Nawara, 2010}), but rather the one related with the knowledge about campaign proposals, attitudes and candidates' competence who participate in the electoral battle (^{Stevens, 2005}). Although there is a great number of research on political sophistication linked with the citizens' factual knowledge, the number of research on the voter's sophistication (^{Van Heerde et al., 2006}) or electoral awareness (^{Schuck et al., 2013}), which in this paper is called electoral political sophistication, is lesser.

Following these premises, this paper sought to determine what importance the different communicative practices performed by citizens have -for instance, attention to political contents, use of web pages about politics, watching electoral debates, use of social networks and political conversation- in both the generation of factual political sophistication and electoral political sophistication. In order to do so, data from the surveys carried out in the Mexican states of Nuevo Leon (2015), Puebla (2016) and Mexico (2017) were used, in the context of electoral campaigns for gubernatorial elections of each State.

Literature used

Explanatory factors of political sophistication

It is usual to find in the literature of disciplines such as political sciences or communications, studies which have tried to establish which factors influence the increase or reduction of the citizens' political knowledge and political sophistication (^{Delli-Carpini and Keeter, 1996}; ^{Hollander, 2014}; ^{Luskin, 1990}; ^{Popa, 2015}; ^{Zaller, 1992}). The results of such researches propose a variety of elements which have an impact on the interest and political knowledge, and determine the degree of political sophistication held by the citizen. Even from his seminal work, Luskin (1990) noted the three main factors which determined the degree of sophistication. In this sense, it can be assumed that the sophisticated citizen is s/he who possess a cognitive ability, that is, tools to acquire and process the information of their sphere as well as an opportunity to be able to acquire that information; factors which determine how easily a citizen learns about their political sphere (^{Elenbaas et al., 2014}; ^{Gordon and Segura, 1997}).

Finally, ^{Luskin (1990}) pointed as a significant factor motivation or the citizen's desire to learn about what occurs in their sphere. This aspect explains to a large extent the individuals' seeking and paying attention to information in order to understand what occurs in the system. The fact of having an inherent motivation to stay informed combined with an acceptable amount of available information through sources such as media, clearly influences the citizen's process of learning, contributing to structure their political knowledge (^{Elenbaas et al., 2014}). For this reason, motivation has been associated as the degree of interest citizens possess regarding politics (Elenbaas et al., 2014; ^{Popa, 2015}). Under this assumption, a lack of motivation would suppose the first barrier to the process of acquiring information (^{Prior, 2005}) rather than the lack of information and abilities held by the citizen to process it.

Now, even though these variables motivation or ability may have a greater impact in the generation of political sophistication than in the one related to the research of information (^{Luskin, 1990}; ^{Prior, 2005}), the citizen, undoubtedly, needs to obtain information to structure their comprehension and knowledge about what occurs inside the system. In this research, Luskin (1990) included the consumption of conventional media, whose influence has been extensively studied.

In the case of the press, it has been constantly perceived that its consumption has a positive effect in the increase of the citizen's sophistication (^{Delli-Carpini and Keeter, 1996}; ^{Gordon and Segura, 1997}; ^{Guo and Moy, 1998}; ^{Hollander, 2005}; ^{Popa, 2015}; ^{Rhee and Cappella, 1997}); in the case of television, it has been observed that its consumption has a negative influence in the generation of political knowledge (^{Gordon and Segura, 1997}; ^{Guo and Moy, 1998}; ^{Rhee and Cappella, 1997}). Nonetheless, other studies have found positive relationships regarding television consumption (^{Hollander, 2005}; ^{Popa, 2015}) mainly due to the sort of contents received through this medium.

Thus, a difference is observed between the exposition to media and the attention to particular contents offered by them, such as news, political programs or electoral debates (^{de Vreese and Boomgaarden, 2006}; ^{Muñiz, 2012}; ^{Rojas, 2006}). In that regard, some authors note that the mere generic exposition to media is a weak sign of the citizen's political learning and knowledge (^{Rhee and Cappella, 1997}). Hence, we must pay more attention to the selective exposition to political contents, that is, attention to politics through conventional media. This variable has been pointed as a strong factor which accounts for the increase of political knowledge in citizens (^{Hollander, 2014}; ^{Rhee and Cappella, 1997}; ^{Rojas, 2006}), and even in audiences rather unmotivated in which a passive learning can be produced from the casual exposition to such contents (^{Prior, 2005}). This positive effect has been also perceived in electoral contexts, usually both attention to politics through conventional media (^{de Vreese and Boomgaarden, 2006}; ^{Thorson, 2014}) as well as attention to electoral debates held during the campaign (^{Drew and Weaver, 2006}; ^{Druckman and Leeper, 2012}) contribute to increase the knowledge about the matters debated and about the personal characteristics of candidates.

The advent of new media, in their different presentations, is slowly producing a migration of audiences from conventional media to online resources (^{Woolley et al., 2010}). Hence, authors such as ^{Hollander (2014}) remind us that in the new studies on political culture it is necessary to include items related to activities such as reading or the use of news and contents online since they constitute factors which explain the citizens’ political attitudes. Thus, several authors say that media and social networks are becoming for a great number of citizens a crucial resource in order to follow the political information and, therefore, to generate political knowledge (^{Cho et al., 2009}; ^{Drew and Weaver, 2006}; ^{Hollander, 2014}). This is because their consumption is not as demanding as other conventional media, which enables the user of such media to look for more active information thanks to a process of comparison between media contents; this has suggested that a levelling effect of the users' political knowledge can be generated (^{Hollander, 2014}; ^{Prior, 2005}). Such positive effect of the consumption of contents of the new media about political knowledge has been also identified in electoral campaigns (^{Drew and Weaver, 2006}; ^{Partheymüller and Faas, 2015}; ^{Woolley et al., 2010}).

Finally, the literature has also suggested an important effect on the knowledge and political sophistication of the variable related to the discussion or political conversation held among citizens, understood as a crucial variable for the political democratic development (^{Cho et al., 2009}; ^{Rojas, 2006}; ^{Thorson, 2014}). It should be reminded that the citizens' communicative practices do not circumscribe to media consumption, but also to the processes of intra and interpersonal reflection, generally understood as the political dialogue through conversation, which can have an impact by managing to make changes in the citizens' political attitudes. Thanks to these strategies of exchange of interpersonal or even interactive information through digital platforms (^{Yamamoto et al., 2015}) the political learning can be increased, serving this conversation as a catalyst factor of the civic deliberative education (^{Gastil and Dillard, 1999}). This is how, for instance, ^{Thorson (2014)} presents it, when finding that the exposure to news and the political discussion had a positive impact in the campaign knowledge held by citizens.

From the empirical records examined, the following questions were determined regarding the explanation of factual and electoral political sophistication:

RQ1: At what extent does attention to politics through conventional media (a), watching electoral debates (b), use of web pages about politics (c), use of the social networks Facebook and Twitter (d), and the interactive political conversation (e) contribute to increase factual political sophistication?

RQ2: At what extent does attention to politics through conventional media (a), watching electoral debates (b), use of web pages about politics (c), use of the social networks Facebook and Twitter (d), and the interactive political conversation (e) contribute to increase electoral political sophistication?

In addition, and bearing in mind that is a comparative exploratory work among three different electoral processes (vid infra), the following research question was applied to explanatory models of factual and electoral political sophistication:

RQ3: Were these influences identical in the different studied electoral contexts?

Methodology

This research presents an explanatory methodological design under the analytical survey method, which not only does it represent broad groups of populations but aims among other objectives, to look for associations between variables and their patterns of behavior, determine their directionality, prove hypotheses and generate new hypotheses in an experimental design way but with statistical controls as well. In order to widen the range of analytic contexts and contrast them, three contexts where elections were held recently, were chosen: the Mexican states of Nuevo Leon, Puebla and Mexico, in all three such analytical survey was carried out in a convenience sample of users of social networks.

In this sense, Facebook users from the three states were invited to participate in an online survey, which was promoted as advertising in this social network. The field work was carried out immediately after the gubernatorial elections in each state: in Nuevo Leon after the elections of June 7th during the months of June and July, 2015; in Puebla after the elections of June 5th, from June to July, 2016; and in the State of Mexico after the elections of June 4th, from June to July, 2017. From the data obtained only the answers from people with legal age (≥ 18 years) and registered as voters in the three states where the research was carried out were taken into account. 51% of the sample in Nuevo Leon were women (n=152); 58% of the sample in Puebla were men (n = 176); in the State of Mexico 50.4% were women (n = 180). The participants in Nuevo Leon were between 18 and 67 years of age (M = 34.06, DE = 12.57), the participants in Puebla were between 18 and 77 years of age (M = 38.90, DE = 14.59) and the participants in the State of Mexico were between 18 and 71 years of age (M = 31.83, DE = 12.51).

Measures used

Attention to politics through conventional media. In order to know the degree of consumption of contents on politics in conventional media during the elections, the participants were asked how much they learned about the campaign through the following: news in newspapers, news in radio, news in television, political radio programs and political television programs. All of them were measured with a Likert scale ranging from nothing (1) to a lot (5). The scale was created from the sum of the five items mentioned above. The scale had a good internal consistency for all cases, the case in Nuevo Leon (α = .76), the case in Puebla (α = .77) and the case of the State of Mexico (α = .80).

Use of social networks during campaign. The survey respondents were asked how much they had used certain social networks to learn about the gubernatorial electoral campaign of each state. In short, the use of Facebook and Twitter was measured with a Likert scale ranging between nothing (1) to a lot (5).

Interactive political conversation. It was intended to know at what extent the respondents had had conversations about politics online, so it was used the scale ^{Yamamoto et al. (2015}) which uses five items, which evaluate activities such as "writing about political matters online" or "joining in political discussions online." All were measured with five-point Likert scales ranging from nothing (1) to a lot (5). The scale was reliable for Nuevo Leon (α = .89), for Puebla (α = .89) and for the State of Mexico (α = .89).

Factual political sophistication In order to generate the indicator of factual political sophistication diverse proposals by prior different authors were used as a reference, those proposals suggested summing the measures of the respondents' interest and political knowledge so as to generate an indicator of political sophistication (^{Catellani and Alberici, 2012}; ^{Guo and Moy, 1998}; ^{Muñiz, 2012}; ^{Schuck et al., 2013}; ^{Singh and Roy, 2014}; ^{Stevens, 2005}). Firstly, it was measured the interest for politics the respondents from the different states had. With the Likert scales which ranged from nothing (1) to a lot (5), respondents were asked about their interest for local or municipal, state, national or federal and international politics. The scale produced by the sum of these items presented good reliability for the three states, Nuevo Leon (α = .72), Puebla (α = .83) and the State of Mexico (α = .71).

Regarding factual political knowledge, six questions were asked concerning the present Mexican and state political reality, determining whether the respondents answered correctly (1) or incorrectly (2) to questions such as: How long is the term for governor? With both variables the indicator of political sophistication was created taking as a reference the proposal of ^{de Vreese et al. (2011}), which consists in duplicating the importance given to the cognitive element compared with the affective element by using the following formula:

Political sophistication= ((political interest x ½) + (political knowledge) / 2

The new indicator had a theoretical rank of variation which ranged between 1.75 to reflect the maximum political sophistication and 0.25 to reflect the minimum political sophistication.

Electoral political sophistication. The same proposal was used to know the degree of the respondents' political sophistication about the electoral battle. The interest shown by the respondents for the gubernatorial elections of each state was measured with a five-point Likert scale which ranged from completely uninterested (1) to completely interested (5). In addition, the degree of electoral political knowledge shown by the respondents was also measured with five questions where respondents were required to choose which candidate had proposed certain thing. The responses were codified as: correct (1) or incorrect (0), following the same criteria for factual political knowledge. The same process for factual political sophistication was followed to create an indicator.

Control variables. Finally, the use of a series of control variables was considered, some sociodemographic and other political. Respondents were asked their gender (1=masculine; 2=feminine) and their age. They were also asked to indicate with a scale which ranged between less than 6,000 pesos (1) and more than 30,001 pesos (4), what their family's monthly income was approximately. And also their education level with a scale which ranged between unschooled (1) to postgraduate studies (7). Finally, they were asked to take an ideological stance, which ranged between leftist (0) to conservative (10).

Results analysis

In order to answer the questions mentioned above the explanatory factors for both the factual and electoral political sophistication were determined, in order to do so several hierarchical multiple linear regressions were used for each state. With this technique, which allows to incorporate the independent variables to the equation in different times, it is possible to create different explanatory models and identify the variations in the effect produced by the different predictor variables on the critically analyzed variables. In both cases, the control variables were incorporated in the first stage, constituted by different sociodemographic measured aspects (model 1), and immediately afterwards the variables related to the communicative habits observed were added, such were attention to politics through conventional media, watching electoral debates, the consumption of web pages about politics, the use of social networks (Facebook and Twitter) and the interactive political conversation (model 2). In this manner a control of impact of the sociodemographic variables was created, considering the results of lack of homogeneity among the samples mentioned.

Regarding the analysis on factual political sophistication (see Table 1)², the variables used in the regression confirmed the independent hypothesis concerning the dependent variable, considering that the value of the Durbin-Watson test was acceptable in all samples. In addition, problems of collinearity were not shown among the variables used in each equation. Concerning the results, in Nuevo Leon was observed that attention to politics through conventional media (β = .123, p = .039), watching electoral debates (β = .142, p = .022), use of Twitter (β = .136, p = .026), and especially interactive political conversation during the electoral campaign (β = .246, p < .001) and the use of web pages about politics (β = .277, p < .001) contributed to increase political sophistication.

Concerning Puebla the explanatory variables were interactive political conversation (β = .175, p = .009), attention to politics through conventional media during electoral campaign (β = .212, p < .001), and mainly, use of web pages about politics (β = .267, p < .001). Finally, in the State of Mexico a positive impact was perceived in the use of Twitter (β = .123, p = .025), use of web pages about politics (β = .200, p < .001), and mainly, political conversation during campaign (β = .208, p < .001).

Now, the analysis of explanatory variables of electoral political sophistication (see table 2) confirmed the independent hypothesis concerning the dependent variable of all independent variables of all samples used, and no problems of collinearity were found either, which allowed to use all variables in the analyses made. Regarding the importance the variables about communicative habits had in explaining the electoral political sophistication in Nuevo Leon, the analysis showed four variables which had a statistically significant impact. Thus, use of Twitter (β = .176, p = .003), use of web pages about politics (β = .218, p < .001), interactive political conversation (β = .252, p < .001) and watching electoral debates (β = .292, p < .001) managed to impact consistently in the increase of political sophistication.

In Puebla, the results showed three explanatory variables which contributed to enhance the respondents' electoral sophistication: use of Facebook during the electoral campaign (β = .161, p = .015), watching debates (β = .212, p < .001) and the use of web pages (β = .299, p < .001). Lastly, watching debates (β = .178, p = .002), use of web pages (β = .222, p < .001) and interactive political conversation (β = .207, p < .001) during the electoral campaign in the State of Mexico contributed positively to increase the respondents' electoral sophistication.

Arguments and conclusion

The main objective of this research was to determine the degree in which each one of the considered communicative practices had an impact in determining the citizens' political sophistication, both factual and electoral. In sum, it was expected to evaluate the influence that factors such as attention to political contents through conventional means, watching electoral debates, use of web pages about politics, use of social networks such as Facebook and Twitter, and interactive political conversation had in the context of gubernatorial electoral campaign in the Mexican states of Nuevo Leon (2015), Puebla (2016) and Mexico (2017). Consequently, the first question was focused on the analysis of explanatory factors of factual political sophistication, which emerges from the degree of interest for politics and the knowledge about the political system citizens have.

The results let us see a decrement in the impact attention to political contents through conventional media has, which was a variable that had usually been among the most important to determine the citizens' knowledge and political sophistication (^{Hollander, 2014}; ^{Muñiz, 2012}; ^{Prior, 2005}; ^{Rhee and Cappella, 1997}; ^{Rojas, 2006}). Although in most elections, except in the State of Mexico, using these contents still contributed to increase political sophistication, it was observed that other communicative variables related to the digital sphere begin to take a highly explanatory place of this cognitive commitment of the citizen.

The use of web pages about politics such as having online conversations about politics contributed significantly to obtain more sophisticated citizens in the field of factual knowledge. Therefore, positive hypotheses proposed by other previous authors about the impact social media have in political knowledge are confirmed (^{Cho et al., 2009}; ^{Hollander, 2014}; ^{Woolley et al., 2010}).

Concerning the second research question, it is observed that three were the communicative variables which are more likely to determine the citizens' political sophistication in the electoral field. In this sense, in all the studied electoral campaigns the political conversation in digital media, use of web pages about politics and watching electoral debates seem to be the factors which help citizens increase their interest for the events during campaign and develop greater knowledge about the candidates' proposals.

Consequently, the influence of using the contents through conventional media seems to decrease again, opposite to what other authors have found in previous studies (^{de Vreese and Boomgaarden, 2006}; ^{Thorson, 2014}). However, the hypotheses previously proposed by other authors about the positive influence of debate are confirmed as long as traditional space of campaign, and of the new media (^{Drew and Weaver, 2006}; ^{Druckman and Leeper, 2012}; ^{Partheymüller and Faas, 2015}; ^{Woolley et al., 2010}).

Along with the study of explanatory elements of both factual and electoral political sophistication, in the third question was proposed whether the explanatory models to be obtained would vary according to the electoral studied context. Despite being from different states and years, the results show that there are similar patterns among the three samples in relation to the explanation of factual political sophistication.

In all three cases the main explanatory variables were those related to the use of web pages and interactive political conversation, and even the influence of attention to conventional media. For the explanatory model of the electoral political sophistication, a pattern which determines the impact of certain communicative variables is also found. The use of web pages and interactive conversation, and in this case, watching debates, all emerge as the same explanatory elements in the three states. There are only small differences among them with regard to the use of some social networks, which are considered in both types of sophistication.

The results reveal the importance Internet and social media are having in determining political attitudes, among them political sophistication. It is more than evident the role conventional media are still playing in the citizens' political learning, yet it is true that increasingly more citizens begin to use different digital platforms to inform themselves about politics (^{Drew and Weaver, 2006}), especially during electoral processes (^{Partheymüller and Faas, 2015}; ^{Woolley et al., 2010}). Therefore, this seems to be the new scenario where citizens become deep processors of political contents mentioned by authors like ^{Luskin (1990}) or ^{Prior (2005}). Thus, the results obtained allow to assume that a great part of highly sophisticated citizens are learning information through web pages about politics, subsequently they reflect upon that information through strategies of political conversation online.

Indeed, Internet has given political actors a great opportunity to spread their political discourse, considering that the new media break the barriers between producer and consumer and the discourses of the different actors are highly interrelated among themselves. Although conventional media are still a main source of information, especially television which is a much consumed medium, Internet has begun to be a mechanism highly effective to transmit political information and have an impact in the citizens’ knowledge.

Particularly in young people, who are more familiar with the use of media and social networks, and for whom a large amount of information about elections comes through Internet (^{Woolley et al., 2010}). It is for this reason that it is necessary to keep doing research about the access to politics through Internet in this sector of society, since it is in this age group that the impact social networks have may be more explicit than what was found in this research concerning population in general.

Scholars have widely argued that those who are more sophisticated tend to change less their opinions by the influence of media, and are less influenced by advertising during campaign since they have richer associative information networks than those who are less sophisticated (^{Lee and Chang, 2011}; ^{Miller, 2011}; ^{Zaller, 1992}). This sort of hypothesis may lead to cast doubt on the results obtained in the research in the sense of whether or not the consumption of politics through Internet is breaking the barriers between high and low sophisticated citizens, in a levelling process as the one proposed by ^{Hollander (2014}) or ^{Prior (2005}). In this sense, it is necessary to keep doing research about the differences shown in this new digital citizenry in order to determine at what extent media is influencing the way political knowledge is learned as well as how it is determining the degrees of political sophistication citizens have.