Introduction

Not all media coverage is the same. The term framing is used in Communication studies to refer to the object produced after the packaging of information in a way that emphasizes a particular element (^{Entman, 1993}). ^{Aalberg et al. (2012)} report that the strategic game frame -which emphasizes who is winning the election and the strategies followed by candidates- predominates in most of the countries where media coverage of electoral campaigns has been analyzed (which, coincidentally, are mostly Western countries). According to these authors, its main determinants are temporal factors, such as the public policy phase. The strategic game frame tends to appear mainly during electoral campaigns, but it can also appear in the media coverage of any political issue (^{Aalberg et al., 2012}; ^{Lawrence, 2000}); in particular, during the climax of the public policy process (for example, on the eve of a legislative vote) and when there is conflict between elites (^{Lawrence, 2000}).

However, not much is yet known about the factors that influence media organizations to prefer strategic game over issue coverage (^{Dimitrova & Kostadinova, 2013}; ^{Dunaway & Lawrence, 2015}; ^{Schmuck et al., 2016}). In addition, most of the evidence refers to media coverage in Western Europe and the United States. Researchers have advised to be cautious before universalizing results (^{Jackson, 2011}) and have raised the need to conduct electoral framing research in contexts other than Western Europe and the United States to find out if the patterns that have been found so far occur in other parts of the world (^{Dimitrova & Kostadinova, 2013}). In particular, ^{Echeverría Victoria (2018)} is skeptical about explaining electoral coverage in the Mexican context by applying models based in highly developed countries.

Hence, we ask: Are trends found in Western countries transferable to the Mexican context? Although frame-building factors have recently begun to be researched by Latin American researchers (^{Rodelo, 2019}: ^{Rodelo & Muñiz, 2016}), we have no news yet of studies that account for the factors that influence the construction of frames during electoral campaigns. Therefore, in this article we examine the role of different influencing factors of electoral coverage in the Mexican context: media sector and media ownership (organizational factors); type of election, political preference of the candidate’s party, and candidate gender (factors related to political actors); and days before election, political events and extra-political events (temporal-contextual variables).

This article adds to the existing literature on frame-building during elections in three ways: First, by testing the universality of the strategic game and issue frames in a Latin American context. In other words, by observing outcomes similar to those observed in Western European countries and the United States, we can infer that analogous media processes are at play. For example, as our data is about coverage of elections at different levels, we replicate the principle that observes that the importance of the election drives strategic game coverage. Second, by testing if media ownership type and media sector reflect in the news frames that are used in electoral coverage. In particular, this paper provides elements to ascertain the potential of public media to distinguish themselves from private media and stimulate the creation of public service contents in the media system in the Mexican context. Third, by improving our understanding of the significance of the candidate’s traits, which include the importance of the election, the degree of their party’s political preference, as well as the candidate’s gender, and contextual factors, such as political and extra-political events, for influencing the framing of electoral coverage.

The structure of the rest of this article is as follows: the theoretical frame guiding framing research is provided, with a special focus on the existing knowledge of the three sets of electoral frame antecedents featured in this study. Next, we explain the particularities of the case study, and provide the necessary details regarding the methodological procedure adopted. Finally, the findings and contributions of this study are discussed.

Media framing of elections and its antecedents

Media coverage of electoral campaigns is relevant in modern democracies mainly because: 1) voters need timely, accurately and reliably reported information on public issues, and the media organizations, even with their many limitations (^{Graber, 2003}; ^{Lippmann, 1922}), are the institutions best equipped to provide it. In addition, 2) the media function as a communicative space for public debate and dialogue -usually construed as a public sphere (^{Habermas, 1974})- that is essential for different social actors to be able to express their ideas and for these to get to be known by the public.

Theories of framing and agenda-setting have been the most successful in explaining the nature of journalistic coverage of public affairs and its effects. Framing theory posits that the form and emphasis of news messages has an impact on the formation or reinforcement of people’s mental frames and, therefore, on social debates (^{Entman, 1993}). Framing theory encompasses both the processes of building news frames (frame-building), and the processes through which these frames are set in people’s minds (frame-setting). For the study of both processes it is also essential to identify and document the frames used in the different journalistic cultures of the world.

The most important frames identified in media coverage of electoral processes have been the strategic game frame and the issue frame. News articles with a strategic game frame are those centered on describing the candidates’ viability and their strategies, such as their use of campaign resources (^{Aalberg et al., 2012}). In contrast, news articles with issue framing present “proposals for the problems, information about who is advocating which policy alternative, and consequences of the problems and proposals” (^{Rhee, 1997, p. 30}).

The predominance of the strategic game frame has been considered problematic because it takes attention away from the candidates’ proposals and platforms (^{Aalberg et al., 2012}). ^{Cappella and Jamieson (1997)} provided evidence that news articles focused entirely on strategy activate political cynicism in people, a condition characterized mainly as an “absence of trust”, based on beliefs about the actors’ motivations, and suspicions about “the lack of honesty of [political] actors and groups” (^{Cappella & Jamieson, 1997, pp. 141, 166}). Despite the latter, other researchers have pointed out that treating elections as sports competitions can make them more interesting and attractive for the electorate (^{Aalberg et al., 2012}), and, in this sense, have proposed distinguishing between the concept of a horse race frame (the treatment of elections as sports competitions, also known as a game frame) and the concept of a strategic game frame (emphasizing the motivations and strategy of political actors), arguing that the latter is the frame responsible for activating political cynicism among voters (^{De Vreese & Elenbaas, 2008}). Likewise, it is assumed that the issue frame is essential to promoting the understanding of public affairs by the public, and that it is necessary during electoral campaigns, if the electorate is to be in a position to cast an informed and reasoned vote (^{Aalberg et al., 2012}).

Predominance of the strategic game frame over the issue frame in the media coverage of electoral campaigns has been documented mainly in the United States and Western European countries (^{Aalberg et al., 2012}; ^{Dimitrova & Kostadinova, 2013}; ^{Zeh & Hopmann, 2013}), possibly associated with the prevalence of majority political systems in these regions and with the macro process of mediatization, as it is argued that presence of the strategic game frame in media coverage is one indicator of mediatization (^{Strömbäck & Dimitrova, 2011}). A longitudinal study (1994-2010) comparing Danish and German coverage of presidential elections reports a stable prevalence of about 60 per cent of horse race coverage in both countries during the period (^{Zeh & Hopmann, 2013}). Studies conducted in Mexico -with an electoral system that can be categorized as mixed because of its combination of relative majority and proportional representation posts- have also found that the strategic game frame prevails during presidential campaigns (^{Echeverría & González, 2018}; ^{Muñiz, 2015}).

To better understand the frames that emerge during the electoral season, it has been considered relevant to investigate which factors are associated with the presence of the strategic game and which are associated with the issue frame. In the categorization of factors explaining frame-building processes in news content, ^{Dimitrova and Kostadinova (2013)} have noted that factors can be classified in two levels of analysis: systemic (features of the country’s media and political system) and organizational (characteristics of the media organization). Temporal-contextual factors and factors related to political actors are two additional sets of determinants featured in frame-building studies. In the next pages, we will focus on three sets of factors: attributes of the media organization (identified in the literature as organizational factors), factors related to political actors, and temporal-contextual factors.

Context of study

Enormous political and technological changes in the Mexican context have increased the relevance of testing the association of attributes of media organizations, political actors’ traits, and context with the presence of strategic game and issue frames. Mexican regulators’ efforts at leveling the field for candidates and parties have impacted the radio and television sectors. The political-electoral reform of 2007 prohibited the purchase of political advertising during campaigns, while guaranteeing fair access by candidates to the electronic media through the equitable distribution of “State times”. Subsequently, the 2013 telecommunications reform prohibited the purchase of advertising disguised as content. Media consumption surveys show that television and radio, in that order, continue to be the most important sources of information about politics in Mexico (^{Instituto Federal de Telecomunicaciones [IFT], 2016}; ^{Secretaría de Gobernación, 2013}). Most radio and television outlets in the country are private; and the few public service broadcasters struggle with underfunding and accusations of being co-opted by the current government.

Furthermore, political reforms have allowed new actors to participate in popular elections under conditions of greater equality: the 2014 Mexican constitutional reform legalized the figure of independent candidates (that is, candidates without the support of a political party). The same constitutional reform turned the gender quota into an obligation of gender parity (fifty percent of each gender) for candidates to the federal legislature. Afterwards, a reform in the state of Jalisco (Mexico) made legislative parity possible at the local level, as well as having vertical parity in the lists of candidates to municipal positions. For the next electoral cycle (2018 local elections), gender parity was extended to the municipal candidates heading the municipal lists, a feature known as horizontal parity. However, huge gender imbalances remain in political participation, as in the gubernatorial election there were seven candidates, six of them male.

In the 2018 local electoral process in Jalisco there were three types of popularly elected positions to be voted for: head of the state government, mayors, and local legislators. The electoral process was concurrent with elections to three federal positions: president of the Mexican Republic, senators and federal legislators. The concurrence of federal elections distinguishes this electoral process from the “intermediate” elections: the perception that there is “more at stake” causes greater competition, which is a factor that influence greater horse race coverage (^{Schuck, Vliegenthart et al., 2013}).

Method

This study relies on the data produced by the monitoring of radio and television coverage of the campaigns of the 2017-2018 electoral process in the state of Jalisco (Mexico). A sample of 48 radio and television programs broadcast by thirteen different media organizations, eleven of them with private use licenses and two operating with public service licenses, was content analyzed. The 48 programs were selected using criteria of relevance (political significance, audience levels) and maximum variability.²

To carry out the analysis, only the items disseminated between March 30^th and June 27^th, 2018 (the official period for electoral campaigns) with a single actor as protagonist were considered (n = 25 663 monitored news items); that is to say, for the purposes of this study, units that mentioned candidates in different elections, from different parties or with different genders were discarded.

Results

The study identified 7 484 news items with a strategic game frame (29.2% of the sampled news items), while the number of news items with an issue frame was 11 163 (43.5%). Table 1 shows the results of the multiple logistic regression applied to predict the use of the strategic game frame in electoral coverage, while the results of the multiple logistic regression applied to predict the use of the issue frame in electoral coverage are shown in Table 2. The significant coefficients reported in both tables reveal that all of the independent variables of interest are significantly related (in a positive or negative way) to the presence of the strategic game and issue frames, with the exception of “day after gubernatorial debate” (a temporal-contextual variable) and “male candidate” (a candidate trait variable), both of which do not predict the use of the strategic game frame.

As previously implied, the variables related to features of the media organization predicted the use of both strategic game and issue frames. Answering the cross-media question (RQ1), Tables 1 and 2 reveal that radio programs are more likely to use the strategic game frame, while television programs are more likely to use the issue frame. On the other hand, results support the public media hypothesis (H1), as we found that, compared with private media, public media are less likely to use the strategic game frame, and are more likely to use the issue frame.

The variables related to candidate traits predicted both strategic game and issue frames, except for “male candidate”, which only predicted the issue frame. We will examine the three variables separately: First, being a gubernatorial candidate was a positive predictor for both the strategic game and issue frame, a result that supports the importance of election hypothesis (H2). Second, being a candidate from a party with greater electoral preference was a positive predictor of the strategic game frame, but a negative predictor of the issue frame; this result supports the viability hypothesis (H3) and adds a twist to it, as in this case viability seemed to penalize substantive reporting. Third, and answering the gender question (RQ2), the findings reveal a gender bias that is detrimental to female candidates, as being a male candidate did not predict the use of strategic game frame, but it positively predicted the issue frame.

The temporal-contextual variables predicted both strategic game and issue frames, except for “day after gubernatorial debate”, that predicted only the issue frame. The complete results are as follows: First, we found, as expected, that the number of days until election day negatively predicted both the strategic game and issue frame (or, in other words, news items closest to the election day were more likely to present both frames). This result supports the election day hypothesis (H4). Second, and answering the political events question (RQ3), the findings show that being published during the day of start/end of campaigns and the day after that was a positive predictor of the strategic game frame but a negative predictor of the issue frame. Third, and answering the extra-political events question (RQ4), the results show that being published the day after an important extra-political event of a sporting nature was a positive predictor of the strategic game frame and a negative predictor of the issue frame. Finally, it was surprising that being published after the mandatory gubernatorial debate was only a negative predictor of the issue frame, as it did not predict the use of the strategic game frame.

Following the latter results, we distinguish several variables that show a similar pattern, positively predicting use of the strategic game frame while negatively predicting the issue frame (hence sort of “penalizing” substantive reporting): 1) candidate from party with greater political preference; 2) day of start/end of campaigns; and 3) day after an important extra-political event; 4) Radio programs; and 5) private media also acted in a manner congruent with this pattern.

On the other hand, two variables show a pattern promoting both the strategic game and issue frames: “days immediately before the election” and “gubernatorial candidate”. In this manner, results indicate that timeliness and relevance were forces that promoted both frames in these elections.

Conclusions

Our findings indicate predominance of the issue frame over the strategic game frame in the coverage of local elections in the Mexican context. Some possible explanations of this outcome are referred to in the literature on political communication: Several researchers have opined that the dynamics influencing strategic game coverage are mostly related to actors and campaigns at the national level (^{Aalberg et al., 2012}; ^{Schuck, Boomgaarden et al., 2013}) -that is, where there is more at stake. It has been affirmed that the use of mass media as information channels is more critical at the national level (^{Dimitrova & Strömbäck, 2012}), which increases the probability of finding actors trying to influence public opinion through the use of strategic game coverage. Additionally, an important trigger of this style of reporting -public opinion surveys- is more frequent at the national level. Other possible explanation derive from a contextual feature of the case study: competitiveness predicts use of the strategic game frame (^{Dunaway & Lawrence, 2015}); however, most newsworthy local elections in our case study were not competitive, as the leading candidates in the largest elections won by a large margin.⁴

The findings of this study account for factors that significantly predict the use of strategic game and issue frames in local electoral coverage in radio and television programs in the Mexican context. For some factors, we observed outcomes congruent with those observed in Western European countries and the U. S. in a way that supports the assumption that analogous media processes are at play. Regarding organizational factors, this article adds to the evidence establishing the potential for cross-media comparisons to find differences in the use of frames (^{Jackson, 2011}; ^{Kim et al., 2010}; ^{Schmuck et al., 2016}). Most of the evidence establishes differences between print and television coverage (^{Jackson, 2011}; ^{Kim et al., 2010}), or between print and digital media coverage (^{Schmuck et al., 2016}). This is the first study ever to compare between radio and television contents: we found that, compared to radio, the news items produced by television were more likely to have an issue frame and less likely to have a strategic game frame. Differences detected by this study may be explained by tighter control in television than radio (^{Shoemaker & Reese, 2014}). In contrast with television, radio tends to be more spontaneous and open to audience participation. It has been noted that the study of radio coverage has been neglected (^{Rafter et al., 2014}). Radio’s higher preference towards the strategic game frame in comparison with television is another call for more research into radio contents.

This study found that the news items produced by public media were more likely to have an issue frame and less likely to have a strategic game frame. Private media’s preference for strategic game framing can be understood as part of the market orientation and commercialism of private media (^{Dimitrova & Strömbäck, 2012}; ^{Schmuck et al., 2016}; ^{Rafter et al., 2014}), whereas public media’s preference for issue framing can be interpreted as the fulfillment of their public service goal (^{Dekavalla, 2016}; ^{Rafter et al., 2014}). This result confirms the potential of public media in a Latin American context to realize a coverage that significantly differs from that made by private media and that provides more and better opportunities to acquire political knowledge (^{Strömbäck, 2016}). However, the context of this case study must be considered before transferring this result to other settings, as Jalisco is home to two relatively strong public service broadcasting systems.

This study confirms that in the news production process candidate traits can influence the selection of news frames. News items about gubernatorial candidates were more likely to have both issue and strategic game frames, whereas news items about candidates from parties with the greatest political preference were more likely to have a strategic game frame and less likely to have an issue frame. Newsworthiness and viability, respectively, may be the drivers that explain both outcomes. On the other hand, this study contributes to the literature on gender biases in electoral coverage by identifying gender differences in frame selection that are detrimental to female candidates, as news items about male candidates were more likely to have an issue frame, while gender was not significant for predicting the strategic game frame. The literature points to two main categories of factors influencing gender biases in electoral reporting: the ones deriving from media organizations (e. g., gender stereotypes among media staff) and the ones deriving from political institutions and that are mirrored by media contents (^{Lühiste & Banducci, 2016}). We can count among the latter gender differences in campaign resources and capacities; for example, gender differences in the placement of candidates. Most of the gubernatorial candidates -a factor positively predicting issue coverage- in our case study were male. This leads us to infer that there is at least one factor deriving from political institutions (candidate placement) that contributes to gender bias in the selection of news frames. Hence, gender bias research that controls for factors associated with relevance and perceived viability of candidates must be conducted to understand if those factors are the drivers of gender biases in the selection of news frames.

Finally, our results confirm the enormous influence of temporal-contextual factors in predicting the selection of news frames. In accordance with previous studies (^{Dekavalla, 2016}; ^{Dunaway et al., 2015}; ^{Schuck, Vliegenthart et al., 2013}; ^{Vliegenthart et al., 2010}), the significance of the proximity of election day for selection of the strategic game frame was supported. This same factor also positively predicted the use of the issue frame. This means that coverage lacking both frames (e.g., mini-stories mentioning candidates but not exploring issues or political strategy) is avoided towards the end of campaigns. In a manner similar to what is seen in the United States’ party conventions (^{Vliegenthart et al., 2010}), this study also found that events of opening and closure of campaigns attract a strategic game coverage. Surprisingly, publication on the day after the gubernatorial debates negatively predicted news items with issue frames and did not predict a strategic game frame. Hence, more research is needed to better understand the role of electoral debates on the media framing of campaigns.

Our findings suggest extra-political events of a competitive kind -e.g., popular sports- may influence the likelihood of a strategic game frame and reduce the likelihood of issue coverage. This may be a consequence of the extra-political event filling the media and the social milieu with ideas and messages related to competition: winning and losing, scoring, strategy, performance, etc. -precisely the same notions that strategic game coverage transfers to electoral settings (^{Aalberg et al., 2012}). In this way, we hypothesize, an important soccer match exacerbates the already popular use of the game metaphor to understand and report the political campaigns. Another explanation is suggested by ^{Schmuck et al. (2016)}: strategic game framing is exacerbated when stories have to compete for attention with sports and entertainment. Hence, important sports events drive journalists who work on electoral topics to compete for attention by choosing a strategic game frame.

The present study was made from a sample of radio and television news items about candidates competing in a local process that involved elections for different positions. Due to the way it was designed, the reader must take into account the possibility that the results reflect the idiosyncrasies of the media organizations and political system of Jalisco, as well as characteristics and processes particularly related to elections of officials at the subnational level.