The negative campaign as an electoral strategy from the perspective of political consultants: who to attack, when to attack and how to attack¹

Abstract

This article analyzes negative advertising as an electoral strategy. The objective is to assess three basic questions: who to attack, when to attack and how to attack. To do so, we conducted a research survey with a sample of Brazilian political consultants in order to understand how negative campaign strategies vary according to the circumstances. The results indicate that the decision to attack the opponent should be used in the final stage of the campaign, focusing on the opponent’s proposals and preferably via social networks. The decision about who to attack is influenced by the expectation of victory and positioning in the polls, but the dynamics of voting intentions change the target of the attacks and the intensity with which negative advertising should be used.

Keywords
Elections; Public Opinion; Negative Campaign; Electoral Campaign; Expert Survey

Resumo

Este artigo analisa a propaganda negativa como estratégia eleitoral. O objetivo é avaliar três questões básicas: quem atacar, quando atacar e de que maneira atacar. Para isso, aplicamos um survey a um grupo de consultores políticos brasileiros a fim de verificar como o uso da estratégia varia de acordo com as circunstâncias. Os resultados indicam que ataques ao adversário devem ser usados na parte final da campanha, tendo como foco suas propostas e, preferencialmente, via redes sociais. A decisão sobre quem atacar é influenciada pela expectativa de vitória e posicionamento nas pesquisas, mas mudanças nas intenções de voto modificam o alvo dos ataques e a intensidade com que a propaganda negativa deve ser usada.

Palavras-chave
Eleições; Opinião Pública; Propaganda Negativa; Campanha Eleitoral; Survey com Especialistas

Resumen

Este artículo analiza la publicidad negativa como estrategia electoral. El objetivo es evaluar tres cuestiones básicas: ¿a quién atacar, cuándo atacar y cómo atacar? Con este fin, aplicamos una encuesta a un grupo de consultores políticos brasileños para verificar cómo varía el uso de las estrategias de campaña negativas según las circunstancias. Los resultados indican que los ataques al oponente deben ser utilizados en la parte final de la campaña, que debe estar centrada en las propuestas del oponente y debe realizarse preferiblemente a través de las redes sociales. La decisión sobre a quién atacar está influenciada por la expectativa de victoria y el posicionamiento en las encuestas, pero los cambios en las intenciones de voto cambian el objetivo de los ataques y la intensidad con la que se debe utilizar la publicidad negativa.

Palabras clave
Elecciones; Opinión Pública; Publicidad Negativa; Campaña Electoral; Encuesta con Expertos

“Those who hit, lose. I have been saying so for a long time

People don´t like that

Duda Mendonça, a political marketing consultant. (Folha de São Paulo, October 1, 2018, our translation²)

Introduction

Responsible for the victorious “Lula Peace and love” campaign that led the ex-president to the highest executive office in Brazil in 2002, the marketer Duda Mendonça has defended, for at least 20 years, that negative advertising must be avoided. Brazilian voters, from Mendonça’s perspective, do not enjoy watching attacks and would rather see political proposals. Although it is a personal perspective on the efficacy of an electoral communication strategy, the marketer’s speech ended up disseminating a general perception about campaigns in Brazil and their characteristics. From this perspective, consultants and strategists should always seek to promote propositional messaging, which, in theory, would produce curious situations in which candidates for the same position would have nothing to say about their opponents and their proposals. The hypothesis suggests that, even when trailing, candidates should still prioritize positive advertising.

Electoral disputes involve a high degree of uncertainty for candidates, consultants and campaign strategists. The need for financial resources, limited time, multiple agendas and types of messages, the choice of communication platforms and the profile of opponents are some of the variables that amplify uncertainty. Even among competitors leading the polls of voting intentions, it is necessary to make decisions that help to maintain or increase electoral chances. For those who depend on sparking the attention and interest of voters, designing powerful and effective communication strategies is the key.

A good part of the candidates’ choices either originate or rely on the participation of campaign consultants. They are the ones who help shape and implement a candidate´s general strategy. Yet little do we know about their thoughts concerning negative advertising in Brazil. Does Duda Mendonça´s widespread notion correspond to what other consultant and strategists think? Should negative advertising be avoided even when trailing behind? What should candidates do if they become the target of negative advertising?

In the last decades, political campaigning in Brazil has become a much more professional affair. The use of specialized electoral marketing, polls and communication experts grows. This trend has led to the production of a series of investigations on how the communication variable is employed or have an impact on electoral outcomes. Many of these investigations, however, tend to focus on the content per se, leaving aside other relevant aspects. This investigation aims to lay out and chart the perceptions of an active Brazilian group of consultants and campaign strategists, in an attempt to understand what is their view concerning negative advertising and its application.

The article is divided into four sections. In the first one, we discuss in general terms the strategic dilemmas associated with negative advertising. Next, we undertake a theoretical-methodological debate on how to empirically measure these dilemmas. The third section presents the main results of the research we conducted with 109 Brazilian political consultants. In the conclusion, we discuss the main implications of this study.

Strategic dilemmas of negative advertising

Negative advertising, understood as a candidate’s decision to create an unfavorable image of his opponents, has political costs. There is plenty documented evidence in the literature showing that attacking an opponent can cause unforeseen and undesirable effects. The main one is known as the boomerang effect: when an advertisement fails to damage the target and backfires (GARRAMONE, 1984, SHAPIRO, RIEGER, 1992, KING, McCONNEL, 2003). Another less frequent risk is double impact: when a negative advertisement invokes negative feelings towards both the attacker and the target of the attacks, simultaneously damaging both (MERRIT, 1984).

Despite the costs, negative advertising is a common strategy in different contexts. In the United States, measurements show the rise of attacks in presidential elections: attacks against opponents comprise as much as half of the spots broadcasted on television (GEER, 2006; WEST, 2009). In Europe, although trends do not point upward, the use of negative advertising is also frequent in strategies pursued by candidates, especially in England (WALTER, 2012; NAI; WALTER, 2015).

In Brazil, the occurrence of attacks is not as high as in the United States, being higher in the second round than in the first. Three reasons could explain the difference: stringent regulation of electoral propaganda, which prohibits attacks between candidates and guarantee the right to reply; the multiparty system, since attacks can benefit not the one who launched them, but a third party candidate; and, lastly, an aversion of part of the electorate regarding this type of rhetorical device (BORBA, 2015).

Why then do candidates use negative advertising despite all the risks? The main explanation is that they do it because they have no alternative if they want to win the election. The device is used when candidates need to overcome unfavorable prospects, otherwise the status quo will remain unaltered until the end, leading to a loss. The hypothesis helps to explain why some candidates attack more and others do it less: those in the lead make less use of negative advertising (SKAPERDAS; GROFMAN, 1995; BORBA, 2015), since there is no pressure to change the game, as the status quo is favorable. Meanwhile, other candidates have no choice but trying to change it, lest they lose the election.

This hypothesis also helps explaining why candidates leading polls might eventually choose to launch attacks against trailing opponents who start gaining momentum. In this specific case, the candidate will attack in order to thwart the opponent´s ascent and secure a position in the dispute, as happened between José Serra (PSDB) and Anthony Garotinho (PSB) in the 2002 presidential election. In the final stretch of the first round, Serra, who had already overtaken Ciro Gomes´s (PPS) second-place spot, found himself threatened by Garotinho. The decision to attack was motivated by the need to secure his political space in the competitive structure. After all, candidates must eliminate those ahead and keep those behind at a distance.

The decision over whom to attack must also take into consideration the candidates’ ideology. In a multiparty system, such as the Brazilian one, ideology can transform a non-cooperative game in which all feel “free” to attack each other into cooperative games in which candidates avoid attacking opponents of the same ideological axis, whether out of the need to secure their support in a second round (BORBA, 2012), or out of the need to form a government coalition with parties with the same ideological proclivities in case they are elected (WALTER, 2012).

In the literature, there is evidence that candidates favor attacking opponents from opposite ideological spectrum (NAI, 2018). Let us take the 1989 presidential elections as an example. Despite Lula (PT) and Brizola (PDT) polarized the dispute for second place, they avoided attacking each other in the official electoral advertising programs (BORBA, 2012). Both candidates coordinated their strategies, as they were aware of the need of mutual support in a run-off against Fernando Collor, an outspoken right-winger.

The timing of an attack is also part of the political calculus of parties and candidates. An important consideration is whether the attack should be launched in the initial, intermediate, or final stages of the campaign. Kern (1989), for example, suggests that each stage of the electoral cycles has a different purpose: the initial one should be used to present the candidates’ biography, the intermediate one to present the agenda and proposals and the final one to attack the frailties of opponents.

Equally relevant is the time of the day one chooses to launch attacks. In Brazil, the Free Electoral Broadcast Airtime (henceforth, HGPE) is shown on television daily during the afternoon and at night. It is possible to speculate that candidates prefer to launch attacks when the audience – and thus the impact – is greater. Also related to timing, another issue has to do with the initiative: is it better to initiate an attack or to be reactive, waiting to become a target and then promoting a counter-offensive? There is not much information about this Brazil. According to a study by Garramone (1985), refuting an attack increased the so-called boomerang attack, although it did not alter the target’s perception.

Another strategic consideration pertains to the content of the attack. Candidates can choose between two types of discourse: criticizing personal traits or investing against political proposals from their opponents (BENOIT, 2015). Possibly, candidates will try to find a balance between the two strategies, as each one has specific goals. The first one aims to draw voters’ attention to the lack of preparation, honesty and experience of a candidate, while the second one focuses on the opponent’s political priorities and their negative impact on voters’ lives.

In an experiment conducted with university students, Shapiro and Rieger (1992) showed that criticism aimed at policy was perceived as fair and therefore resulted in a competitive advantage for the author of the attack. Meanwhile, attacks against the personal image of the adversary were evaluated as unfair and resulted in a boomerang effect. In Brazil, the perception of electors is similar. Borba, Veiga and Martins (2018) studied the conditions for the acceptance and rejection of negative advertising in the 2014 presidential election and found similar results.

Lastly, the final decision has to do with the vehicle of the attack. Nowadays, candidates can access multiples channels of communication in order to attack adversaries: press interviews, advertising broadcast in the traditional HGPE model of radio and television, commercial spots during regular programming, debates promoted by stations, and, more recently, social media such as Facebook and Twitter.

The 30-second spots are considered dynamic because they appear while voters are watching their favorite show – that is the moment they are caught off guard (LAVAREDA, 2009). The HGPE, broadcast in two fixed slots in the programming of radio and television stations, although considered static and often tiresome, provides a space where candidates can expose their proposals in detailed fashion (ALBUQUERQUE; TAVARES, 2018). In turn, debates provide a unique chance for voters to evaluate candidates in a live broadcast with no disguises, in a direct confrontation between adversaries (VASCONCELLOS, 2013). Finally, social media – especially Twitter and Facebook – lowered campaign costs and, at the same time, amplified the possibilities of online engagement and interaction between candidates and voters (STROMER-GALLEY, 2014; AGGIO, 2010).

Considering the relevance of attack strategies in electoral disputes, the goal of this study is to understand how campaign professionals apply them. Unlike studies that focus on the analysis of speeches, this study observed how professionals running campaigns in Brazil ponder the use of negative advertising, its conditions and the means of producing and disseminating attacks. In the next section, we explain in detail the advantages and disadvantages of our methodology.

How to evaluate campaign strategies? Content analysis versus surveying with experts

There are two viable methodological paths to analyze the strategic behavior of parties and candidates. One of them is to chart the discourse during the campaign based on an analysis of the campaign materials produced. The other one is questioning the professionals and political consultants who are involved on a day-to-day basis in the decision making process of a campaign. Each one of these methodological paths has its advantages and disadvantages that must be considered by the investigator. Among the main ones, there are decisions related to cost, speed in collecting and analyzing data and the scope of the research.

The traditional method of studying campaign strategies is the content analysis of candidates’ speeches. A cornerstone of the disciplines of Communication and Political Science, it allows the researcher to verify, among other issues, the themes and political proposals defended by the candidate, the personal image she or he intends to project and their stance towards other candidates. In Brazil, several theses, dissertations, books and articles could be cited as examples of researched based on some form of content analysis – qualitative or quantitative – to study municipal, state, or nationwide elections.

Different methodological issues must be considered in the content analysis. The first one is the material to be analyzed. Nowadays, candidates can count on a variety of channels to express their political preferences and try to convince the electorate that they are the best choice, such as government plans, manifestos, pamphlets, debates, posts on social media and advertising on radio and television.

This first choice has direct consequences on the quality of the research. The selection of just one of the channels may be insufficient, since, for example, a privileged candidate can use social networks to dialogue with the platform audience, normally more adjusted to this type of platform, while prioritizing televised debates to reach people like with advanced age. Analyzing one more material increases the research researcher’s ability to know in detail a candidate’s set of proposals, but on the other hand makes it less speedy and considerably more expensive, as it increases the costs related to data collection and analysis and training of coders of messages.

Content analysis has the main merit of being a direct data analysis procedure. In this approach, the researcher can look at what was in fact said by candidates and this is why, if costs and time for data collection are not a factor, this is the recommended method (GÉLINEAU; BLAIS, 2015).

The survey with specialists is an alternative method that consists of sending digital questionnaires to a specific set of people who have a profound knowledge of the subject as a result of their experience as academics or professionals. This modality of research has been growing and has been applied for different purposes. Its origin is related to researchers who started surveying policy specialists to uncover their ideological positions in parties, replacing the analysis of political statements (BUDGE, 2000; KEMAN, 2006; BENOIT; LAVER, 2007). Its utility has become more wide-ranging and has been employed to evaluate other subjects, such as the quality of democracies and the integrity of elections (NORRIS; GRÖMPING, 2019), the dominant tone of campaigns (GÉLINEAU; BLAIS, 2015), and even to assess the personality or communication style of political leadership (NAI; MAIER, 2018; NAI; MARTÍNEZ I COMA; MAIER, 2019).

The studies on the strategies for negative advertising are of particular importance for the purposes of this article. Pioneers on the subject, Theilamm and Willhite (1998) sent surveys – at the time through mail – to 246 consultants of the American Association of Political Consultants, inviting them to imagine themselves in a series of hypothetical electoral scenarios to find out which advice they would offer to the candidates. Their findings support the claim that the decision to attack an opponent depends on the candidate´s placement in the race. More recently, Nai (2018) sent a questionnaire to 675 experts in 35 countries, confirming outside the North American context and in a comparative perspective the expectation that negative advertising is usually aimed at leading candidates. Furthermore, Nai showed that extremists attack more, and incumbents attack less.

The main advantage of the survey with specialists is, undoubtedly, the low cost. Different platforms such as LimeSurvey and Google Docs are available free of charge and similar ones are available at relatively low costs. The speed of data collection and its scope are also considerable advantages, making it easier to conduct comparative international research. The Electoral Integrity Project, conducted by political scientist Pippa Norris, has long been collecting data on the integrity of elections at a global level based on surveys with specialists from different countries. According to the last report, for the period from 2012 to 2018 (NORRIS; GRÖMPING, 2019), the survey was able to collect responses from 3,861 specialists from 166 countries, providing information on 336 elections.

An additional advantage of the survey with specialists is its capacity to analyze several elections simultaneously. Content analysis normally favors the observation of one type of media in a single election. In the Brazilian case, scientific output has been historically concentrated on the study of the discourse of presidential candidates during the HGPE. In the last decade, the focus has become more diverse, with studies being carried out focusing on other levels of the federation and other channels, such as debates (VASCONCELLOS, 2013) and social media (AGGIO, 2020; AGGIO; LUCAS, 2013; EVANS; CORDOVA; SIPOLE, 2014; JOATHAN; MARQUES, 2020; MASSUCHIN; TAVARES, 2015; ROSSINI et al., 2018). However, analyses that incorporate and compare discursive strategies in more than one media are still rare (BORBA, 2019).

Surveys with specialists make it possible to overcome this limitation by expanding the scope of the research since these professionals are active in a wide variety of campaigns. It is thus possible to extract information about campaign strategies in different scenarios such as municipal, state and nationwide elections; with or without run-offs; according to the size of the municipality; with or without HGPE. The method also enriches the research, as these professionals possess different expertise. There are specialists in TV advertising as well as in social media, so the data collected does not refer to the analysis of a single channel of communication during a single campaign.

However, the survey with specialists is not free of inconveniences. Among the main disadvantages are the sampling process, the low response rate and the fact that this is an indirect form of data collection. To send a form, it is necessary to have a contact list of specialists (which is not always available) and “hope” that people will fill them in. The literature on online surveys reports that the response rate usually varies according to the type of respondent and purpose of the research, although it is usually lower than surveys carried out in person or over the telephone (FOWLER, 2009). In the research concerning the integrity of elections, Norris and Grömping (2019) informed that the response rate was 38% - our own survey achieved 28%.

Certainly, the main weakness of the survey with professionals is the indirect form of data collection, meaning that the researcher does not observe the phenomenon closely and has to rely on the report of others. However, this lack of precision is compensated by the inclusion of a large number of respondents. On one hand, the inclusion of participants compensated for the imprecision entailed by a single response. On the other hand, specialists base their answers on the experience of acting in many campaigns and not just one, as in content analysis, offering a much more diversified tableau (GÉLINEAU; BLAISS, 2015).

Methodology

In this study, the electronic questionnaire was sent through the SurveyMonkey platform to more than 377 Brazilian political strategists between June and November 2018. The list of the names was formed from different sources, but mainly from the list of affiliates of the Brazilian Association of Political Consultants (ABCOP), which was kind enough to encourage its members to answer the survey. Other names were included based on the suggestion of consultant friends or by other respondents.

The questionnaire comprised 28 questions concerning different issues discussed in the theoretical section above. In the invitation email, the consultants were prompted to imagine that they were involved in hypothetical campaigns for majority office and to respond what kind of advice they would give to candidates in different situations. In total, we received responses from 109 consultants, who together reported having been involved in 1,959 campaigns. This number is certainly an underestimation, as many consultants reported that they had worked for “hundreds” of campaigns or that “there were so many” that they could not inform the number. On average, each consultant had participated in 22 campaigns. The most experienced one informed 300 campaigns, and the least experienced, only one.

Of the 109 respondents, 80% was male and 20% was female. The average age of consultants was 48 years old, being that the oldest was 78 and the youngest 29. They stated that they work in the five regions of the country, although most are from the Southeast region (65%), followed by the South (21%), Northeast (19%), Midwest (11%) and North (8%). Considering experience in the job, the most experienced respondent reported that he had been working in campaigns for 50 years, and the least experienced was only working for a year.

Results

General perception of negative advertising: when, how and where to attack

The first question investigated regarded whether there was a positive perception of the use of negative campaigning. As we know, candidates are not expected to recognize the adoption of attacks in their campaigns. This very likely happens due to the fear of being negatively evaluated by voters, causing a boomerang effect. The same does not occur among strategists: 81% agrees or partially agrees that this is a legitimate strategy (Table 1). At the core of campaign coordination, there is the perception that this device can be employed, suggesting that this strategy is always or almost always part of the tactical repertoire that can be used.

Therefore, the logical question is: when to attack? Among respondents, the overall perception is in line with the literature: 59% agrees or partially agrees that attacks should be launched in the final stage of a dispute. There is also a consistent perception that it is better to fight back than to launch an attack: 54% agrees or partially agrees that the risk of negative advertising is lower when counter-attacking.

However, it is less consensual among consultants whether it is better to attack in the first or second round. As mentioned above, the Brazilian multiparty system tends to produce electoral disputes with lots of candidates, rendering the effect of attacks much more uncertain. The difficulty in gauging these effects reflects among strategists. 46% states that they agree or partially agree with the statement that the chances of success of negative advertising are greater in the second round. Another 39% states that they disagree or partially disagree and 14% neither agrees nor disagrees.

Table 1
General perceptions of negative advertising

	Agree	Partially agree	Neither agree nor disagree	Partially disagree	Disagree	Total
Negative advertising is a legitimate tool in electoral campaigns	44.4%	36.8%	1.9%	10.4%	6.6%	100%
Negative advertising is more efficient in the final stage of the campaign	23.8%	35.3%	10.5%	13.4%	17.2%	100%
The chances of success of negative advertising are greater in the second round than in the first round	20.0%	26.7%	14.3%	17.2%	21.9%	100%
Negative advertising is more efficient when it aims at proposals rather than personal traits	30.5%	35.3%	9.5%	11.4%	14.3%	100%
The risks of negative advertising are smaller when an opponent initiates them	26.7%	27.6%	13.4%	12.4%	20%	100%

• Source: the authors.

Regarding the type of attack, the consultants’ perception is in line with what the literature has revealed: attacks aimed at political proposals tend to be more efficient (65% agrees of partially agrees) that those aimed at personal traits (25% disagrees or partially disagrees). Although contemporary electoral disputes revolve around the construction and maintenance of the image of candidates as a function of the centrality of means of communication such as television (WATTENBERG, 1991; LEAL; VIEIRA, 2009), the data suggests that experts prefer launching attacks aimed at proposals, something that deserves a closer look in future investigations.

Perhaps the type of dispute (majority or proportional) also has an impact on the responses of strategists, since in Brazil the allotment of time for candidates in proportional disputes is much smaller compared to majority disputes. One must also consider the possibility of using direct communication, via social media, to disseminate political proposals and aspects relative to personal traits.

Because of the electoral legislation adopted in Brazil, candidates have access to a series of channels in order to disseminate their campaign messages. Unlike the model adopted in the US, in which competitors must buy time slots, in Brazil candidates can use mandatory HGPE time and spots to attack opponents. There is also the possibility of concentrating attacks during the course of televised debates, which are more common for majority disputes, or even in social media, a cheap resource that has been increasingly used according to a very unique logic: content can generate organic engagement (i.e., users become disseminators of the original message) increasing its reach.

Both the multiplicity of channels to promote attacks and the centrality of social media reflect the differences identified in the survey. When questioned about the best vehicle to promote attacks, 52% answered social media, 38% television, and 21.5% debates. It is interesting to observe that for 17% of consultants, radio remains the best vehicle for negative advertising, while 12% stated they were indifferent. This finding suggests that Brazilian campaigns are in line with the perception that social media have a relevant role in sparking engagement among voters in order to propagate messages (WLEIZEN, 2014), as well as attacks aimed at opponents.

The preference for a determined channel is connected to how each strategist believes an attack is efficient. While in social media attacks can be a powerful element in engaging voters, somewhat obscuring the authorship of a message, in debates candidates must own up to a direct attack. In HGPE and spots, it is also possible to identify the author of the messages more easily, which increases the chance of a boomerang attack. The choice of the best channel can also be representative of the environment in which consultants work. In Brazil, elections in small municipalities rely on a precarious network of internal communication, most without HGPE or local television news, leaving social media and radio as the only available means to disseminate information.

When specifically questioned about the best moment to launch attacks, 50% responded that the best moment to attack is at night, while 42% considered themselves indifferent to the time. The remaining 8% informed that the afternoon is the best. Although the preference for the night, it is not clear for strategists if there are advantages at this or that period. Perhaps part of this question is associated to another question asked to respondents regarding their preference for the HGPE or spots as a channel to promote attacks: 61% prefers spots, 19% prefers HGPE and 21% considers the channel irrelevant. The preference for spots reflect what the literature has pointed as one of the main advantages of this type of communication: it is short and catches the voter off guard (LAVAREDA, 2009), which increases the likelihood of efficiency.

Electoral scenarios and attack strategies

In this section, the focus is on measuring even more specific questions on the use of negative advertising. We want to test how strategists consider diverse scenarios since, as already mentioned, the employment of this device always involves risks.

The first question assessed how experts evaluate the use of attacks in a first round scenario in which there is a large number of candidates. There appears to be a consensus. Around 61% agrees or partially agrees that the risks of negative advertising are greater in this scenario, practically twice as much (32%) as those who disagree or partially disagree. The results reflect an already discussed aspect: when there are many competitors, it is not clear if the author of the attack will necessarily benefit from the strategy, something that ends up being an incentive against the use of negative advertising in the first stage of the campaign.

Strategists must also consider situations in which it is worth investing in negative advertising, even if at risk of losing support among voters. The field is divided when asked if they would recommend attacks when their candidate is in the lead of polls in the first round. As already mentioned, the attacks are preferentially employed by competitors who are behind in the polls. Around 34% of experts say that they do not recommend attacks in the first round. On the other hand, 36% affirm that they would recommend attacks if their candidate were placed third, followed by those who would do it if their candidate were placed second (27%). For those leading, attacking is recommended by only 6.8% of respondents.

However, campaigns are very dynamic affairs. Fluctuation in public opinion can prompt strategic shifts: a candidate who led the polls and announced a clean and attack-free campaign might have to reconsider. We created some hypothetical dynamic scenarios to verify how campaign strategies change according to circumstances.

In the first scenario, the expert advised a candidate in first place in the polls with a guaranteed spot in the run-off who was being attacked by opponents. 62% of the experts said that they would not advise attacking any of the opponents, 5% would attack only the one placed second, 15% almost exclusively the one placed second, and 18% would indistinctly attack those placed second and third in the polls.

In the second hypothetical scenario, the candidate being advised was placed second and the polls were pointing towards the election of the one placed first in the first round. 63% would advise attacking the one placed first exclusively, 15% both first and third place, and only 7% said they would not advise attacking anyone. The third scenario still had the candidate in second, but polls did not show the candidate first in polls winning in the first round. The strategies changed significantly: 20% of the experts said they would advise attacking almost exclusively the one placed first, 16% would attack the one placed first more, and the one placed second less, 11% would attack first and third place equally and 31% would not attack any of the candidates.

In the fourth hypothetical scenario, the candidate was placed second, but the one placed third in the polls threatened their chance of going to the second round. Only 2% said that they would advise attack almost exclusively the one placed first and 6% would attack the first one more and the third one less. The main target became the candidate gaining momentum: almost 40% would attack the third one more, and the one placed first less, while 23% would prioritize almost exclusively the third. In other words, the traditional target is no longer the leader and becomes the candidate threatening more directly to take their spot. Table 2 summarizes these results.

Table 2
Election scenarios and targets of attacks

Scenario	Strategy	Percentage
Your candidate leads the polls and has the right presence in the second round. The remaining contestants have resorted to attacking your candidate. In this case, you would advise a strategy of attacks aimed at:	1. Almost exclusively the candidate in second place 2. Second place more, third place less 3. Second and third place equally 4. Third place more, second place less 5. Almost exclusively third place 6. Neither	5.3 14.7 17.9 2.1 0.0 62.1
Your candidate is in second place and polls indicate that the opponent leading the dispute will be elected in the first round. In this case, you would advise a strategy of attacks aimed at:	1. Almost exclusively the first place 2. First place more, third place less 3. First and third place equally 4. Third place more, first place less 5. Almost exclusively third place 6. Neither	63.4 12.9 9.7 5.4 2.2 7.5
Your candidate is placed second and the polls indicate that the leading candidate will not win in the first round. In this case, you would advise a strategy of attacks aimed at:	1. Almost exclusively the first place 2. First place more, third place less 3. First and third place equally 4. Third place more, first place less 5. Almost exclusively the third place 6. Neither	20.2 15.9 10.6 20.2 4.3 30.8
Your candidate is placed second and the polls indicate that the candidate placed third has become a threat and might risk your place in the second round. In this case, you would advise a strategy of attacks aimed at:	1. Almost exclusively the first place 2. First place more, third place less 3. First and third place equally 4. Third place more, first place less 5. Almost exclusively the third place 6. Neither	2.2 6.4 20.4 39.8 22.6 9.7
Now your candidate is in third and with chances of advancing to the second round. In this case, you would advise a strategy of attacks aimed at:	1. Almost exclusively the first place 2. First place more, second place less 3. First and second place equally 4. Second place more, first place less 5. Almost exclusively the second place 6. Neither	4.3 1.1 34.4 38.7 18.3 4.3

• Source: the authors.

The last hypothetical scenario invited experts to imagine themselves working for a candidate placed third in the polls, but with chances of overtaking the one in second. According to our initial expectations, the situation turns the candidate in second into the prime target. 39% of the respondents would attack more the candidate placed second, and less the first one, while 18% would almost exclusively attack the second. Data on candidate position and decision on whether attack or not in the first round suggests that, although most acknowledge the risks of launching attacks in the first round, this decision might change depending on the competitor´s placement in the polls.

To test in another way the relationship between chances of victory and negative advertising, a new set of scenarios was presented. The goal was to observe the ideal equilibrium between stressing one´s qualities and detracting opponents according to the situation in the polls (Table 3). Strategists were asked what dose of negative advertising they would use in an ideal situation in which their candidate leads the polls and could win with a comfortable margin in a campaign marked by the lack of attacks among competitors. 79% stated they would prefer a campaign comprised almost exclusively of positive advertising, against 21% who would prefer a mostly positive campaign with only a small amount of negative advertising.

Table 3
Election scenarios and intensity of attacks

Scenarios	Almost exclusively positive	Mostly positive, a little negative	Half positive, half negative	Mostly negative, a little positive	Almost exclusively negative
The candidate you work for has a comfortable lead according to the polls. The campaign has been relatively attack free. If you can win by a comfortable margin, what kind of strategy would you advise:	79.4%	20.6%	1.0%	0.0%	0.0%
The candidate you work for has a comfortable lead in the polls, but his adversary has initiated a series of attacks that might jeopardize the victory. You can react using positive or negative advertising, or a combination of both. In this situation, what kind of strategy would you advise:	31.9%	58.7%	8.3%	1.0%	0.0%
Your adversary has initiated a series of attacks against the candidate you work for and the effects have been devastating. Your candidate, once in the lead, now trails. You can react using positive or negative advertising, or a combination of both. In this situation, what kind of strategy would you advise:	14.4%	25.7%	43.3%	11.3%	5.2%
Your candidate is far behind in the polls with virtually no chances. In this situation, what kind of strategy would you advise:	30.9%	14.4%	32.9%	15.5%	6.2%
Your adversary is not as strong as in the last scenario. You are trailing but the gap is not so large and there is still a chance of victory. In this situation, what kind of strategy would you advise:	17.7%	53.2%	19.8%	9.4%	0.0%

• Source: the authors.

The situation changes if the candidate remains in the lead, but the opponent initiates a series of attacks that jeopardize victory. 59% would recommend a strategy with mostly positive advertising and just a little negative advertising, and 31.9% would prefer a strategy almost entirely based on positive advertising. In other words, although the recommendation for candidates in the lead is not to launch attacks, this changes when there is a menace.

The third hypothetical scenario evaluated what is recommended when attacks from opponents start undermining the chances of victory of a candidate who at one point led in the polls and has been overtaken. A half-positive, half-negative communication strategy was recommend by 42%. The second most preferred strategy was one in which most advertising is positive with some negative, with 25%. This shows that even in situations in which attacks generate damages, professionals still prefer combining positive and negative strategies. The employment of negative advertising, although acknowledged as legitimate and always part of the toolkit of campaign, is not an easy decision, probably because of lack of control over unexpected reverse effects.

What about the scenario in which the candidate has no chance of winning, that is, has nothing to lose compared to the leader? Even in this situation, consultants are very cautious in the use of negative advertising, with a division among those who would rather advise an almost exclusively positive approach (31%) and those who prefer a half positive, half negative advertising (33%). Once again, there is a concern with the possible reverse effects of an entirely negative campaign.

The fifth and last hypothetical scenario asked experts to imagine that the candidate they were working for trailed in the polls, but not by a large margin and with victory within range. 53% said they would favor a mostly positive strategy. The others practically were divided among the other options: 18% answered they would choose the almost entirely positive alternative while 20% would choose the half-positive, half-negative alternative.

Lastly, we attempted to evaluate the decision to attack according to the ideological position of the competitors. As mentioned earlier, candidates from the same ideological field face some dilemmas before commenting on each other, whether because they share a similar electorate, because they need to negotiate political support in a run-off, or because they may need to form a government coalition in the future.

Strategists responded what they would do in case they were on the verge of making it into the second round but had to choose between attacking a candidate on the center, right, or left, being that the candidate they worked for had a leftist history. They were also informed that, regardless of the choice, the electoral outcome would be the same. As expected, ideology matters in multiparty scenarios. For 49%, the attacks would have to be aimed at the right-wing candidate, while 27% chose the center candidate as the target and 22% would attack the left-wing candidate.

Conclusion

This study sought to chart the perception of a group of electoral campaign consultants and strategists in Brazil concerning the use of negative advertising. To this end, we conducted a survey with 109 campaign experts in 2018. In addition to our findings, this study is an attempt to fill in a gap in the literature on electoral strategies, many of which are focused on the analysis of the content of messages. We defend that chartering the perceptions of campaign experts is a relevant set of data in order to understand the uses and application of the communication variable in an electoral dispute.

The first block of data in the research analyzed general aspects about the use of negative advertising. The first finding was the consultants’ broad approval of the employment of attacks as a legitimate candidate strategy. Despite the widespread notion spread by some marketers with wide access to the press that negative advertising must be avoided, the professionals who responded to the questionnaire were more favorable to its employment: 81% acknowledged the legitimacy of this kind of strategy.

Although there is favorable perception concerning its use, there is not much consensus among consultants about the best time of the day – whether in the morning or in the afternoon – to launch the attacks. However, there is a common understanding that it is better to attack in the first round than in the second round. The reason is that when there are many candidates, the attacks may not have the intended impact.

Connected to new technologies, strategists declared that social media are the most important platform when launching attacks. This perception can be related to the high potential of engaging voters on social media, that is, the ability to interact with content, share and expand the visibility of an electoral message. Television, however, still has strength and shows up as the second preferred vehicle, but strategists believe that spots are more appropriate for attacks than advertising broadcast in the traditional exhibition in time blocks.

Regarding the content of the attacks, consultants consider that the messages must focus more on criticizing the proposals presented by the opponents. 66% agrees or partially agrees that negative advertising must target proposals and not the image or personal attributes of competitors. This finding deserves further inquiry in future studies, as the literature has shown that the image of candidates has been increasingly used in electoral campaigns, and not the contrary.

The second block of the data mobilized by this investigation sought to test different electoral scenarios in order to identify the targets and intensity of negative advertising. Data shows that the position of a competitor in the polls is an important element for decision-making. Those who lead tend to use less or no negative advertising, but this perception changes when the candidates are in a situation of disadvantage, that is, when prospects of election are gone or when they start losing support among voters. Therefore, Duda Mendonça´s hypothesis (“whoever hits, loses”) is not supported by what experts think. Mendonça’s causal nexus, in fact, seems to be inverted: “whoever loses, hits”.

• 1
  This article was suppoted by funding from the Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (Faperj). The authors would also like to thank the Brazilian Association of Political Consultants (ABCOP-Associação Brasileira de Consultores Políticos).
• 2
  From the original: “Quem bate, perde. Digo isso há muito tempo. O povo não gosta dessas coisas” (Duda Mendonça, Folha de São Paulo, October 1, 2018).

Referências

• AGGIO, C. Campanhas online: O percurso de formação das questões, problemas e configurações a partir da literatura produzida entre 1992 e 2009. Opinião Pública, v. 16, n. 2, p.426-445, 2010.
• AGGIO, C. Comunicação eleitoral “desintermediada”, mas o quão realmente interativa? E-Compós, v. 23, p. 1-18, 2020.
• AGGIO, C.; LUCAS, R. Campanha eleitoral no Facebook: usos, configurações e o papel atribuído a esse site por três candidatos eleitos nas eleições municipais de 2012. Revista Compolítica, v. 3, n. 2, p.155-188, 2013.
• ALBUQUERQUE, A.; TAVARES, C. Q. Horário Gratuito de Propaganda Eleitoral: estilo, estratégias, alcance e os desafios para o futuro. In: FIGUEIREDO, A; BORBA, F. (org.). 25 anos de eleições presidenciais no Brasil Curitiba: Appris, 2018. p.147-169.
• BENOIT, K.; LAVER, M. Estimating party policy positions: Comparing expert surveys and hand-coded content analysis. Electoral Studies, v. 26, p. 90-107, 2007.
• BENOIT, W. Functional Theory: negative campaigning in political television spots. In: NAI, A.; WALTER, A. (ed.). New perspectives on negative campaigning: why attack politics matters. Colchester: ECPR, 2015, p. 36-47.
• BORBA, F. Propaganda negativa: estratégia e voto nas eleições brasileiras. 2012. 197f. Tese (Doutorado em Ciência Política) – Instituto de Estudos Sociais e Políticos, Universidade do Estado do Rio de Janeiro.
• BORBA, F. Propaganda negativa nas eleições presidenciais brasileiras. Opinião Pública, v.21, n.2, p. 268-295, 2015.
• BORBA, F. Medindo a propaganda negativa na TV, rádio, debates, imprensa e Facebook: o caso das eleições presidenciais de 2014. Intercom, Rev. Bras. Ciênc. Comun, v. 42, n. 1, p. 37-56, 2019.
• BORBA, F.; VEIGA, L.; MARTINS, F. B. Os condicionantes da aceitação e da rejeição à propaganda negativa na eleição presidencial de 2014. Revista Brasileira de Ciência Política, n.25, p. 205-236, 2018.
• BUDGE, I. Expert judgements of party policy positions: uses and limitations in political research. European Journal of Political Research, v. 37, p. 103-113, 2000.
• EVANS, H. K.; CORDOVA, V.; SIPOLE, S. Twitter style: an analysis of how house candidates used twitter in their 2012 campaigns. Political Science and Politics, v. 47, n. 2, p. 454-462, 2014.
• FOWLER, F. J. Survey research methods California: Sage Publications, 2009.
• GARRAMONE, G. 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.
• GARRAMONE, G. Voter response to negative political ads. Journalism Quarterly, v. 61, n. 2, p. 250-259, 1984.
• GEER, J. In defense of negativity: attack ads in residential campaigns. Chicago: University of Chicago Press, 2006.
• GÉLINEAU, F.; BLAIS, A. Comparing measures of campaign negativity: expert judgements, manifestos, debates, and advertisements. In: NAI, A.; WALTER, A. (eds.). New perspectives on negative campaigning: why attack politics matters. Colchester: ECPR, 2015, p. 63-74.
• JOATHAN, I.; MARQUES, F. P. J. A. Emotion, reason, and political attacks on Facebook: the use of rhetorical appeals in the 2014 Brazilian presidential race. In: FELDMAN, O. (org.) The rhetoric of political leadership Northampton: Edward Elgar Publishing Limited, 2020.
• KEMAN, H. Experts and manifestos: Different sources – Same results for comparative research? Electoral Studies, v. 26, p. 76-89, 2006.
• KERN, M. 30-second politics: political advertising in the eighties. New York: Praeger Publisher, 1989.
• 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: Objetiva, 2009.
• LEAL, P. R.; VIEIRA, M. O fenômeno da personalização da política: evidências comunicacionais das campanhas de Dilma e Serra em 2010. Teoria e Cultura, v. 4, n. 1, p. 41-50, 2009.
• MASSUCHIN, G. M.; 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.
• MERRITT, S. Negative political advertising: some empirical findings. Journal of Advertising, v. 13, n. 3, p. 27-38, 1984.
• NAI, A. Going negative, worldwide: towards a general understanding of determinants and targets of negative campaigning. Government and Opposition, v. 55, n. 3, p. 1-26, 2018.
• NAI, A.; MAIER, J. Perceived personality and campaign style of Hillary Clinton and Donald Trump. Personality and Individual Difference, v. 121, p. 80-83, 2018.
• NAI, A.; MARTÍNEZ I COMA, F.; MAIER, J. Donald Trump, populism, and the age of extremes: comparing the personality traits and campaigning styles of Trump and other leaders worldwide. Presidential Studies Quarterly, v. 49, n. 3, p. 609-643, 2019.
• NAI, A.; WALTER, A. The war of words: the art of negative campaigning. In: NAI, A.; WALTER, A. (ed.). New perspectives on negative campaigning: why attack politics matters. Colchester: ECPR, 2015, p.1-34.
• NORRIS, P.; GRÖMPING, M. Electoral Integrity WorldWide. Electoral Integrity Report, 2019.
• ROSSINI, P. et al Social media, opinion polls, and the use of persuasive messages during the 2016 US election primaries. Social Media and Society, v. 4, n. 3, p. 1-11, 2018.
• SHAPIRO, M; RIEGER, R. Comparing positive and negative political advertising on radio. Journalism & Mass Communication Quarterly, v. 69, n. 1, p. 135-145, 1992.
• SKAPERDAS, S.; GROFMAN, B. Modeling negative campaigning. American Political Science Review, v. 89, n. 1, p. 49-61, 1995.
• STROMER-GALLEY, J. Presidential campaigning in the internet age Oxford, UK: Oxford University Press, 2014.
• 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.
• WATTENBERG, Martin. The rise of candidate-centered politics Cambridge: Harvard University Press, 2001.
• WEST, D. Air Wars: television advertising in election campaign, 1952-2008. Washington: Congressional Quarterly Press, 2009.

Edited by

Editor: Maria Ataide Malcher

Editorial assistant: Weverton Raiol

Publication Dates

• Publication in this collection
  23 May 2022
• Date of issue
  2022

History

• Received
  07 July 2020
• Accepted
  28 Feb 2022