Measuring Negative Campaigning on TV, radio, debates, press and Facebook: The Case of 2014 Brazilian Presidential Elections

Abstract

This article analyzes the discursive strategies of the candidates Dilma Rousseff, Aécio Neves and Marina Silva during the presidential election of 2014. The objective is to measure the tone of the presidential campaign based on television ads, radio ads, election debates, printed and aired candidate declarations and party messages placed on Facebook, in the first and second rounds. A review of the literature shows that this approach is unprecedented in studies of rhetoric in campaigns. In general, studies measure the incidence of negative television advertising, neglecting other channels of communication. The analysis shows that the intensity of negative advertising varies in these multiple channels according to the degree of control of the candidates on their messages, the level of interaction between the candidates and the chance of the candidate being identified as the perpetrator of the attacks.

Keywords
Public opinion; Presidential elections; Electoral campaign; Political Communication; Negative Campaign

Resumo

Este artigo analisa as estratégias discursivas dos candidatos Dilma Rousseff, Aécio Neves e Marina Silva durante a eleição presidencial de 2014. O seu objetivo é medir a variação da propaganda negativa veiculada no Horário Gratuito de Propaganda Eleitoral do rádio e da televisão, nos debates eleitorais, nas declarações dos candidatos publicadas na imprensa escrita e televisiva e nas postagens do Facebook, no primeiro e no segundo turnos. Uma revisão da literatura mostra que essa abordagem é inédita nos estudos sobre retórica em campanhas. Em geral, os estudos medem a incidência da propaganda negativa com foco na televisão, negligenciando outros canais de comunicação. A análise mostra que a intensidade da propaganda negativa varia nesses múltiplos canais de acordo com o grau de controle dos candidatos sobre as suas mensagens, o nível de interação entre os candidatos e a chance de o candidato ser identificado como o autor dos ataques.

Palavras-chave
Opinião Pública; Eleição presidencial; Campanha eleitoral; Comunicação Política; Propaganda Negativa

Resumen

Este artículo analiza las estrategias discursivas de los candidatos Dilma Rousseff, Aécio Neves y Marina Silva durante la elección presidencial de 2014. Su objetivo es medir la variación de la propaganda negativa transmitida en el “Horário Gratuito de Propaganda Electoral” de la radio y la televisión, en los debates electorales, en las declaraciones de los candidatos publicadas en la prensa escrita y televisiva y en las mensajes de Facebook, en el primer y segundo vueltas. Una revisión de la literatura muestra que este enfoque es inédito en los estudios sobre retórica en campañas. En general, los estudios miden la incidencia de la propaganda negativa con foco en la televisión, descuidando otros canales de comunicación. El análisis muestra que la intensidad de la propaganda negativa varía en estos múltiples canales de acuerdo con el grado de control de los candidatos sobre sus mensajes, el nivel de interacción entre lós candidatos y la posibilidad de que el candidato sea identificado como el autor de los ataques.

Palabras clave
Opinión Pública; Elección presidencial; Campaña electoral; Comunicación política; Propaganda Negativa

Introduction

In electoral campaigns, candidates use multiple channels of communication when it comes to criticizing opponents¹. Prior to the electoral period, when TV and radio ads are prohibited, candidates publish documents, make public appearances, and quite often attack their opponents through the press. Take, for instance, presidential candidate Aécio Neves´s claim that the propaganda techniques used by then president Dilma Rousseff echoes those used by Nazi Germany in its use of lies in a “criminal attempt to deconstruct adversaries²”.

Although the mechanisms of attack and defense identified above are common in electoral disputes, an overview of the literature shows that studies primarily measure the use of negative campaigning in just a single channel. The US academic literature, the pioneer on the subject, focuses primarily on television spots (KAID; JOHNSTON, 2001; BENOIT, 1999; GEER, 2006; WEST, 2009). Outside the US context, communication channels are more plural. In Europe, it is common to find analyses of negative campaign focusing on debates, press releases, manifestos, and others (NAI; WALTER 2015)³. Comparisons of negative ads in different channels are a rare exception. These studies indicate that the amount of negative publicity varies among different channels according to the degree of control that candidates have over their messages and the level of direct confrontation (ELMELUND- PRÆSTEKÆR, 2010; WALTER; VLIEGENTHART, 2010; WALTER, 2012).

In Brazil, studies on negatives ads have gained traction in the last years. Generally, this is a strategy seldom used by presidential candidates (BORBA, 2015a). The decision to attack tends to be made by those who are trailing in polls and opposition candidates, and is more common in reelections years and second round votes (CARVALHO, 1994; LOURENÇO, 2009; BORBA, 2015a; BORBA, 2015b). These studies, however exclusively investigate television ads in Brazil, in fact, the predominant mode of communication between candidates and voters, given the rules that grant political parties unpaid access to television and also the high degree of television viewership in Brazilian society. Comparative studies generally contrast the rhetoric strategies employed in television and radio, indicating that negative ads tend to be more recurrent in radio, in addition to the fact that this mode of communication is more prone to approach sensitive and personal topics, unlike television, such as the criticism made by José Serra against Dilma Rousseff about her stance on abortion and gay marriage and her historical involvement with groups that resisted against the military dictatorship in the 1970s (BORBA, 2013; PANKE; WIGGERS; OKIDO, 2013).

The goal of this article is to contribute to the national and international literature by measuring the patterns of negative ads in different channels during the 2014 presidential elections in Brazil. The article simultaneously considers attacks made in television and radio spots, debates, Facebook posts and statements quoted by printed newspapers and statements broadcast by television networks in the first and second round. The study includes the coding of all discourse transmitted through these different channels of communication, however limiting the range of the analysis to the three main candidates: the (then) president and reelected candidate Dilma Rousseff (PT) and opposition candidates Aécio Neves (PSDB) and Marina Silva (PSB), which together collected 96.4% of the votes. The analysis shows that there are significant differences in the amount of attacks per channel, but the ranking is the same: that candidate which attacks over a channel attacks invariably more on the others.

The 2014 Brazilian electoral campaign turned out to be a very interesting one to study negative ads. The dispute was characterized by high volatility of voting intentions due to the amount of attacks launched by candidates. President Dilma Rousseff sought reelection in a context of low popularity and was threated by the considerable growth of Marina Silva in the first round and Aécio Neves in the second, interrupted by direct actions by the Dilma campaign. The attacks served to discredit Marina Silva as an authentic representative of change, by successfully labeling her as a shaky, volatile candidate, often changing her opinions and without a solid parliamentary foundation to assure governability. Against Aécio Neves, the negative ads launched by Dilma Rousseff was effective in tarnishing his personal image and reputation as an efficient administrator (BORBA; VEIGA; MARTINS, 2015).

This article is organized in the following manner: it begins with a review of the literature on negative ads found in the national and international literature. In this section, we discuss certain factors that influence the occurrence of attacks in different media channels, such as the level of interaction between candidates and degree of control over the massages. In the following section, two other variables are suggested, the anonymity of the author of the messages and the level of regulation in Brazilian campaigns, focusing on the channels of communication included in this study. In the following section we present the methodology and the content analysis of the campaign of the candidate up for reelection, Dilma Rousseff (PT) and the opposition Aécio Neves (PSDB) and Marina Silva (PSB). In the last section we discuss the main findings.

Negative ads: review of the literature and expectations

Negative ads became an object of systematic study in the beginning of the 1980s in the United States. Today, it is possible to state that the studies on negative ads have evolved based on three distinct perspectives. The main one sought to understand its impact on the democratic process. The starting point of this line of research is the hypothesis of demobilization, which assumes that negative ads are harmful to the political system, as it correlates negatively with electoral turnout. The hypothesis of demobilization appeared in controlled experiments conducted by Ansolabehere and Iyengar et al (1994, 1995), but their findings were later put into question (MARTIN, 2004; GEER, 2006; FINKLE; GEER, 1998).

The other lines of research display a closer relationship with the electoral dynamic. One of them aims to measure the impact of attacks on voting intention rates. The findings once again proved controversial. Studies concluded that negative ads are efficient in taking away votes from the opponent, but it also has a collateral effect on the attacker. The socalled “boomerang effect” of negative ads has been identified in several studies, all of which show that the chance of it happening is greater when the attack is launched by the candidate against the opponent personally without the presentation of “evidence” supporting the content of the message (GARRAMONE, 1984, 1985; JOHNSON-CARTEE; COPELAND, 1989; RODDY; GARRAMONE, 1988; KING; MCCONNEL, 2003; PAINTER, 2013; DOWLING; WICHOWSKY, 2015).

Lastly, there is the line of research that seeks to understand the determinants of negative ads. In this set of studies, the main concern is identifying who attacks, when and with how much intensity. The studies carried out in the US demonstrate the increasing use of negative ads in presidential elections and estimate that currently approximately half of all ads were intended as attacks against the adversary (BENOIT, 1999; KAID; JOHNSTON, 2002; GEER, 2006; WEST, 2009). The decision to attack is predominantly used by opposition candidates and those trailing in polls and who resort to attacks to reverse the situation (SKAPERDAS; GROFMAN, 1995; THEILMANN; WILHITE, 1998). In Brazil, negative ads are not on an upward trajectory. They tend to show up more when candidates are up for reelection and in run-offs (CARVALHO, 1994; LOURENÇO, 2009; BORBA, 2015a, 2015b).

Studies seeking to measure the amount of attacks in different communication vehicles are relatively recent and few in number. This gap is relevant since voters decide based on a plural market of political information (FIGUEIREDO, 2000), composed by the official ads promoted by candidates on radio and television, debates promoted by television and radio broadcasters, press coverage, and, more recently, digital media such as Facebook and Twitter. Thus, to ignore information circulated in these channels can compromise a full understanding of how candidates interact with voters and thus how public opinion is formed in electoral times.

The international literature suggests that the degree of negativity in different media vehicles varies due to two basic reasons. The first one is the control of candidacies over their messages. The level of control is defined by the capacity of candidates in transmitting their messages as originally intended (WALTER; VLIEGENTHART, 2010). This is the case for example of the Horário Gratuito de Propaganda Eleitoral (HPGE)– the obligatory slots assigned for electoral ads -, in which candidates are able to pass on their messages unfiltered. Parties and candidates, however, can only partially control what is published by the press. Journalists are gatekeepers who act as filters for what the candidates declare, selecting news that suits their interests, which may or not coincide with those of the candidate. This concern must be greater in contexts in which the media can be characterized as “polarized pluralism”, to use the term adopted by Hallin and Mancini (2004), such as the Brazilian case, in which a small diversity of opinions prevail, and where there is a high level of interest in negative ads and the media displays a conservative bias (AZEVEDO, 2006).

Negative ads also vary according to the level of interaction between candidates. This is the particular case of electoral debates. The debates are in fact the only moment of the campaign in which there is direct and unmediated interaction between candidates. In debate, candidates must stand face to face, enjoying equal amount of video time. Given the confrontational character of debates, analyses have therefore been showing that the tone of campaigns tends to be relatively more negative in debates than in other communication channels. This conclusion can noticed in a comparison of different studies (NAI; WALTER, 2015), such as in those comparing the intensity of negative campaigns in different channels (ELMELUND-PRAESTEKAER, 2010; WALTER, 2012).

In light of the above, it is possible to affirm that the literature has established that the interaction among candidates and the level of control over their messages as the two main variables capable of explaining varying levels of negative ads among different channels. In the next section, I will discuss the regulation of campaigns with regard to content control and the so-called boomerang effect as two additional variable capable of influencing the strategies of candidates when attacking opponents.

The Regulation of the Brazilian Media System and its Unexpected Effects

In Brazil, Electoral Law 9,504 of 1997 regulates electoral campaigns. It establishes, among other provisions, the rules that establish how political parties and candidates access radio and television, the norms of use of the internet and even how the press covers the campaigning. In this section, we present the regulation regarding the channels investigated in this study and next the occurrence of unexpected effects, such as the socalled boomerang effect⁴.

According to Electoral Law 9,504/97, political advertising is exhibited in radio and television broadcasts in specially designated time slots – the Horário Gratuito de Propaganda Eleitoral (HPGE). The HPGE is a slot guaranteed by law intended for political parties to inform the electoral about their proposals, free of cost. The HPGE was established by Law 4,737 (July 1965), laying out the Brazilian Electoral Code during the military government and creating time slots guaranteeing parties free access to electronic media. It is divided into two distinct formats: continuous block-format slots and short 30-second spots shown during the regular programming of networks. Spots are recent additions and their broadcast began in the 1998 presidential campaign, with promulgation of Law 9,504/1997. Each model has its advantages and disadvantages, but political analysts have as of late argued against the traditional mode, considered long and expensive, and have positioned themselves in favor of short spots, which can catch viewers off guard (LAVAREDA, 2009).

One of the central problems faced by the Brazilian electoral ad model is how to control content. Unlike the US, Brazilian electoral legislation prohibits ads that may “degrade or ridicule a party, candidate or coalition”, with two possible consequences for parties who disrespect this norm: the ads can be removed from broadcasts or the concession of time for the attacked candidate provide a response. The right to a response provides the attacked candidate the right to use part of the time allotted to the opponent who launched the attack, being that the response´s time will equal the time used to make the attack and cannot be less than one-minute long. Despite not being usual (STEIBEL, 2007), the right to respond mobilizes legal teams and affects the decision to attack as it is accompanied by the risk of losing ad time to the opponent.

Regulation of digital campaigning through the Internet is relatively recent. It has been established by Law 12,034/2009, which has been incorporated into the Electoral Code. Ads or messages can be placed in the websites of candidates, parties or coalitions and must be informed to the Electoral Justice system and must be hosted in servers based in Brazil. As for ads exhibited in television and radio, the TSE ensures the right to respond on the Internet. The legislation established that the response to an offence can by displayed on the same web page used for the attacked, being shown at the same time, size and forma as the attack. The response must remain available for web users during no less than twice that the message considered offensive remained available. Although the internet has been used in electoral campaigns since 2002 (ALDÉ; BORGES, 2006), Facebook is a relatively new tool. This social media was not integrated into campaigns until 2012 and the national academic output so far is scarce therefore. Findings show that the strategy of mobilization, in tandem with agenda promotion, is predominant on Facebook (MASSUCHIN; TAVARES 2015). There are no records of response being conceded, according to this literature.

The Electoral legislation also regulates the organization of debates. In 2014, television stations were required to invite all candidates whose parties had at least one representative in the Chamber of Deputies, but were free to lay down their own rules of questions, answers and the right to answers. On the right of reply, broadcasters typically maintained a team of technicians willing to analyze the campaign response request with a moderate impact on the voter because it occurs at low audience hours (late at night on a weekday) and reaches a spectator with specific profile: voters with already defined votes, high schooling, high purchasing power and strong interest in politics. Its effect on the electorate depends to a large extent on the repercussion in the news (which points to winners and losers) and also the repercussion on the electoral propaganda itself. It is common for candidates to select best-performing segments and advertise in their advertisements in the days following the debate (VASCONCELLOS, 2013). The analysis of the content of the debates, which took place between 1989 and 2010, confirms the high degree of confrontation. They are campaign events with a high incidence of attacks and are more negative in the seconds than in the first shifts, such as television advertising (LOURENÇO, 2007; VASCONCELLOS, 2013).

Lastly, the electoral law also regulated how radio and television networks cover the elections. This is because in Brazil television and radio broadcasting rights are a public concession. In other words, in order for a radio or television network to operate, it requires public authorization, establishing the duration of the concession, which must be renewed in order to remain active. The legislation defines the balanced news coverage of candidates. In fact, in Globo’s (the main television network in Brazil and the object of this investigation) coverage in the last election, the three main candidates had exactly 20 seconds of coverage daily, with the order of their appearance being alternated: the last to appear on one day was the first to appear the following day and so on. However, although the legislation secures equal treatment, electoral coverage has been far from fair and balanced. It is well known that the selection of news and framing have historically been biased against candidacies of the PT (MIGUEL, 1999; MIGUEL, 2003; PORTO, 2012). Regarding printed news, as this media is not a public concession, there is no specific regulation concerning the criteria of news coverage. Here, yet again, the literature has identified a strong bias against the PT in national press (ALDÉ, 2003; ALDÉ; MENDES; FIGUEIREDO, 2007; FERES JR; SASSARA; BARBABELA, 2014).

The boomerang effect, another variable neglected by international studies, is the fourth factor that influences the level of negativity detected in different channels. The boomerang effect occurs when the attack does not impact the target, but rather the sender of the message. The probability of the boomerang effect happening is greater when the voter identifies the author of the message and perceives it as out of bounds or fraudulent. In the US, candidates avoid this in two ways. One is not showing the candidate making the attack – the task is delegated to a surrogate, either a hired person or a party leader. Another common way is allowing an association to carry out the attack, sine in the US electoral ads can be freely disseminated by anyone willing to purchase time on television. The advantage is the lack of clarity, from the voter´s perspective, regarding who is doing the dirty work, making it hard to attribute blame and therefore avoiding the boomerang effect (DOWLING; WICHOWSY, 2015).

In the Brazilian case, some circumstances affect the identification of the authorship of attacks. The electoral legislation stipulates that candidates must include subtitles in ads indicating all parties that are part of the coalition. However, this obligation is relativized in two manners. First, the electoral legislation does not require the identification of parties that are part of the coalition in 15-second spots on the radio. This legislation came into in effect during the 2006 presidential election, based on a ruling by the Supreme Electoral Court (TSE) issued August of the same year. Second, although the law requires the identification of advertising in a “clear” and “legible” fashion, it does not specify what these terms mean. As a result, the task of camouflaging the author of attacks is facilitated with the use of very fine printing the subtitles, often in white fonts against a light-colored backdrop⁵, artifices which compromise legibility and the comprehension of subtitles (BORBA, 2012)⁶.

In light of the above, it is thus possible to state that negative ads tend to occur more in environments that allow direct confrontation, with little regulation, and when candidates do not control their messages and are able to elude self-identification. In the next section the methodology is presented.

Methodology

The study involved the collection of data relative to the candidates considered relevant in the 2014 presidential dispute: Dilma Rousseff, Aécio Neves and Marina Silva, who together collected 96.5% of the votes from a pool of 11 candidacies. These candidates participated in all debates, their Facebook pages counted an elevated number of visits, and were widely covered by the written and televised daily press, in addition to receiving the largest amount of television and radio spots.

This study included all television and radio spots broadcasted by the three candidates, all Facebook posts, the eight debates organized by television networks and all the statements quoted by the printed newspaper “O Globo” and the televised news show “Jornal Nacional,” both a property of Globo Organizations, the biggest media conglomerate in Brazil. The “O Globo” newspaper, although based in Rio de Janeiro, is one of the only Brazilian newspapers with national circulation, while Jornal Nacional has been, since the 1960s, the ratings recoldholder in Brazil (PORTO, 2012). The collection of data occurred between August 19 and October 2 (first round) and October 9 and 24 (second round) – coinciding with the official campaign period on television and radio.

Chart 1, following, sums up the units of analysis and the total 4,098 incidences of observation:

Chart 1
Units of Analysis of Communication Channel

Channel of Communication	Unit of Analysis	Observations
TV spots	Text components (images were ignored)	1033
Radio spots	Text	1297
Debates	The candidates words in each intervention (question, answer and following back and forth)	450
O Globo	The candidates words quoted by the newspaper	626
Jornal Nacional	The candidate’s words quoted in the news	133
Facebook	Post text. Shared content not considered	543

• Source: the author

The content analysis unfolded in two stages. In the first one, each message was coded by a team of three research assistants into three categories according to Benoit’s (1999) functional theory: acclaims, attacks and defenses. “Acclaim” is the type of message that can also be called “positive advertisement” and is used by candidates to present their government platform, personal qualities and any other type of information that can help drive one´s campaign. “Attack” or “negative advertisement” are the terms used to describe messages that are intended to highlight a negative trait of an opponent. In this study we based our classification on Geer’s (2006, p.23 – Our translation) definition: “a negative ad is any criticism directed at a candidate by an adversary”. This definition therefore includes any reference to an opponent´s personality, their political convictions, their track record as a politician and public administrators, their political party, associates, family members, friends and support groups. The third category, “Defense” is a response/explanation given to an attack one receives. In the second one, the messages were classified according to their function as political, personal or a combination of both. Messages of political character involve the presentation of proposal, whereas personal one aim to highlight the personal qualities of candidates, such as administrative experience, preparedness and competence. Lastly, “combination” refers to the strategy of mixing in the same message, personal and political content.

Results

The intensity of negative propaganda varies according to the channel of communication used by the candidates. Considering all channels simultaneously, the attacks corresponded to 36.5% of all messages in the 2014 presidential elections. The debates, as expected due to degree of confrontational level, appear as the site of greater occurrence of attacks. Approximately 55% of the statements by candidates were intended as attacks against opponents. The statements highlighted by the O Globo newspaper appear next, with 43.1%, thus confirming the expectation that low regulation and the lack of control in addition to the press´s proclivity to prefer negative messages increased the occurrence of attacks. However, this expectation was not observed in the Jornal Nacional. In this media vehicle, the attacks between candidates was the lowest among all channels (6.8%). This finding is probably the result of more restrictive regulation towards public concessions, something that does not apply to written press. Televised and radio spots appear as the third and fourth channels with more negative ads, with 41.7% and 35.6%, respectively. However, this result is the opposite of what was observed in Borba (2013), regarding the 2010 presidential elections, when the radio spots were more heated than those exhibited on television. Finally, posts on official Facebook pages added up to only 13.3% of the attacks, confirming the low degree of negativity in this vehicle (MASSUCHIN; TAVARES, 2015; SOUSA; MARQUES, 2017).

Table 1
Strategies of Political Communications in different channels (%)

	TV Spot	Radio Spot	O Globo	Facebook	Debate	Jornal Nacional	All Channels
Acclaim	51.5	64.4	32.7	82.5	36.0	88.0	56.3
Attack	41.7	35.6	43.1	13.3	54.9	6.8	36.5
Defence	6.8	0.00	24.1	4.2	9.1	5.3	7.2
N	1033	1297	626	543	450	133	4098

• Source: the author.

The division of results by electoral round reaffirms the findings of previous studies (BORBA 2015a, 2015b), concluding that the second round is more negative than the first. Exceptions are the quotes of candidates selected and transmitted in Jornal Nacional where the percentage of attacks decreases, from 7.5% to 3.7%. The growth is expressive in radio spots (from 26.1% to 45.4%), on Facebook (8.4% to 26.2%) and, most of all, in debates (43% to 66.4%), which once again confirms their confrontational aspect due to greater interaction among candidates. In television spots, the percentage of negative ads soared from 35.9% to 46.3%, while in statements published in O Globo they climbed from 41.1% to 49.1% (Table 2).

Table 2
Frequency of Attack by Communications Channels (%)

	TV Spot	Radio Spot	O Globo	Facebook	Debate	Jornal Nacional	All Channels
1 Round	35.9	26.1	41.1	8.4	43.0	7.5	28.8
2 Round	46.3	45.4	49.1	26.2	66.4	3.7	46.5
Both Rounds	41.7	35.6	43.1	13.3	54.9	6.8	36.5

• Source: the author.

In Brazil, the second rounds are more negative as a result of a set of factors. The Brazilian political multi-party system in which the choice for president undergoes two electoral rounds certainly influences the volume of attacks given the cost-benefit dilemma faced by parties, which is distinct from that in bipartisan systems with simple majorities, such as in the US system. In the dispute with more than two competitors, the benefits of negative propaganda are dispersed in the sense that they can be divided among parties and not benefit the one who launched the attack – while the cost of attacking is borne by the attacker alone. In other words, in a plural dispute, the change of accusations among two candidates can hypothetically benefit a third candidate, a restraint that does not apply to bipartisan systems, in which disputes are a zero-sum game in which the percentage lost by one candidate automatically benefits the opponent. At the same time, the existence of a second round influences the decision to attack due since the support of an attacked opponent might be needed in the second round. The consolidation of this agreement will be hindered by the intensity of attacks in the first round, thereby increasing the costs of bargaining. In the second, these constraints cease to exist (BORBA, 2015a).

Individual Variations in Campaign Strategies

There are considerable variations in the strategy of candidates (Table 3). In the first round, the poised for reelection president, Dilma Rousseff, attacked less than her opponents in the first round (18.1%), followed by Marina Silva (22.7%) and Aécio Neves (44.6%). Dilma targeted mainly Marina, and almost ignored Aécio: 87.0% ofher attacks targeted the PSB candidate. The attacks against Marina, considering all channels of communication, began in early September, in line with electoral growth of the adversary in the polls. In the second round, the president attacks more than her opponents, possibly due to the fact she begins this round trailing in polls, an unprecedented fact in Brazilian presidential disputes – the winner of the first round has always led polls from the beginning to the end of the second round (BORBA; VEIGA; MARTINS, 2015). The peak occurs in the first week of the campaign and the intensity of attacks decreases after Dilma reassumes the lead.

Table 3
Frequency of Attacks by Candidates and Communication Channels

	1 Round			2 Round
	Dilma	Aécio	Marina	Dilma	Aécio
TV spot	22.4	58.9	34.7	49.5	43.5
Radio spot	10.6	42.3	17.9	54.4	35.9
O Globo	29.2	56.9	37.4	45.9	52.7
Facebook	7.5	12.4	6.6	31.3	17.0
Debate	30.7	63.0	35.6	58.3	74.6
Jornal Nacional	0.0	10.3	11.8	0.0	7.1
All channels	18.1	44.6	22.7	49.4	43.4

• Source: the author.

The percentage of attacks by candidates varied according to the network. In the first round, the attacks launched by Dilma Rousseff were observed mainly during electoral debates (30.7%), in statements highlighted by the O Globo newspaper (29.2%) and in televised spots (22.4%). In his turn, Aécio Neves used 63% of his interventions in debates to attack adversaries. His percentages are also above the average (44.6%) considering televised spots (58.9%) and statements relayed by O Globo (56.9%). The attacks by Marina follow stable percentages from televised spots (34.7%) to statements in O Globo (37.4%) and debates (35.6%). In the run-off, an inversion can be noted in the pattern compared to the first round, although the debates remain as the foremost locus of negative ads. Dilma begins using the radio as one of the main weapons of attack: 54.4% of spots in this vehicle served to disqualify, one way or another, the PSDB candidate. There is also a significant increase of negative post on Dilma´s official Facebook profile, increasing from 7.5% to 31.3%. Aécio, on his turn, considerably decreased his attacks on radio and television, yet sustained them in debates. It is interesting to observe that none of Dilma´s statements shown on Jornal Nacional were aimed at disqualifying opponents, either during the first round or the run-off.

Different Kinds of Attacks

Candidates strategically chose to attack from where to attack opponents, but also dose the kind of attack depending on the channel being used. According to Benoit`s (1999) functional theory, the attacks may assume a political character, a personal one or be a blend of both. Personal attacks highlight, for example, the individual shortcomings of candidates, their lack of preparedness, experience or dishonesty. Political attacks involve first and foremost criticism toward proposed policies, bringing attention to the fact some might be technically unfeasible or impossible considering budget restrictions, while so-called “combined” attacks seek to balance these two kinds of criticism in a single message. Every campaign will be invariably the result of a balance between these strategies, but the aim of this section is to figure out whether there are variations depending on the type attack and chosen media vehicle.

In general terms, the candidates discuss more political issues (69.7%) than personal ones (14.9%), or the combination of both (15.5%). However, it was possible to observe significant variation when the aim of the message is considered according to vehicle (Table 4). It is possible to observe an increase in messages that discuss the image of candidates on radio (34.2%). The news channels – O Globo and Jornal Nacional – privilege coverage that is mainly political, highlighting pieces in which candidates talk about their proposals. It is also possible to notice that Facebook and electoral debates also appear as privileged spaces for the discussion of proposals. In the case of debates, one could hardly expect a different result since, in most occasions, the issues are chosen by the organizers.

Table 4
Topic of messages by communication channels (%)

	Spot TV	Spot Radio	O Globo	Facebook	Debate	JN	All Channels
Político	68.0	48.0	82.6	90.8	85.1	95.5	69.7
Pessoal	7.9	34.2	8.3	4.6	0.9	0.0	14.9
Misto	24.1	17.8	9.1	4.6	14.0	4.5	15.5

• Source: author.

The candidates varied their choice of attack according to the employed media vehicle (Table 5). Personal attacks were especially employed on radio, confirming the more personal-driven character of this media in Brazil (BORBA, 2013). Approximately half of all spots focused on strictly personal attacks (25.5%) or combined personal and political attacks (23.6%). In addition to radio, televised spot also were used to make personal criticism, although mixed messages stood out (36%) compared to strictly personal ones. In other channels - O Globo, Facebook, Jornal Nacional and debates – the candidates and the news vehicles prioritized political attacks.

Table 5
Frequency of attacks according to function and channel of communication (%)

	TV Spot	Radio Spot	O Globo	Facebook	Debates	JN	All channels
Political	55.0	50.9	73.3	81.9	83.4	77.8	63.2
Personal	9.0	25.5	11.5	9.7	1.2	0.0	13.3
Mixed	36.0	23.6	15.2	8.3	15.4	22.2	23.5
Total	100	100	100	100	100	100	100

• Source: the author.

Lastly, the candidates varied their communication strategy according to the channel being used (Table 6). Of the attacks launched by Dilma on the radio, 35.9% were political, 32.1% personal and 32.1% mixed. In no other vehicle did the then incumbent president and reelection candidate distribute so evenly her attacks. In televised spots, on the other hand, few personal attacks were observed, with Dilma switching back and forth between mixed and political attacks. Aécio sustains his stable pattern, with an absolute predominance of political attacks in all channels, with mixed attacks standing out in televised spots. Marina, on her turn, favored political attacks in all vehicles, with radio being an exception, where she launched personal attacks in the majority. In this vehicle 60% of the attacks focused on the image of opponents, while in other vehicles, the percentages of attack were non- existent.

Table 6
Frequency of individual variations of attacks according to channels of communication (%)

		TV Spot	Radio Spot	O Globo	Facebook	Debate	JN
Dilma	Political	49.5	35.9	77.3	84.2	82.2	0.0
	Personal	5.7	32.1	14.8	10.5	1.1	0.0
	Mixed	44.8	32.1	8.0	5.3	16.7	0.0
Aécio	Political	56.3	64.2	63.1	72.7	84.0	80.0
	Personal	12.6	18.5	13.8	13.6	1.5	0.0
	Mixed	31.1	17.3	23.1	13.6	14.5	20.0
Marina	Political	100,0	40,0	92,3	91,7	84,6	75,0
	Personal	0,0	60,0	0,0	0,0	0,0	0,0
	Mixed	0,0	0,0	7,7	8,3	15,4	25,0

• Source: the author.

In summary, Dilma Rousseff during her balanced mixed attacks against her opponents during the campaign. Marina was characterized as fragile, self-victimizing, incoherent and, inexperienced and without political support for a future government. In parallel, she was attacked for her stance in relation to oil exploration on the Brazilian coast and her proposal to grant the Central Bank autonomy. Aécio on his turn was labeled as sexist, incompetent as the ex-governor of Minas Gerais and the defender of the rich in detriment of the poor. In this context the motto “Aécio: those who know him don´t vote for him” was widely used everytime her campaign exploited his mishaps as the former governor of the state of Minas Gerais. The opposition candidates, in turn, placed their bets on a negative campaign against a president, with political topics standing out, especially corruption and the economic crisis. Aécio also used part of his campaign to depict marina as a former member of the PT in order to identify himself as the only true opposition.

Conclusion

This article aimed to discuss the importance of considering different communication channels in the measurement of the intensity of negative ads. Today, with the increasing modernization of campaigns, voters try to get to know political parties and candidates in a complex information market. Voters read newspapers and manifestos, browse the Internet, listen to radio, watch the ads on television, talk with friends and colleagues. Neglecting this plurality of channels can, in a sense, compromise the understanding of how public opinion is formed in electoral campaigns.

In this study, we have seen that the negative ads varies from channel to channel according to the degree of interaction among the candidates, the ability to control the content of the messages, the level of regulation, and the chance of being identified with the attack. It is therefore more intense in electoral debates where direct interaction is the hallmark and, on a smaller scale, in television and radio spots. Facebook proved an instrument of mobilization and awareness campaign agenda, with little room for the attacks, perhaps due to the extreme ease in linking the attack to its author, which, in turn, could have a boomerang effect. Printed press and television, despite belonging to the same economic group, promoted coverage of the electoral process very differently. While the printed newspaper O Globo allowed criticism from opponents, especially the opposition candidates to President Dilma Rousseff, Jornal Nacional chose to promote a more purposeful debate. It is important to point out that these candidates have no control over what comes out published in the media, unlike other vehicles, which makes the comparison of these means with each other a little complicated task.

Finally, this article contributes to the literature that investigates the determinants of negative ads. Although it is not possible to compare the performance of Dilma Rousseff with previous presidents because of the absence of measurements of intensity of negative ads in all the communication channels included in this study, the data suggests that Dilma Rousseff surpassed the rates of observed attacks in previous elections compared with candidates who were also pro-government and research leaders. If in 2014 the president used 22% of her TV spots to attack opponents, in the 2010 election, when she ran for his first term, Dilma used 18% of her spots for this purpose, while Lula in 2006, only 3%. The tie with Marina Silva in the first round and the initial favoritism Aécio Neves in the second, a low scenario popularity and calls for changes (BORBA; VEIGA; MARTINS, 2015), help to explain the change in strategy. Therefore, studies that try to understand why candidates attack should include the relative strength of the incumbent government as a variable.

A more complete panorama would obviously have to include radio and TV propaganda broadcast in bloc-spots, debates promoted by radios, weekly magazines that circulate nationally, and Twitter posts, among manner other possible means of communication. In fact, the plurality of sources of information is enormous is a country with continental dimensions such as Brazil and for this very reason it impossible to include all these sources of information in a single study. However, this study may contribute to studies concerning the formation and dynamics of public opinion during elections.

• 1
  This article had resources from Foundation Carlos Chagas Filho for Research Support of the State of Rio de Janeiro (Faperj).
• 2
  Statement published in newspapers October 16, 2014.
• 3
  For a complete review of the literature on the intensity of negative ads in different countries see Nai and Walter (2015).
• 4
  In 2015, electoral legislation was modified by Law 13,165, changing, among other things, the date of beginning of campaigns, the rules of access and the distribution of time in radio and television broadcasts. In this article we will only consider the legislation in effect in 2014.
• 5
  This information was provided to the author by a publicist active in electoral and product marketing during an interview
• 6
  In this regard, I reccomend reading the decisions regarding the identification of advertising in TSE rulings.

Referências

• ALDÉ, A. As Eleições Presidenciais de 2002 nos Jornais. Alceu Revista de Comunicação. Cultura e Política, v.3, n.6, p.93-121, 2003.
• ALDÉ, A.; BORGES, J. Internet, the press and Brazilian elections: Agenda-setting on real time. Journal of Systemics, Cybernetics and Informatics, v.4, n.6, p.61-66, 2006.
• ALDÉ, A.; MENDES, G.; FIGUEIREDO, M. Tomando partido: imprensa e eleições presidenciais em 2006. Política & Sociedade, v.6, n.10, p.1-20, 2007.
• ANSOLABEHERE, S.; IYENGAR, S.; SIMON, A; VALENTINO, N. Does attack advertising demobilize the electorate? American Political Science Review, v.88, n.4, p.829-838, 1994.
• ANSOLABEHERE, S.; IYENGAR, S. Going negative: how attack ads shrink and polarize the electorate. New York: Free Press, 1995.
• AZEVEDO, A. Mídia e democracia no Brasil: relações entre o sistema de mídia e o sistema politico. Opinião Pública, v.12, n.1, p.88-113, 2006.
• BENOIT, W. Seeing spot: a function analysis of presidential television advertisement, 1952-1996. Westport: Praeger Publisher, 1999.
• BORBA, F. A propaganda negativa: estratégia e voto nas eleições brasileiras. 2012. 208 f. Tese (Doutorado em Ciência Política) – Instituto de Estudos Sociais e Políticos, Universidade do Estado do Rio de Janeiro.
• ______. Rádio e Televisão na Eleição Presidencial de 2010. In: PANKE, L.; GONDO, R. (Ed.). HGPE: Desafios e Perspectivas nos 50 anos do Horário Gratuito de Propaganda Eleitoral São Paulo: Editora Nova Consciência, 2013, p.105-126.
• ______. Propaganda negativa nas eleições presidenciais brasileiras. Opinião Pública, v.21, n.2, p.268-295, 2015a.
• ______. The Strategy of electoral spot in Brazilian presidential campaign: The decision on when and where to broadcast an attack. In: NAI, A; WALTER, A. (Ed.). New Perspectives on Negative Campaigning. Colchester: ECPR, 2015b, p.181-198.
• BORBA, F.; VEIGA, L.; MARTINS, F. Propaganda negativa na eleição Presidencial de 2014. Ou como tudo que é frágil se desmancha no ar. Revista Estudos Políticos, v.6, n.1, p.171-189, 2015.
• CARVALHO, F. L. O centro do labirinto: um estudo sobre a competição eleitoral na TV. 1994. Dissertação (Mestrado em Ciência Política). Instituto Universitário de Pesquisas do Rio de Janeiro, Rio de Janeiro.
• DOWLING, C. M.; WICHOWSKY, A. Attacks without Consequence? Candidates, Parties, Groups, and the Changing Face of Negative Advertising. American Journal of Political Science, v.59, n.1, p.19-36, 2015.
• ELMELUND-PRÆSTEKÆR, C. Beyond American Negativity: Toward a General Understanding of the Determinants of Negative Campaigning. European Political Science Review, v.2, n.1, p.137-156, 2010.
• FERES JR, J.; SASSARA, L.; BARBABELA, E. et al. A (In)clement mídia das eleições. O lado político dos jornalões. Insight Inteligência, v.17, n.67, p.46-59, 2014.
• FIGUEIREDO, M. Mídia, Mercado de Informação e Opinião Pública. In: GUIMARÃES, C.; JUNIOR, C. (Ed.). Informação e Democracia Rio de Janeiro: Editora da UERJ, 2000.
• FINKLE, S.; GEER, J.A spot check: Casting doubt on the demobilizing effect of attack advertising. American Journal of Political Science, v.42, n.2, p.573-595, 1998
• GEER, J. In Defense of Negativity: Attack Ads in Presidential Campaigns. Chicago: University of Chicago Press, 2006.
• GARRAMONE, G. Voter Response to Negative Political Ads. Journalism Quarterly, v.61, n.2, p.250-259, 1984.
• _____. Effects of Negative Political Advertising: The Roles of Sponsor and Rebuttal. Journal of Broadcasting and Electronic Media, v.29, n.2, p.147-159, 1985.
• HALLIN, D. C.; MANCINI, P. Comparing media systems: three models of media and politics. New York: Cambridge University Press, 2004.
• KAID, L.; JOHNSTON, A. Negative Versus Positive Television Advertising in U.S. Presidential Campaigns, 1960-1988. Journal of Communication, v.41, n.3, p.53-64, 1991.
• ______. Image ads and issue ads in U.S. presidential advertising: using video style to explore stylistic differences in televised political ads from 1952 to 2000. Journal of Communication, v.52, n.2, p.281-300, 2002.
• KING, J; MCCONNEL, J. The Effect of Negative Campaign Advertising on Vote Choice: The Mediating Influence of Gender. Social Science Quarterly, v.84, n.4, p.843-857, 2003.
• LAVAREDA, A. Emoções ocultas e estratégias eleitorais Rio de Janeiro: Editora Objetiva, 2009.
• LOURENÇO, L. C. Abrindo a caixa-preta: da indecisão à escolha – a eleição presidencial de 2002. 2007. 319 f. Tese (Doutorado em Ciência Política). Instituto Universidade de Pesquisas do Rio de Janeiro.
• ______. Propaganda negativa: ataque versus votos nas eleições presidenciais de 2002. Opinião Pública, v.15, n.1, p.133-158, 2009.
• JOHNSON-CARTEE, K.; COPELAND, G. Southern voter’s reaction to negative political ads in the 1986 elections. Journalism Quarterly, v.66, n.4, p.888-893, 1989.
• MASSUCHIN, M.G.; TAVARES, C.Q. Campanha eleitoral nas redes sociais: estratégias empregadas pelos candidatos à Presidência em 2014 no Facebook. Revista Compolítica, v.5, n.2, p.75-112, 2015.
• MARTIN, P. Inside the Black Box of Negative Campaign Effects: Three Reasons Why Negative Campaigns Mobilize. Political Psychology, v.25, n.4, p.545-562, 2004.
• MIGUEL, L.F. Mídia e eleições: a campanha de 1998 na Rede Globo. Dados, v.42, n.2, p.253-276, 1999.
• ______. A eleição visível: A Rede Globo descobre a política em 2002. Dados, v.46, n.2, p.289-310, 2003.
• NAI, A.; WALTER, A. The war of words: The art of negative campaigning. In: NAI, A; WALTER, A. (Ed.). New Perspectives on Negative Campaigning Colchester: ECPR, 2015, p.1-34.
• PAINTER, D. L. Collateral Damage: Involvement and the Effects of Negative Super PAC Advertising. American Behavioral Scientist, v.58, n.4, p.510-523, 2013.
• PANKE, L.; WIGGERS, M. L.; OKIDO, S.H. Eleições presidenciais brasileiras em 2010: uma análise comparativa do HGPE em rádio e televisão. In: PANKE, L.; GONDO, R. (Ed.). HGPE: Desafios e Perspectivas nos 50 anos do Horário Gratuito de Propaganda Eleitoral. São Paulo: Editora Nova Consciência, 2013, p.127-143.
• PORTO, M. P. Media Power and Democratization in Brazil: TV Globo and the Dillemmas of Political Accountability. New York: Routledge, 2012.
• RODDY, B.; GARRAMONE, G. Appeals and Strategies of Negative Political Advertising. Journal of Broadcasting and Electronic Media, v.32, n.4, p.415-427, 1988.
• SKAPERDAS, S.; GROFMAN, B. Modeling negative campaigning. American Political Science Review, v.89, n.1, p.49-61, 1995.
• SOUSA, I. J.; MARQUES, F. P. J. O Facebook como mídia de ataque – um estudo sobre a campanha negativa nas eleições presidenciais brasileiras de 2014. In: MOREIRA, A.; ARAÚJO, E; SOUSA, H. (Eds.). Comunicação e Política: tempos, contextos e desafios. Braga: CECS, 2017, p.249-289.
• STEIBEL, F. Feios, sujos e malvados: políticos, juízes e a campanha eleitoral de 2002 na TV. Rio de Janeiro: E-paper, 2007. THEILMANN, J.; WILHITE, A. Campaign tactics and the decision to attack. Journal of Politics, v.60, n.4, p.1050-1062, 1998.
• VASCONCELLOS, F. Do confronto à conciliação: debates presidenciais na TV como eventos persuasivos de campanha. 2013. 245f. Tese (Doutorado em Ciência Política) – Instituto de Estudos Sociais e Políticos, Universidade do Estado do Rio de Janeiro.
• WALTER, A. Negative Campaigning in Western Europe: Beyond the vote-seeking perspective. Zutphen: Wohrmann Print Service, 2012.
• WALTER, A.; VLIEGENTHART, R. Negative Campaigning across Different Communication Channels: Different Ball Games? International Journal of Press Politics, v.15, n.4, p.441-461, 2010.
• WEST, D. Air Wars: television advertising in election campaign, 1952-2008. Washington: Congressional Quarterly Press, 2009.

Publication Dates

• Publication in this collection
  15 Apr 2019
• Date of issue
  Jan-Apr 2019

History

• Received
  25 Mar 2018
• Accepted
  25 Jan 2019