Abstract
Most of the e-campaigning literature claims that the Internet can reinforce political attitudes, but not change them. In this article, I analyze the issue through the receive−accept−sample (RAS) theory, which postulates that messages can change attitudes if they are both received and accepted by audiences. Based on qualitative elite interviews with 31 consultants and operatives involved in the 2008 United States presidential election, I argue that campaigns are finding new avenues to improve both reception and acceptance of their messages, and the implications merit close empirical scrutiny. The probability of message reception can be increased by citizens’ propensity to seek out issue positions online and by diffusion through low-threshold activities by supporters; the probability of message acceptance can be augmented by video and the targeting of content; finally, indirect persuasion through interpersonal communication can increase the probability of both reception and acceptance.
Keywords
‘The holy grail that everybody is looking for right now is how can you use the Internet for persuasion.’ 1 This statement, made during the 2006 US midterm elections by Jerome Armstrong, a blogger and political consultant, reflected not only the conventional wisdom among professionals at the time, but also most of the research, which has so far concluded that Internet campaigning can reinforce, rather than change, political attitudes. This study assesses whether developments in the social shaping of technology might have expanded the Web’s persuasive power in politics, which depends, among other things, on how the producers of online campaigning evaluate its electoral potential. In order to address these issues, I analyze e-campaigning in the 2008 United States presidential election through qualitative interviews with 31 consultants and operatives involved in the race. The data suggest five hypotheses regarding conditions under which online electioneering may be conducive to persuasion.
Online campaigning effects: A question of intensity
The idea that online campaigning constitutes an ‘echo chamber’ that only reinforces preexisting attitudes has been established by Bimber and Davis’s path-breaking study of the 2000 US campaign cycle. This research was premised on the idea that the contents to which citizens are exposed online result from users’ rather than producers’ decisions; ‘selectivity’ was thus identified as one of the main properties of the Internet. The authors concluded that ‘Although those who are undecided or who oppose the candidate may constitute a minority of the audience, they are not the objects of the campaign’s attention in its online presentation … The main message of candidate Web content is reinforcement’ (Bimber and Davis, 2003: 144). Therefore, digital media were seen as helpful for ‘preaching to the converted’ (Norris, 2003: 24), but not for changing attitudes. Recent reviews have similarly concluded that online electioneering has limited scope for persuasion (Davis et al., 2009: 23; Davis and Owen, 2008: 97–98). As a result, online campaigns are expected to exclusively focus on the reinforcement and mobilization of supporters (Vaccari, 2008), a goal which has nevertheless become strategic in contemporary elections because ‘loyalty to the party cannot be assumed, but must be constantly reinforced’ (Gibson and Rommele, 2001: 33).
The debate on the effects of online campaigning is part of a wider discussion inspired by Bennett and Iyengar (2008), who have suggested that theories on the effects of political communication have not fully addressed the implications of phenomena such as audience fragmentation, the demise of inadvertent exposure, and increased partisan selectivity. These developments ‘rais[e] questions about the effects of conventionally conceived persuasion campaigns’ (2008: 714) and may foster a return to the minimal effects research tradition (see also Prior, 2007). However, Holbert, Garrett and Gleason have suggested that selectivity may not necessarily prevent attitude change because ‘individuals exhibit a stronger bias toward attitude-consistent information than against attitude-discrepant information’ (Holbert et al., 2010: 19): thus, the ability to select content does not reduce encounters with political disagreement per se. In sum, selectivity may expand opportunities for attitude reinforcement, but it does not necessarily impede attitude change, or at least the conditions for it.
As noted, a minimal effects approach has been dominant in the Internet politics literature, but the idea that online campaigning only reinforces preexisting attitudes has resulted more from absence of evidence than from evidence of absence. It is the cumulative received wisdom of studies that have tried to detect the effects of e-campaigning, such as changes in attitudes and behaviors resulting from exposure to online politics, and, failing to uncover them while finding evidence for reinforcement, has concluded that only reinforcement occurs. However, a theory that predicts lack of effects is only a second-best paradigm because it cannot be falsified. 2 Moreover, claims that reinforcement is the only outcome of Internet campaigning sometimes border on technological determinism when they refer to selectivity as an intrinsic property of the medium. Better understanding the effects of online electioneering requires placing them within a testable theoretical framework rather than assuming that reinforcement is an inevitable outcome resulting from innate qualities of the Web.
Zaller’s (1992) receive−accept−sample (RAS) model of attitude change can be particularly useful to broaden the debate on the effects of e-campaigning. Zaller proves that, for attitude change to occur, change-producing messages must be first received, then, given reception, accepted (Zaller, 1992: 118–128). When they encounter counter-attitudinal messages that may foster change, individuals can resist in three ways: because they recognize received messages to be inconsistent with their predispositions and thus reject them (partisan resistance), because they possess stores of preexisting considerations that dwarf the new information (inertial resistance), and because they can provide counter-arguments based on internalized opposing considerations encountered in the media (countervalent resistance). Political awareness − that is, people’s attentiveness to and understanding of politics − is a crucial intervening variable: while the most politically aware are most likely to receive messages, they are also least likely to accept them due to the resistance mechanisms described above; by contrast, the least politically aware are most likely to accept messages due to weak resistance mechanisms, but also least likely to receive them because they do not pay attention to politics; thus, moderately aware voters are the most persuadable because they are quite likely to both receive and accept messages. A relevant independent variable is message intensity, which determines how many people are reached by the message and is a property of both the message and the audience that becomes interested in it (Zaller, 1992: 153). When intensity is high, the least aware are most likely to be influenced; when it is moderate, moderately aware voters are the most affected; when it is low, only the most aware citizens change attitudes (Zaller, 1992: 155–158; 267).
The RAS model does not make specific predictions about different sources of political information (Zaller, 1992: 43–44). While its main purpose is to account for information flows in the mass media age, Zaller maintains that its principles can apply to other domains such as face-to-face interaction (1992: 287–289). The claim that e-campaigning cannot produce attitude change can thus be understood within the RAS model: because exposure to online politics is defined as intentional, its audience is predicted to be mostly composed of ‘those who tend toward moderate to high levels of political awareness’ (Bimber and Davis, 2003: 135). Leaving aside the fact that moderately aware citizens are often the most persuadable sector of the electorate according to the RAS theory, reinforcement in e-campaigning can be understood as an empirical case in which a low-intensity message reaches only highly aware voters who select online sources that are consistent with their political preferences; these preferences, in turn, inoculate them against attitude change: hence, attitudes can only be reinforced. Thus, empirical evidence gathered while e-campaigning was at the dawn of its development has been generalized to a theory that the Internet cannot channel medium-intensity messages reaching beyond the most aware citizens. This, however, is a particular empirical configuration, not an intrinsic property of the medium. The RAS model would predict a different outcome if online campaigning achieved greater intensity.
Various developments suggest that e-campaigning could be conducive to at least medium-intensity messages that reach moderately aware voters, particularly in the United States. First, the size of the audience has grown substantially. During the 2008 US presidential election, 55 percent of Americans with Internet access used the Web for campaign-related activities (Pew Internet & American Life Project, 2009). In September 2008, more than a month before the vote, Barack Obama’s website had almost eight million separate monthly visitors and John McCain’s had more than four million. 3 While reaching a larger portion of the public does not guarantee catering to a politically broader population, modeling these developments through the RAS theory suggests that, as e-campaigning becomes more intense, it can reach persuadable moderately aware voters.
Second, e-campaigning has changed qualitatively as well as quantitatively. Web 2.0 environments, based on low-threshold contributions by large numbers of users (Chadwick, 2009), have expanded the content and interaction opportunities available to online audiences, thus increasing the probability that citizens encounter unexpected messages. Since citizens are more likely to find attitude-discordant political content in non-political online environments than in overly political spaces (Wojcieszak and Mutz, 2009), Web 2.0 platforms, being typically non-political, may enable serendipitous exposure to politically dissonant content among a much larger number of citizens than the less populated chat rooms and forums of yesteryear. The RAS model would predict that these developments imply an increase in the intensity of online political messages.
Moreover, the dialogic and networking opportunities of contemporary Internet platforms may augment their credibility, as noted by various scholars of rhetoric. Gronbeck and Wiese argue that digital media contribute to ‘de-massification and, hence, repersonalization of campaign processes’ (2005: 529), thus facilitating connections between politicians and citizens. E-campaigning may be more likely to persuade citizens because it enriches their experiences and takes their identities into account. Relatedly, Warnick claims that online interactivity ‘plays a role in persuasion by bringing users to identifying themselves with the speakers’ interests’ (2007: 71). From this perspective, the dialogic and relational affordances of e-campaigning environments may promote mutual identification, which rhetoric scholars consider a prerequisite for persuasion.
Another qualitative change is the increasing centrality of video in e-campaigning, which, in terms of the RAS model, may also imply an increase in message intensity. While early normative views of the Internet emphasized text-based deliberation, the growth of audio-visual content may ‘democratize political expression by creating a new grassroots outlet for the affective in politics’ (Chadwick, 2012). This emphasis on emotional aspects of political communication should be welcomed in a field that has emphasized reason over affect (Richards, 2004). It may also encourage researchers to reconsider questions of attitude change in light of the specific opportunities of audiovisual compared to textual content (Graber, 1990).
Research questions
This study aims to reassess the claim that e-campaigning has intrinsically little potential to enable attitude change and that it can only be conducive to attitude reinforcement. Instead, I argue that the reinforcement-only scenario is just one possible configuration in the RAS theory and that e-campaigning is evolving in ways that can be conducive to persuasion, understood through the RAS theory as the result of message reception and acceptance. I develop these hypotheses based on qualitative data on the 2008 US presidential election, which show that recent developments in online campaigning might challenge empirical findings and conclusions from the early age of Internet electioneering.
My study addresses two main questions. The first is whether and to what extent high-profile contemporary campaigns see Internet tools as capable of achieving a sufficient intensity to reach and persuade voters. The second is what conditions may be required to achieve such intensity, understood as a property of both the message and the audience paying attention to it. Answering these questions will help expand the debate on the effects of online campaigning by integrating it within the theoretical framework of the RAS model and advancing hypotheses regarding empirical conditions for online persuasion.
Method
As Foot and Schneider have observed, ‘developments in the Web practices of campaign organizations constitute the impact of the Web on politics’ (2006: 13). In other words, whether online campaigning is employed for persuasion or reinforcement depends on decisions by campaign staffers and organizations. Therefore, while I will also provide micro-level data to support some of my findings, I mostly address these questions from the meso, organizational level, and through qualitative data on the attitudes, perspectives, and experiences of e-campaigning specialists. The method of qualitative interviewing (Rubin and Rubin, 2005) is particularly useful to investigate these issues, as it illuminates experts’ assumptions and operational habits as well as their techniques and behaviors.
In order to understand the philosophies, methods and practices of online campaigners, I have followed a grounded theory approach (Glasser and Strauss, 1967) and have constructed a theoretical sample suitable for my research questions. To include different perspectives and approaches, I defined a broad target population that comprised: (a) consultants and operatives who had a high-profile role in the online campaigns of presidential candidates; (b) directors of online communications for the major parties; and (c) notable e-campaigning figures. While I focused mostly on digital media, I also increased the diversity in my sample by including perspectives from field organization and campaign management.
The resulting sampling frame consisted of a total of 66 consultants and operatives, who were contacted a maximum of five times. Twenty-eight of the 66 people, or 42 percent of the sampling frame, agreed to be interviewed. I also adopted a snowball-sampling technique by which I included six more prospective respondents, three of whom agreed to be interviewed. Thus, from a sampling frame of 72 elite informants, 31 total interviews were conducted, for a response rate of 43 percent. My sample includes 18 political consultants (nine Democrats, eight Republicans and one Independent) and 13 political operatives (seven Republicans and six Democrats). 4 The interviews were conducted between November 2008 and April 2009 and lasted on average about one hour each. All conversations took place via phone or Skype apart from two which took place in person. The interviews were conducted based on a semi-structured questionnaire where half the questions were identical for all sources, one-fourth were designed to match interviewees’ skills and campaign roles, and one-fourth were reserved for follow-up and impromptu inquiries (see Vaccari, 2010 for further information).
Before illustrating my main findings, a note on the validity of my data is appropriate. Interviewing campaign professionals may introduce three sources of bias. The first is that consultants might not be willing to reveal their ‘trade secrets’, especially when they think that they are valuable to their clients and unknown to their competitors. I attempted to avoid such a bias by emphasizing the research nature of my endeavor and by asking professionals to focus on the contextual limitations and organizational consequences of their techniques rather than prompting statements on their impact. The second source of bias is conceptual: while the RAS model focuses on attitude change, campaign professionals aim to increase their candidates’ tally of votes and employ less precise notions of ‘persuasion’ and ‘mobilization’ than political scientists do. As Mayer (2008) notes, definitions of ‘swing voters’, ‘undecided voters’, ‘independent voters’ and ‘party switchers’ correspond to different models of the relationship between campaign messages and citizens’ predispositions and behavior. However, to the extent that these definitions overlap to a certain degree, campaigners’ assessments of online persuasion help generate hypotheses that scholars can then formalize and test with greater conceptual precision. The third source of bias stems from the fact that professionals may have a vested interest in exaggerating the impact of the tools they produce. However, this has turned out not to be the case because, as I showed in another study (Vaccari, 2010), most interviewees argued that organizational and contextual factors, rather than the adoption of specific technologies, determined the effectiveness of online campaigning in 2008. Moreover, consultants interviewed with the same method in 2004 claimed that e-campaigning had little potential to change attitudes (Vaccari, 2008). Finally, consultants’ assessments are not taken at face value in this study, but their informed perspectives are evaluated in the context of established theories of media effects and employed as first steps to generating testable hypotheses regarding the conditions under which e-campaigning may be conducive to attitude change.
Findings
Most professionals interviewed for this research agree that the Internet’s main campaign functions are opinion reinforcement and supporter mobilization. Such an assessment is premised not only on the characteristics of the medium, but also on the dominant role that television still plays as a persuasion tool. For instance Eric Frenchman, who oversaw various aspects of McCain’s e-campaign, claims: ‘From a messaging perspective I still think TV is king, it’s still the place where you can move the most voters, you can still do the most persuasion marketing.’ Conversely, most interviewees identify mobilization as the key function of online campaigns, but also note that persuasion and outreach to undecided voters were not marginal in 2008. For example Thomas Gensemer, managing partner of Blue State Digital, the company that contracted for most of Obama’s online infrastructure, claims that ‘the way you do online advertising will enable more and more persuasion, but the core of what we do is still mobilization’. Thus, the main target audience of e-campaigns is still composed of supporters, and persuasion is better achieved through other channels, particularly television.
Various interviewees, however, noted that the relevance and scope of online persuasion and mobilization might vary at different campaign stages. For instance, Tracy Russo, John Edwards’s deputy online communications director, claims that ‘it’s really on those last few months of the campaign before Election Day or before a primary in certain States, that you go into organizing mode, but I think the bulk of the time is really always spent on persuasion’. In other words, persuasion must precede organization if online audiences are to be convinced to support the campaign; from that perspective, persuasion does not only entail changing voters’ minds before election day, but also reinforcing supporters’ attitudes early on, to later mobilize them in field activities.
A similar, extended definition of persuasion is offered by Michael Turk, who directed George W. Bush’s 2004 online campaign and advised Fred Thompson during the 2008 primaries. Turk claims that: persuasion just for persuasion’s sake is probably not in the best interest of the campaign early on. So what they are doing very early on [is] looking at not just persuading somebody, but persuading somebody to actually get involved.
Similarly, Joe Rospars, Obama’s new media director, claims that the distinction between online persuasion and mobilization may be false because: there were people who were undecided who would come to the website because they were interested in the issues, then they discovered that they liked Barack Obama, and then they would see all this interesting content about what was happening at the campaign and get involved from there.
Thus, the Obama campaign did not see persuasion and mobilization as mutually exclusive goals, but operated on the premise that ‘you could get people going on that escalator of the relationship really quickly if you sort of put it all in one place’, as Rospars explains. Because e-campaigning platforms integrate information and participation opportunities, the boundaries between persuasion and mobilization are more blurred online than in other media, an aspect that has not been considered by theories that assume persuasion and reinforcement to be mutually exclusive on the Web.
Moreover, consultants highlighted that persuasion may be more common in primary than in general elections because primary voters cannot rely on partisan cues to make their decisions, but need to learn about candidates’ issue positions, character and electability (Geer, 1989). As voters have to pay attention to the campaign in order to get information on these aspects, primary campaigns can affect highly aware voters more than general election campaigns. The implications are accounted for by the RAS model: during primaries, particularly in the crucial early stages, the most aware voters are more open for persuasion because they are likely to both receive messages, since they pay attention to the campaign, and accept messages, as there are few sources of countervalent information and partisan predispositions are not activated in an intra-party competition (Zaller, 1992: 253–258). In such a context, even if online campaigning only reaches the most aware voters, it can still lead to attitude change among them because this is the most persuadable group at that stage. This possibility is well understood by professionals: for instance Patrick Ruffini, who advised Rudolph Giuliani, argued that e-campaigns must be aware that primary voters are particularly keen on getting information online. Indeed, a November 2007 survey found that as many as 30 percent of eligible primary voters in Iowa and 29 percent in New Hampshire had visited one of the candidates’ websites. 5
Interviewees also claimed that e-campaigning can lead to attitude change because it can target voters that may both receive and accept messages. As I will show, various factors can be conducive to persuasion: they have to do with users’ online behavior, the targeting of persuadable voters with relevant content, the growth of the diffusion scale that can be achieved through low-threshold activities by supporters, video and rich media, and indirect persuasion through interpersonal communication. I will discuss each aspect in the following pages, while in the concluding section I will place these considerations within the RAS model to suggest testable hypotheses that specify conditions under which e-campaigning can change attitudes.
Users’ online behavior
Getting information about the candidates and the election is the most common activity among Internet users (Pew Internet & American Life Project, 2009). Various consultants interviewed for this study emphasized that candidate websites receive a significant amount of traffic on their issues and biographical pages, especially around election day, and argued that this is a sign that their audience includes many voters open to attitude change, as shown in this statement by Matt Lira, the McCain’s webmaster: ‘in the final few days before the vote you have a crush of traffic into your issues pages … The website was becoming much more of a tool for persuasion than it was in the rest of the campaign’. Similar considerations were offered by Ben Olson, McCain’s deputy e-campaign director: ‘A lot of people weren’t coming to the site to take action, they were coming to the site to learn.’ Moreover, Rospars observed that ‘if you talked to undecided voters in battleground States, a double-digit percentage of them was actually on our email list and getting updates on the campaign’, which suggests that even a subscription-based tool such as email may help campaigns address voters whose attitudes are open to change.
The growth of their online audiences is prompting e-campaigns to treat users differently than as core supporters. Scott Goodstein, Obama’s external online director, emphasizes that campaigns’ ability to reach persuadable voters online depends on how much effort each platform demands of users. While signing up on Obama’s volunteer network required a certain commitment and was presented as a first step toward greater campaign engagement, connecting with the candidate on social media was less demanding. As a result, many voters who engaged on these low-threshold platforms were less than committed supporters, and thus more open to attitude change in terms of the RAS model.
Empirical data on user behavior corroborate these statements. As Figure 1 shows, while the majority of people who visited presidential candidates’ websites and received their emails were in-partisans, sizeable minorities were out-partisans and Independents, particularly for websites. Among those who visited Obama’s website, almost one-third were Independents and one-fifth Republicans. The degree of crossover viewing is similar for McCain’s website, more than one-fourth of whose audience included Democrats and a little less than one-third Independents. While the degree of partisan trespassing is smaller for emails, it is still far from negligible, particularly for McCain, more than one-third of whose email recipients were Independents.
Partisan composition of voters who visited campaign websites and received emails from candidates in 2008 (N in parenthesis).
These data show that e-campaigning in 2008 was more intense in terms of the RAS model than earlier studies found, as it reached large sectors of the electorate and targeted sizeable numbers of Independents, who were less likely to resist messages, and of out-partisans, who were conceivably exploring the possibility of voting across party lines. These results confirm Garrett’s (2009) findings that strong supporters of 2004 presidential candidates were not more likely than weak supporters to visit the opponent candidate’s website. A similar pattern has been found in Belgium among visitors to party websites, four in ten of whom were either undecided or considering voting for a different party than the one whose website they went to (Vissers, 2009).
Targeting persuadable voters with relevant content
Campaigns increasingly employ database technologies to target voters with personalized messages (Howard, 2006; Nielsen, 2012: ch. 6). Professionals interviewed for this study argue that targeting techniques, particularly online advertising, help e-campaigns reach moderately aware voters with contents that they are likely to find relevant. From the standpoint of the RAS model, these opportunities increase the probability of attitude change to the extent that targeted messages have a greater probability of being accepted than undifferentiated content.
One key component of campaigns’ online targeting is ‘search engine optimization or search engine marketing’, as highlighted by Justine Lam, who ran Ron Paul’s online operations. Most citizens do not access campaign websites directly, but through search engine queries for topics they find relevant. For instance, Americans looking for the right government website to solve a particular problem are more likely to use search engines (44 percent) than to go directly (16 percent) to a site (Pew Internet & American Life Project, 2010). While data on the amount of search-generated traffic to e-campaign platforms are not available, most professionals interviewed for this study argued that search engine optimization and advertising were key avenues through which campaigns provided targeted voters with information that could potentially change their attitudes and behavior. As explained by Olson: ‘Whether or not someone is on our website, when they were doing a Google search [related to the campaign], they were served with non self-selected McCain online advertisements’, which the campaign had devised for that target. Tim Tagaris, who managed Chris Dodd’s online primary campaign, highlights that through Google ‘you can make sure that the people who are looking for information on certain issues are directed to pages that relate to your campaign’, which indicates that online advertising can help campaigns reach voters based on the issues they are interested in rather than their partisan predispositions. Frenchman, who oversaw McCain’s online advertising, highlights that this was the most relevant online tool for persuasion: ‘Towards the end the most persuasion marketing we were doing was … in the paid search because we knew people were looking for information on McCain’s positions … and we wanted to get the messages in front of them.’
As these statements show, the communication logic of online advertising is not a simple ‘push’ relationship by which campaigns impose content on unwilling users, but a mix of push and ‘pull’ interactions by which campaigns try to intercept citizens’ requests with targeted content. In other words, campaigns try to ensure that voters who are potentially eager to receive and accept certain messages are exposed to them. Rospars explains that the Obama campaign stimulated this interactive mechanism by ‘making sure that there was a lot of content for undecided voters … so that … there was a destination for everybody that would be more or less what they were looking for’ as a result of search engine queries.
Campaigns also have opportunities to reach specific audiences through Web 2.0 environments. Goodstein mentions that Obama targeted undecided voters through outlets such as the professional social network LinkedIn, where he first released his business policies. The conclusion is that ‘Strategically, the question is what audience you want to reach and how to reach it. The Internet has made it very easy to communicate to niches of society in places where they are comfortable.’ Thus, e-campaigns may persuade voters by addressing relatively small numbers of citizens with personalized content in relevant online platforms, which may increase these audience’s likelihood of receiving and accepting messages.
Distribution through large-scale, low-threshold activities
Citizens who engage in e-campaigning can be involved in low-threshold activities such as following candidates on social networking sites, forwarding messages, and sharing video content (Chadwick, 2009). These endeavors are not so demanding that moderately aware and less aware voters necessarily avoid them, and they can result in exposing supporters’ online social contacts to messages generated, or suggested, by campaigns, thus extending their reach beyond candidate-controlled, higher-threshold platforms.
Web 2.0 tools allow e-campaigns to rely on their supporters to distribute messages to their contacts easily and at little cost, thus creating opportunities for inadvertent exposure, which in turn contributes to increasing reception of messages. Because it is very easy to share content online, particularly on social networking sites, a new type of two-step flow occurs by which opinion leaders use digital media not only to comment on messages that they and their acquaintances have received via the mass media (Schmitt-Beck, 2003), but also to distribute messages, or, to quote Holbert et al., ‘to play a role in the filtering of political messages’ and ‘become agenda setters’ (2010: 24). This process can be initiated autonomously by users (as when a supporter forwards an email or recommends a campaign video or application), with some guidance from the campaign (as when candidates ask supporters to forward messages or post contents), through purpose-built automatic software (applications that, once installed, interact automatically with social networking sites to help users distribute campaign information and involve their contacts), or due to the default functioning of Web 2.0 platforms (as when interactions with a candidate appear in a user’s status updates).
Chris Hughes, co-founder of Facebook and director of online organizing for Obama, claimed that online word of mouth was a relevant part of the 2008 e-campaign: Any time there was a policy announcement or major campaign event we emailed on that and asked people to forward the email to their friends. That ensured that many more people were exposed to our message than would have been the case without forwarding, and that is a step towards persuasion.
Olson explains that social networking sites were used by the McCain campaign to propagate messages through supporters’ online contacts: We would build applications that, when somebody would add them, their friends would see it and this application that they were adding … were selling points [explaining McCain’s positions], and so somebody is seeing that whenever they look at their friends’ profile or in friends’ updates.
Furthermore, Turk claims that campaigns can leverage on tools such as social bookmarking, really simple syndication (RSS) feeds and embeddable video to ‘extend the headline of your site across literally thousands and thousands of other sites’.
In sum, campaigns can use social networking sites and other online tools to enroll their audiences as message multipliers, thus increasing the intensity of their communications. As Lam argues, ‘Having your supporters become your word-of-mouthpieces, going around telling their friends and emailing things to them about your candidate can lead to persuasion through the Internet.’ Web 2.0 platforms have decreased the costs of online indirect mobilization for campaigns, as their followers can spread messages quickly, cheaply and as part of their online social lives. Although campaigns cannot control how their content will be re-distributed by the online public, which includes the possibility that it will be criticized or distorted, most professionals are convinced that the benefits outweigh the cost and that, more broadly, decentralized message diffusion is inevitable. Figure 2 shows that in 2008 a significant portion of the online public engaged in digital-word-of-mouth transmission of campaign messages at both the sending and receiving ends: one-quarter of Americans who use social networking sites revealed which presidential candidate they voted for and more than four in ten discovered whom their friends voted for. Moreover, political commentary and audiovisuals were forwarded to other people by more than one-third and one-fourth of online political users, respectively. While these groups are predominantly composed of voters who identify with a party, sizeable minorities of Independents also engaged in online propagation of political information.
Percentage of voters who shared content in 2008 (N in parenthesis).
Video and rich media
Video has become a central component of e-campaigning. In 2008, 50 percent of Americans who engaged with the election online watched a Web video from a campaign or news organization and 43 percent watched one from other sources (Pew Internet & American Life Project, 2009). Cyrus Krohn, director of e-campaigning at the Republican National Committee, suggests that the diffusion of audiovisual content requires a revision of paradigms elaborated when the Web was a text-only environment: ‘the notion that the Internet is not good for persuasion’ was an argument ‘made in the late 1990s when dial-up [connections] made it difficult for creative design to flourish, but with broadband and beyond, the medium is incredibly useful for persuasion’. Justin Germany, who directed online media for McCain, suggests that ‘there is nothing more powerful than the moving image in being able to persuade someone … because it adds talk to motion’. Hughes argues that, while video is still mostly used for mobilization, it is now so widespread that it can reach undecided voters as well: A lot of the evidence suggests that [video is] really a tool for the committed to gin up their enthusiasm and get more committed, and I think that’s true, but even if that is 90% of the people who are watching videos, there is still a sizeable 10% that may very well be undecided and may get a better impression of a candidate or a campaign. And when you talk about one million views on a video, a hundred thousand of [undecided] people are a lot of people.
Moreover, Sam Graham-Felsen, Obama’s director of online outreach, emphasizes that the campaign produced over 1800 videos, many of which were targeted to different constituencies, including various types of undecided voters, so that ‘there was really something for everybody’. Online video thus contributes to increasing the intensity of campaign messages because of both its wide availability and its targeted nature; moreover, because many citizens share video through email and social networking sites (see Figure 2 above), a large amount of online video views are generated by referrals from friends and acquaintances, consistent with the bottom-up diffusion dynamics described earlier.
Indirect persuasion through interpersonal communication
An evaluation of the effects of online campaigning requires assessing the intensity its messages can achieve not only directly and online, but also indirectly and offline. Many studies (e.g. Beck et al., 2002) have shown that interpersonal communication and offline social networks strongly affect voting behavior and campaigns are increasingly aware that ‘indirect mobilization through their supporters’ social networks is cheaper and more effective than direct mobilization’ (Haynes and Pitts, 2009: 58). Vissers similarly notes that Belgian parties online ‘no longer only preach to the converted, but they also preach through the converted’ (2009: 27), using the web to motivate their supporters to persuade their acquaintances. In turn, a study of British voters found that people who engaged with the election on the Internet were more likely than the rest of the population to talk to others about it (Norris and Curtice, 2008). Thus, what superficially looks like online reinforcement could be a first step in a causal sequence whereby online mobilization spawns offline persuasion through interpersonal communication. This is all the more likely given that opinion leadership in interpersonal networks has been found to be positively correlated with interest in politics and online information seeking (Shah and Scheufele, 2006).
According to most interviewees, indirect persuasion is one of the most important potential effects of e-campaigning. Kevin Thurman, Hillary Clinton’s deputy Internet director, claimed that the Internet ‘changes persuasion in the sense that it allows us to more effectively use old ways to persuade people’ through the mobilization of supporters who in turn address other citizens; Gensemer argues that ‘the creation of surrogates as email subscribers and volunteers can become a persuasion agent because they are out there representing their support’. While no interviewee could estimate the impact of this indirect persuasion route, most emphasized that, in the 2008 campaign, online mobilization was not treated as a stand-alone activity, but as a precursor to offline persuasion via interpersonal communication. The logic of this approach is summarized by Jascha Franklin-Hodge, chief technology officer at Blue State Digital: I think the audience of websites is of course going to be mostly oriented towards supporters. Those supporters have social networks and I’m not only talking about Facebook, I’m talking about real-world social networks. These voters are the best advocates through their networks because they have a personal relationship there and so much of what this campaign tried to do came to reaching outside the limited space of the website to enable people to tie to their social networks … Those are the kinds of contexts when you rely on people, you reach out to networks of supporters and reach out of that little close circle to these people’s personal networks and I think this will be revolutionary for persuasion.
In terms of the RAS theory, indirect offline persuasion can be understood as increasing both reception and acceptance of messages, because online supporters are capable of both spreading messages to their acquaintances and using their personal influence to increase the credibility of the messages they disseminate. Campaigns have only a limited control on these endeavors and a significant part of indirect mobilization may be triggered by the interactions among the mass media, digital media, and interpersonal communication rather than campaigns’ online efforts per se. However, as previously noted, to the extent that exposure to e-campaigning correlates with opinion leadership and propensity to discuss politics, campaigns have a better opportunity to reach influential supporters online than through other media.
Conclusions
This essay has sought to make three interrelated contributions to the field of online political communication.
The first is theoretical and situated at the micro (individual) level: I suggested that the impact of online campaigning can be understood through the receive−accept−sample theory of media effects, which states that a message can cause attitude change if it is both received and accepted, and that intensity, understood as a property of both the message and the audience, determines whether effects are concentrated among voters who had little awareness, were moderately aware, or were highly aware. Within this framework, attitude change or reinforcement can be understood not as intrinsic properties of online campaigning, but as empirical outcomes conditional on observable conditions that can be specified and tested through appropriate causal models. Moreover, from a conceptual point of view consultants’ observations that the boundaries between persuasion and reinforcement are more blurred and nuanced in online campaigning than in offline campaigning are worth considering.
The second contribution is empirical and situated at the meso (organizational) level: I employed qualitative data to understand whether e-campaigning professionals treat online tools as capable of achieving a high enough message intensity to generate attitude change. The findings indicate that e-campaigns now see themselves close to the goal of reaching large numbers of voters who are at least somewhat likely to accept their messages. In other words, professionals believe that online tools can spread messages that are so intense that they affect not only highly aware, but also moderately aware voters.
The third contribution is theoretical and derives from situating the findings from the meso level into the RAS model to formulate testable hypotheses regarding conditions that can increase reception and acceptance of online messages and thus be conducive to attitude change. It is important to emphasize once again that, while I have provided some micro-level data and references to other studies in support of these arguments, I employed campaign professionals’ insights as first steps towards the building and refinement of hypotheses to be tested in future empirical research. Table 1 provides a summary of the hypotheses derived from my findings.
Summary of theoretical background and hypotheses.
Consistently with the RAS model, the tendencies highlighted by consultants can be understood as pertaining to intensity as a property of messages or audiences. Thus, users’ online behavior and distribution through large-scale, low-threshold activities may increase the number of people who can potentially receive e-campaigning messages. Conversely, the targeting of persuadable voters with relevant contents and the increasing centrality of video might affect message properties to the extent that they could enhance the probability that messages are accepted upon reception. Indirect persuasion through interpersonal communication can affect both message and audience properties: on the one hand, it may allow campaign messages to reach broader audiences due to interpersonal diffusion; on the other hand, it might increase the probability that messages are accepted because they come from a trusted source.
Future research should build on the present work by refining and testing the hypotheses summarized in Table 1, ideally through a multi-method approach. Surveys may be used to more precisely understand whether users’ online behavior leads to moderately aware and weakly affiliated voters’ being exposed to counter-attitudinal information and to more closely scrutinize the correlation between online and offline interpersonal diffusion dynamics. Experiments may be the most appropriate method to evaluate whether targeted and audiovisual political contents distributed online are more likely to be accepted than undifferentiated and text-only messages. Analysis of transactional behavioral data generated in real online interactions is the most promising avenue towards understanding whether low-threshold distribution can lead to inadvertent exposure to political messages among moderately aware voters via their Web 2.0 nonpolitical interactions. Content analysis and network analysis may also add valuable insights by improving our knowledge of how, respectively, campaigns employ online tools to increase message intensity and citizens relay messages to one another both online and offline.
The findings and hypotheses presented in this article contribute to the developing debate on media effects in the information age. While political communication research is assessing the increased costs of attitude change and the larger scope for reinforcement and minimal effects, the hypotheses developed here suggest that online political communication research, while not dismissing the notion that reinforcement and mobilization remain the cornerstones of e-campaigning, should also acknowledge that some conditions for attitude change may be emerging that are worth studying. A recognition that, while the main effect of e-campaigning is still to reinforce preexisting attitudes, some avenues for persuasion may become available, indicates a convergence between the effects of digital and mass media, which is consistent with their increasing integration into citizens’ everyday lives.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Notes
Cristian Vaccari (PhD, Università IULM, Milan, 2006) is Assistant Professor in Political Communication at the Department of Political and Social Sciences of the University of Bologna. He studies political communication in comparative perspective, with a particular focus on digital media. He has authored and co-authored four books in Italian and his work has been published in Political Communication, Party Politics, New Media & Society, Information, Communication and Society, Journal of Information Technology and Politics, European Journal of Communication and French Politics, as well as in various international edited volumes. He has been a visiting scholar at the University of Oxford, Columbia University, Massachusetts Institute of Technology, and American University.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
