Is the equalization/normalization lens dead? Social media campaigning in US congressional elections

Jason Gainous (Department of Political Science, University of Louisville, Louisville, Kentucky, USA)
Andrew Segal (Department of Political Science, University of Texas, Austin, Texas, USA)
Kevin Wagner (Department of Political Science, Florida Atlantic University, Boca Raton, Florida, USA)

Online Information Review

ISSN: 1468-4527

Publication date: 10 September 2018

Abstract

Purpose

Early information technology scholarship centered on the internet’s potential to be a democratizing force was often framed using an equalization/normalization lens arguing that either the internet was going to be an equalizing force bringing power to the masses, or it was going to be normalized into the existing power structure. The purpose of this paper is to argue that considered over time the equalization/normalization lens still sheds light on our understanding of how social media (SM) strategy can shape electoral success asking if SM are an equalizing force balancing the resource gap between candidates or are being normalized into the modern campaign.

Design/methodology/approach

SM metrics and electoral data were collected for US congressional candidates in 2012 and 2016. A series of additive and interactive models are employed to test whether the effects of SM reach on electoral success are conditional on levels of campaign spending.

Findings

The results suggest that those candidates who spend more actually get more utility for their SM campaign than those who spend less in 2012. However, by 2016, spending inversely correlates with SM campaign utility.

Research limitations/implications

The findings indicate that SM appeared to be normalizing into the modern congressional campaign in 2012. However, with higher rates of penetration and greater levels of usage in 2016, the SM campaign utility was not a result of higher spending. SM may be a greater equalizing force now.

Practical implications

Campaigns that initially integrate digital and traditional strategies increase the effectiveness of the SM campaign because the non-digital strategy both complements and draws attention to the SM campaign. However, by 2016 the SM campaign was not driven by its relation to traditional campaign spending.

Originality/value

This is the first large N study to examine the interactive effects of SM reach and campaign spending on electoral success.

Keywords

Citation

Gainous, J., Segal, A. and Wagner, K. (2018), "Is the equalization/normalization lens dead? Social media campaigning in US congressional elections", Online Information Review, Vol. 42 No. 5, pp. 718-731. https://doi.org/10.1108/OIR-08-2017-0247

Download as .RIS

Publisher

:

Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited


Since the introduction of the internet, scholars have debated the efficacy of this new technology in altering that political landscape (Bimber and Davis, 2003; Chadwick, 2006). Traditionally, politics had few easy entry points, and those often required significant resources including access to existing political networks and financial means for mass outreach. Even in the USA, where the press has been, and is, open and unrestricted, it is rarely free. Obtaining significant media coverage to advocate for candidates or support a policy position has been difficult and costly. Efforts to limit political influence of monied interests have met both political and judicial limitations. The introduction of the internet presented a possible new paradigm as it was a medium with great reach and relatively low entry cost. As a result, some scholars predicted that the internet would have an equalizing effect and lead to greater political access and a democratization of our political process (Barber, 2001; Corrado and Firestone, 1996; Gainous and Wagner, 2011, 2014).

However, other scholars suggested that the new medium would be captured by traditional political players resulting in normalization and no significant change in the basic power balance of the political systems (Bimber and Davis, 2003; Hindman 2008; Margolis and Resnick, 2000; Stromer-Galley, 2014; Ward et al., 2003). The results have been mixed (Boulianne, 2009). Even studies that support the proposition that the internet equalizes access and increases political participation are inconsistent across years, and show an effect that is often small in magnitude (Bimber and Copeland, 2013). Resultantly, much of the research has moved past the equalization vs normalization debate to focus more on the mechanisms of change.

In this research, we reconsider the equalization vs normalization debate, but refocus this frame onto how people and politicians are evolving in their use of online technologies over time, which is shifting the existing media environment and political process. More specifically, we seek to account for the evolving role of the internet as it moves from being seen and understood as a complimentary element in the existing campaign into a more independent medium. We consider its importance as part of a new hybrid media environment and propose that inconsistent findings about the role of the internet in remaking the political system and shaking the control of vested interests are because the environment itself is evolving as people shift where and how they consume information. Snapshots of social media (SM) influence in the early years of adoption are likely to be inconsistent as the media sphere is unsettled and campaigns are still adapting. A single year in isolation is an incomplete picture. Thus, an evaluation of the internet’s effect on politics must account for an evolving media environment that includes a longer measure of the effect of traditional media and online approaches.

To test this notion, we measure the importance of candidate expenditures on the success of an online SM campaign in 2012 and 2016. While SM outreach is less expensive in isolation, we theorize that successful offline campaigns are largely just reinforced by SM, making resources still a significant factor in successful political outreach. However, we expect this relationship to decline as SM becomes a primary source of information in subsequent elections. Our findings in 2012 show that candidates who spend more on their campaign actually get more utility for their SM campaign. However, this pattern does not hold in 2016, where the candidate spending is inversely related to the success of the SM campaign. In 2012, a campaign that integrated the digital with the traditional increases the effectiveness of the online efforts, as the non-digital strategy both complemented and drew attention to the SM campaign. However, by the 2016 campaign, this relationship more than dissipated. It appears now, that money does not drive online success. The implication of this finding is that in more recent election cycles, digital campaign can and does stand on its own. As the online component does not require the same level of resources as that of the traditional campaign, SM and the internet may actually provide a basis for greater equalization than earlier data and research had suggested.

Equalization, normalization and hybridity

It is understandable why many scholars and observers initially thought that the internet would disrupt traditional campaigns. The internet and SM are a significant change to traditional political outreach. The internet offers the means by which those disadvantaged under the current system can economically access the political sphere to influence policy and run for office. With the ability to allow any person to access political information and organize political activities, the internet can be seen as “equalizing” the playing field, and as a democratizing force and a driver of political participation (Bakker and de Vreese, 2011; Bimber, 2001; Bimber and Davis, 2003; Boulianne, 2009, 2015; Cho et al., 2009; Gainous and Wagner, 2011; Gibson et al., 2005; Mossberger et al., 2008; Shah et al., 2007; Xenos and Moy, 2007). Moreover, some types of online engagement, such as the use of SM platforms, may be as significant a predictor of political participation as socio-economic status (Xenos et al., 2014). A meta-analysis of the work in this area showed that 80 percent of coefficients in the measured studies show a positive relationship between SM and political participation or engagement (Boulianne, 2015). The effect of the internet on the political process in developed nations has been significant and measurable in a number of areas. The internet has had some leveling or equalizing effects in some important ways, such as campaigning, fund-raising, advertising and even political organization (Bimber and Davis, 2003; Mossberger et al., 2008; Wagner and Gainous, 2009).

However, there are significant questions about this view of the internet. Other studies have shown an inconsistent, or in some years, absent relationship between internet use and political participation (Bimber and Davis, 2003; Tolbert and McNeal, 2003). The record is even more mixed outside the USA (Gainous et al., 2015). Despite its rapid growth, the internet is not available to everyone, nor do all people even have the skills to get online (Norris, 2001). Further, while the internet is still largely open to new political players, the rules of its usage can be restricted by the state or by larger political actors. Government regulations are already limiting content and access in places such as China, East Asia, the Middle East and Russia (Gainous et al., 2015; Wagner and Gainous, 2013). Even in the USA, concerns about “fake news” are leading to proposals and debates over the role the US Government or leading SM platforms should have in contesting propaganda and misinformation available online (Coats, 2018). Finally, even in the absence of the above concerns, some scholars have concluded that digital tools will be studied and eventually captured by the dominant political players, making the influence of the new medium transient at best. (Bimber and Davis, 2003; Hindman, 2008; Margolis and Resnick, 2000; Stromer-Galley, 2014; Ward et al., 2003). In this view, the internet would simply be normalized into existing uneven political structures.

Ultimately, studies of the effect of the internet on political participation are often inconsistent across election cycles, and small in magnitude (Bimber and Copeland, 2013). Some have argued that inconsistent findings may be due in part to measurement and/or specification issues (Boulianne, 2009). Others have argued that results are dependent on the particular elections cycle, and sometimes established patterns of turnout (Tolbert and McNeal, 2003). The importance of the internet may be subject to environmental concerns as well. The technology can be limited by cultural context that limits or filters the impact of its uses (Gainous et al., 2015). The political environment, the means of access and distribution, and the level of trust in the particular media platform mediate the effectiveness of digital and traditional media across different states and systems (Müller, 2013). In a closed state, the internet, as the only open political platform, may well be the only significant source of dissident information raising the importance of the medium in that context. Scholars have already shown that technology can have an outsized influence in autocratic states where communication and dissemination platforms are at a premium (Wagner and Gainous, 2013).

In the end, many have moved from the paradigm of equalization vs normalization to focus on these underlying mechanisms, psychology or even specific platforms (see Gainous and Wagner, 2014). In attempting to understand this evolving environment, Andrew Chadwick suggested that new media is best understood through its interdependence with old media. The internet is not separate from older media, but rather it is a hybrid of newer and older media, creating new bundles of political practices and paths. This theoretical lens allows for a different view of the normalization and equalization debate. If the internet and SM are interacting with traditional media and creating a new system, then this is more akin to an evolution and changes are likely to be adapted over election cycles as the media consumers shift their behaviors and the campaigns adjust to this change. Consider that the penetration rates of SM are very different across even recent election cycles. Just 5 percent of American adults used any SM in 2005, but by 2012 it was nearly half of all Americans. Today, it is nearly 70 percent (Smith and Anderson, 2018).

In this new dynamic hybrid system, SM is evolving into a primary medium and not just another means to reinforce familiar forms of traditional messaging. At times, SM originates and amplifies campaign messages and it is the legacy media that is secondary. This evolving hybridity is a more realistic picture of the current media environment than the image of two static and competing forums. When US President Donald Trump uses the SM platform Twitter to make a statement, his “tweet” becomes part of the media coverage on multiple platforms, including broadcast and print outlets. In fact, the President often uses SM to engage with traditional media groups across multiple platforms. The effect works in reverse as well. A comedy sketch on broadcast television can, and is, distributed later through SM and other online platforms. In this way, SM is starting to subsume the traditional media making it a far more encompassing medium. It is this comprehensive approach along with the immediacy and reach of the SM that makes it increasingly viable as a primary source of information for many voters.

As SM becomes more prominent, its importance in campaign messaging should increase as well. This presents an opportunity structure for political campaigns that are under resourced to compete as the entrance requirements are less and the reach is great (Gainous and Wagner, 2014). However, to the extent that consumers are still gathering their information from traditional sources and are using SM to reinforce traditional messaging, than the power of SM will be married to the same traditional resource equation. This is the foundation of the old normalization paradigm. The strength in outreach through traditional campaign outlets should reinforce SM, making the new medium more effective as part of this combined effort. Under this static structure, there is a limit to the ability of the internet in isolation to equalize the system. However, the structure is not static, and we expect this relationship to dissipate over time as the SM becomes a primary source of information and voters become accustom to this new online environment. As consumers move to SM as a primary source of political information and engagement, the connection between the resource-driven traditional messaging and SM frays. Money should drive success initially, but as the SM environment grows, the money equation should change and greater equalization should be possible. We test this proposition below.

Data, methods and measurement

We rely on data collected by Bode and Epstein (2015) for this study for the 2012 portion of the analysis, and we replicated the data collection, with some adjustments, for the 2016 portion of the analysis. Both data sets included measures of all the concepts necessary to test whether SM campaigning is an effective low-cost means of securing votes. More directly, we test, among other things, whether those candidates with low spending able to combat high-spending candidates with an influential SM campaign. The data included both measures of SM influence (Klout scores) and candidate/district level electoral data from the 2012 and 2016 US congressional elections. While they collected data for presidential candidates, US. House candidates and US Senate candidates, we use only the data on House candidates because we wanted to the candidates in our analysis to face the same institutional electoral conditions. This allows for meaningful comparisons, and gives us the largest comparable sample size in 2012 (n=273) and 2016 (n=870). Bode and Epstein collected data across a timeline of over a year leading up to, and directly following, the election on November 6, 2012, for those candidates who had an active Twitter campaign account. By 2016, nearly all major-party incumbents and challengers had an SM presence, so we collected data on every candidate increasing the sample size dramatically. It is important to note here that while we think that we can make a valid comparison across years because data from both years capture the SM active candidates, some caution should be applied to the inferences. Nearly all incumbents and challengers were active in 2016 (84 percent had a Klout score), while in 2012, the sample was incumbent heavy as they were more likely to adopt Twitter.

Our testing strategy here is fairly straightforward. After providing some background descriptive statistics on the sample, including candidate characteristics and the distribution of SM influence, we offer a series of models aimed at examining the relative and interactive effects of campaign spending and SM influence on electoral success. We begin by modeling as our first dependent variable, SM influence (using several variations of Klout scores in 2012 and a single measure in 2016), as a function of campaign spending, controlling for incumbency status, candidate partisanship, candidate seniority, candidate age, whether the candidate was in a leadership position, candidate gender, race, candidate racial identification and the median age of their district. These models provide the foundation for our argument that traditional campaign spending actually helps digital presence. Next, we model candidate vote share (the dependent variable) as a function of SM influence (each of the various measures, respectively) with the same controls above, adding a control for the percent vote for Barack Obama in the previous election (to account for district competitiveness – higher percentages suggest a less competitive district) but excluding campaign spending. We then re-estimate the same models introducing campaign spending. This provides an initial look at whether the effect of SM influence is sensitive to levels of campaign spending. We follow this with our primary test: the same models of vote share, but with the inclusion of an interaction term between SM influence and campaign spending. This allows us to test whether the effects of SM influence are conditional on levels of campaign spending.

There are numerous measures of online influence out there, but Klout is clearly the most sophisticated and popular indicator (Bode and Epstein, 2015). Klout scores range from 0 to 100 and are based on three components: true reach (how many people a user’s account influences), amplification (how much the account influences people) and network impact (the influence of the account’s network) (Edwards et al., 2013; also see https://klout.com/corp/score). This is a robust and multi-faceted measure that captures influence by essentially calculating the reach of an account. The 0–100 score (of which we rescale to range continuously from 0 through 1 maintaining the original intervals) is based on an algorithm including inputs from hundreds of signals about a person’s social network activity. These signals are derived from eight different social sites: Facebook, Twitter, Google +, LinkedIn, Foursquare, Wikipedia, Bing and Klout itself. Klout has over time increased the transparency of their methodology, but given the proprietary nature of their product, they keep the exact details of their algorithm private. Bode and Epstein (2015) presented solid empirical evidence of the validity of the measure, and also offer a strong argument for its superiority to other SM influence indicators out there (for a thorough description of the measure, see both Bode and Epstein and https://klout.com/corp/score).

Bode and Epstein (2015) collected Klout scores for each candidate who had an active campaign Twitter account beginning on December 9, 2011, and collected data through the last day of 2012. These were all the major-party candidates in the general election (who had a Twitter account at the time of the primaries). They continued collecting data after Election Day (November 6, 2012) to take into account the residual online influence registered in Klout scores because of the rolling 90-day window used by the company. Because the data were collected across time, they were able to create several different indicators of SM influence, including the average Klout score across the duration of the campaign, the highest Klout score over the same period and the Election Day Klout score. We also use these three indicators for 2012. The first gives us an indicator of sustained influence, the second gives us an indicator of maximum influence and the last gives us a measure where influence was on the day most important to our study – the day people actually vote. Because Klout would not let us go back beyond one year, we were only able to collect the Election Day Klout Score for 2016 (again these data were collected in December to account for the rolling 90-day calculation Klout employs). This rolling 90-day calculation actually helps with the validity of the measure as related to our ultimate dependent variable, congressional vote share. Because many states allow early voting, some percentage of voters are casting a ballot before the end of the election. As such, any SM influence would have taken place before election day. Fortunately, the Klout score at the time of the election is not based solely on data from that day, but rather would be from data that also pick up where the score was in the time leading up to the election.

The data for the outcome, explanatory and control variable measures come from several sources: The Federal Elections Commission (election returns – Congressional and Obama), Opensecrets.com (campaign spending), candidates’ personal websites and biographical websites (incumbency, partisanship, seniority, age, leadership, gender, race), and US Census (median household income, median age). Both elections returns measures were coded as percentages and scaled to range continuously from 0 through 1. Campaign spending was coded in dollars and then rescaled to range continuously from 0 through 1 maintaining the original intervals. Incumbency was coded as a dummy variable (1=incumbent, 0=challenger). Partisanship was also coded as a dummy variable representing Democrats (Democrats=1, Republicans=0 with one independent in this category as well). Seniority and age were both measured in years, and then rescaled to range continuously from 0 through 1 maintaining the original intervals. Leadership was coded as a dummy variable (1=in a leadership position in Congress, 0= not in a leadership position). Gender was coded as a female dummy (1=female, 0=male). Race was coded as a dummy variable representing minority identification (1=Asian, black, Latino, other, 0=white). Median household income was coded in dollars and then rescaled to range continuously from 0 through 1, maintaining the original intervals. Finally, the median age in the district scaled to range continuously from 0 through 1.

Results

The graphs in Figure 1 show the distribution of the three respective measures of Klout score in 2012 and the single measure in 2016 across a dichotomized measure of electoral success (whether the candidate won or lost the election). The first observation that can be made from looking at these distributions is that there is considerable variation across each of the Klout measures in 2012 and 2016. This is true regardless of whether candidates were winners or losers. The next observation is that the distributions are skewed toward higher Klout scores for winners and lower Klout scores for losers, and this is particularly pronounced in 2016 (the first evidence that the 2016 results diverge from those in 2012). Further, t-tests suggest that these observed differences are not statistically significant in 2012 but the difference is significant in 2016 (p=0.00). So, an initial bivariate uncontrolled glimpse indicates that the relationship between Klout and electoral success is not strong 2012 and it is in 2016. Bode and Epstein (2015) showed significant effects in controlled models for both average and maximum Klout scores. Perhaps most importantly though, these conditional distributions do not test whether there is an effect on vote share.

Before moving to our models of vote share, we first look at the relationship between campaign spending, our other control variables and each of the Klout score measures (as dependent variables). This helps us build some context before looking at Klout scores as a predictor; we can first examine what predicts Klout scores (in particular, whether campaign spending predicts Klout). It is important to note here that we standardized all of the ordered variables in the OLS model estimates presented in Figure 2. This allowed us to compare the relative effects of campaign spending on Klout scores to all of the other controls. The results of this comparison are quite stark. The campaign spending effects dwarf those of all the other variables included in the specification in 2012 and this is not true in 2016 where incumbency is the largest effect by a huge margin. The only statistically significant variables across all four models (Klout scores – Average, Maximum and 2012/2016 Election Day) are campaign spending and seniority – both positive relationships. Leadership is positive significant in the Average model, and racial minorities tend to have higher Klout scores in 2016. Again, though what is clear here is that the magnitude of the effects of campaign spending on Klout scores – all types – is relatively large in 2012, and still matters in 2016 but the effect dramatically dissipates. As apparent in Figure 2, in fact, in 2012 it is more than 20 times larger than that of seniority, on average (see www.academia.edu/37210912/Appendix_Is_the_Equalization_Normalization_Lens_Dead_Social_Media_Campaigning_in_U.S._Congressional_Elections for a table of all the estimates), and in 2016 though, while campaign spending still influence Klout score, the incumbency effect is roughly seven times larger than campaign spending.

Simply, the more money candidates spend, the more SM influence they achieved; this was just more pronounced in 2012 which foreshadows the results that come. This clearly suggests that SM is not the great equalizer in the modern campaign in 2012, but points to a potential divergence in 2016. The competitive edge that it may provide seems to be normalized in 2012, but in 2016 the Klout edge may be driven by incumbency as opposed to money. In 2012, it is clear that who can typically raise large sums of money, seem to be building more SM influence, perhaps, as a product of more staff, heightened attention from a traditional media campaign, and the hybridity of their strategy. By 2016 though, perhaps money is not at the root of SM influence. Perhaps those with less money are able to capitalize on the relative low cost of mounting an SM campaign.

The real test of the equalization/normalization lens, though, centers on whether SM influence translates into votes, and whether this influence is conditioned by spending. If hybridity characterizes success, then we would expect this to be the case. Our first look at this is the results presented in Table I. Here, we can see that in 2012 Klout score effects on candidates’ share of the vote are, indeed, not independent of campaign spending. In fact, we have some reverse mediation in the models. While in 2016, it is only SM influence as opposed to campaign spending that matters and this effect is completely independent. The positive significance of Klout scores is contingent on controlling for campaign spending in 2012 but not in 2016. Klout Average, Klout Maximum and Klout Election Day are fairly strong positive predictors of vote share (0.24, 0.11 and 0.16, respectively), but only when controlling for campaign spending. Campaign spending is the largest effect in all of the 2012 model specifications, again by a huge margin. Interestingly though, this effect is negative. This may seem counterintuitive, but it is not, given that the sample is largely made up of incumbents in 2012. Given incumbency advantages, including name-recognition and gerrymandered districts, they do not need to spend as much money. Thus, challengers are spending more money (probability the difference >0=0.06) and are less likely to win, 44 percent relative to 92 percent (χ2=53.3, p<0.00). Incumbents simply did not need to spend as much money to win in 2012.

The only other consistent predictors in the models are leadership, where those in leadership positions are more likely to generate votes in all the 2012 models but not in 2016, and the higher the median age of the districts in 2012 the lower the percentage of vote candidates received. The latter was also not the case in 2016. Finally, candidates from high-income districts are less likely to generate votes in 2012 and 2016. This is probably because high-income districts tend to be swing districts (those with relatively balanced distribution of Democrats and Republicans) resulting in a lower vote share for candidates of either party[1]. Altogether, the most important point of these models, again, is that the Klout effects seem to be contingent on campaign spending being in the model specification in 2012 but not in 2016. Removal of the other significant variables does not significantly change this result; Klout scores still fail to reach significance at the 0.05 level unless campaign spending is in the specification in 2012.

The findings presented in Table I provide evidence suggesting that SM influence on electoral results is conditional on campaign spending in 2012 but not in 2016, but interacting the two gives a much stronger sense of exactly how this relationship works. Those interactive effects are presented in Figure 3 (and the full table of the estimates in www.academia.edu/37210912/Appendix_Is_the_Equalization_Normalization_Lens_Dead_Social_Media_Campaigning_in_U.S._Congressional_Elections). The results here are informative. In 2012, the positive effect of Klout scores on vote share is stronger for those candidates who spent more money. This is true for all three measures of Klout score in 2012. In fact, the effect of Klout for those on the lowest end of the spending measure is completely absent (see the solid line in each of the graphs in Figure 3). Conversely, the positive effect of all three Klout scores on vote share for those candidates who both were on the midpoint and highest levels of spending were strongly positive, with the strongest effect for those who spent the most (see that dashed and dotted lines). On the other hand, the effect of our Klout measure on vote share in 2016 is positive for those at the lowest level of campaign spending, and is negative for those both at the midpoint and highest levels of spending.

Our findings show that SM influence on vote share seemed to be conditional on campaign spending in 2012 but not in 2016. While at first, it seemed that those candidates who could generate large sums of money were more able to build a hybrid campaign, and as a result get a bigger bang for their SM implementation, the internet is an ever-evolving tool, and perhaps, those with fewer resources still have the opportunity to capitalize on these changes.

Conclusion

Our findings present a window into why the normalization approach to the role of the internet in the political sphere appeared feasible. Even after the rise of the internet, the legacy value of the traditional campaign and the success of well-resourced candidates were evidence that online campaigns, including SM, were not overcoming the importance of these traditional measures of campaign success. The limited role of SM in campaign success was due in part to the fact that there are not enough Twitter users and followers of candidates to make Twitter campaigns alone the driving force behind electoral outcomes. Yet, even in this environment, SM still mattered (Boulianne, 2009). Large changes in the public sphere rarely happen immediately, and the implications of the change on behavior are a further lag. Nonetheless, the initial clues were still there. If SM is already having an impact on elections, there was always extant, great potential for a more massive impact, as we got closer to universal adoption (Gainous and Wagner, 2014).

However, this answer alone would suggest that increases in internet penetration should produce a larger effect with each election cycle. Research has not shown that to be the case (Bimber and Copeland, 2013). The underlying mechanisms that are driving the change need more study. The context likely matters and the more subtle changes in voter behavior are likely being adopted unevenly across the nation. Candidate adoption in various elections is also likely to be inconsistent as the online learning curve can differ and the stimulus to adapt may be less urgent in some campaigns. Candidates have a choice about how and when to use Twitter, e-mail and a variety of hybrid media system components to capitalize on the opportunity to control the flow of information.

The implications for this research are substantial as campaign actors and political scientists continue to develop their understanding of the evolving world of political campaigns. As more variables are added to the equation of campaign success, it is important to understand not only how those variables affect campaign success individually, but also how these variables, old and new, interact with one another. While money appears less important to online success by 2016, it is not clear that this will be a persistent condition. As technology, platforms and behavior continue to change, the importance of financial resources in the efficacy of SM in political campaigns may shift as well. For illustration, SM itself is now awash with misinformation by political and state actors, and this may lead to a shift in how consumers view it. In future elections, it mike take significant resources to distinguish campaign messaging in SM from the other noise. Indeed, heaps misinformation is a rising SM strategy in campaigns (Gunitsky, 2015). Ultimately, this research indicates that the equalization theory is not dead, but the power of the internet to continue as a democratizing vehicle is still very much unsettled.

Further studies attempting to shed light on this knowledge gap could attempt to break down campaign spending by where the money is spent – on events, television ads, mailers – and seeing how money and Klout score interact when such variables are introduced. Additionally, by comparing the influence of other traditional indicators of election performance, such as opinion polling at crucial inflexion points of the campaign, to Klout score in the time after the inflexion points could allow for insight into the ability of SM campaigning to augment or overcome the highs and lows of a campaign.

Figures

Distribution of Klout scores across electoral success

Figure 1

Distribution of Klout scores across electoral success

OLS estimates of Klout scores

Figure 2

OLS estimates of Klout scores

The moderating effects of campaign spending on electoral SM influence

Figure 3

The moderating effects of campaign spending on electoral SM influence

Consequences of controlling for spending on the effects of SM influence on vote share

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Klout average 0.24* (0.10) 0.16 (0.10)
Klout maximum 0.11* (0.06) 0.09 (0.06)
Klout election day 0.18** (0.08) 0.14 (0.08) 0.18** (0.02) 0.21** (0.02)
Campaign spending −9.04** (3.42) −7.95* (3.35) −8.35* (3.36) 0.02 (0.05)
Incumbent 0.12 (0.17) 0.14 (0.17) 0.12 (0.17) 0.13 (0.17) 0.12 (0.17) 0.14 (0.17) 0.09** (0.01) 0.09** (0.01)
Democrat 0.05 (0.03) 0.05 (0.03) 0.05 (0.03) 0.05 (0.03) 0.05 (0.03) 0.05 (0.03) −0.00 (0.01) 0.00 (0.01)
Seniority 0.09 (0.09) 0.09 (0.09) 0.10 (0.09) 0.10 (0.09) 0.10 (0.09) 0.09 (0.09) 0.04 (0.03) 0.03 (0.03)
Age 0.07 (0.07) 0.08 (0.07) 0.06 (0.07) 0.07 (0.07) 0.06 (0.07) 0.07 (0.07) 0.00 (0.02) 0.01 (0.02)
Leadership 0.19* (0.07) 0.15* (0.07) 0.21** (0.07) 0.16* (0.07) 0.21** (0.07) 0.16* (0.07) −0.01 (0.03) −0.01 (0.03)
Female 0.02 (0.03) 0.01 (0.03) 0.02 (0.03) 0.01 (0.03) 0.02 (0.03) 0.01 (0.03) −0.01 (0.01) −0.01 (0.01)
Racial minority 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.05 (0.04) 0.00 (0.01) 0.00 (0.01)
%Obama 2012/2016 −0.04 (0.09) −0.02 (0.09) −0.04 (0.09) −0.02 (0.09) −0.05 (0.09) −0.03 (0.09) 0.03 (0.02) 0.02 (0.02)
Median age −0.14* (0.06) −0.15* (0.06) −0.15* (0.06) −0.15* (0.06) −0.14* (0.06) −0.15* (0.06) −0.04 (0.03) −0.04 (0.03)
Median income −0.21** (0.06) −0.22** (0.06) −0.20** (0.06) −0.22** (0.06) −0.20** (0.06) −0.22** (0.06) −0.09** (0.02) −0.08** (0.02)
Observations 253 254 253 254 252 253 643
R2 0.23 0.21 0.22 0.21 0.22 0.20 0.52

Notes: Standard errors are in parentheses. *p⩽0.05; **p⩽0.01

Note

1.

See US Census data estimates at https://medium.com/third-way/the-story-of-americas-swing-districts-90e809bbf864 (accessed March 13, 2018).

References

Bakker, T.P. and de Vreese, C.H. (2011), “Good news for the future? Young people, internet use, and political participation”, Communication Research, Vol. 38 No. 4, pp. 451-470.

Barber, B.R. (2001), “The uncertainty of digital politics”, Harvard International Review, Vol. 23 No. 1, pp. 42-47.

Bimber, B. (2001), “Information and political engagement in America: the search for effects of information technology at the individual level”, Political Research Quarterly, Vol. 54 No. 1, pp. 53-67.

Bimber, B. and Copeland, L. (2013), “Digital media and political participation over time in the US”, Journal of Information Technology & Politics, Vol. 10 No. 2, pp. 125-137.

Bimber, B.A. and Davis, R. (2003), Campaigning Online: The Internet in US Elections, Oxford University Press, Oxford and New York, NY.

Bode, L. and Epstein, B. (2015), “Campaign Klout: measuring online influence during the 2012 election”, Journal of Information Technology & Politics, Vol. 12 No. 1, pp. 133-148.

Boulianne, S. (2009), “Does internet use affect engagement? A meta-analysis of research”, Political Communication, Vol. 26 No. 2, pp. 193-211.

Boulianne, S. (2015), “Social media use and participation: a meta-analysis of current research”, Information, Communication, and Society, Vol. 18 No. 5, pp. 524-538.

Chadwick, A. (2006), Internet Politics: States, Citizens, and New Communication Technologies, Oxford University Press, Oxford.

Cho, J., Shah, D.V., McLeod, J.M., McLeod, D.M., Scholl, R.M. and Gotlieb, M.R. (2009), “Campaigns, reflection, and deliberation: advancing an O-S-R-O-R model of communication effects”, Communication Theory, Vol. 19 No. 1, pp. 66-88.

Coats, D. (2018), “Worldwide threat assessment of the US intelligence community”, available at: www.dni.gov/files/documents/Newsroom/Testimonies/2018-ATA---Unclassified-SSCI.pdf. (accessed August 9, 2018).

Corrado, A. and Firestone, C.M. (Eds) (1996), Elections in Cyberspace: Towards a New Era in American Politics, Aspen Institute, Washington, DC.

Edwards, C., Spence, P.R., Gentile, C.J., Edwards, A. and Edwards, A. (2013), “How much Klout do you have … a test of system generated cues on score credibility”, Computers in Human Behavior, Vol. 29 No. 5, pp. A12-A16.

Gainous, J. and Wagner, K.M. (2011), Rebooting American Politics: The Internet Revolution, Rowman and Littlefield, Lanham, MD.

Gainous, J. and Wagner, K.M. (2014), Tweeting to Power: The Social Media Revolution in American Politics, Oxford University Press, New York, NY.

Gainous, J., Wagner, K.M. and Abbott, J.P. (2015), “Civic disobedience: does internet use stimulate political unrest in East-Asia?”, Journal of Information Technology & Politics, Vol. 12 No. 2, pp. 219-236.

Gibson, R.K., Lusoli, W. and Ward, S. (2005), “Online participation in the UK: testing a ‘contextualised’ model of internet effects”, The British Journal of Politics & International Relations, Vol. 7 No. 4, pp. 561-583.

Gunitsky, S. (2015), “Corrupting the cyber-commons: social media as a tool of autocratic stability”, Perspectives on Politics, Vol. 13 No. 1, pp. 42-54.

Hindman, M. (2008), The Myth of Digital Democracy, Princeton University Press, Princeton, NJ.

Margolis, M. and Resnick, D. (2000), Politics as Usual: The Cyberspace Revolution, Sage, London.

Mossberger, K., Tolbert, C.J. and McNeal, R.S. (2008), Digital Citizenship: The Internet, Society and Participation, MIT Press, Cambridge.

Müller, J. (2013), Mechanisms of Trust: News Media in Democratic and Authoritarian Regimes, Campus Verlag, Frankfurt.

Norris, P. (2001), Digital Divide, Civic Engagement, Information Poverty, and the Internet Worldwide, Cambridge University Press, New York, NY.

Shah, D.V., Cho, J., Nah, S., Gotlieb, M.R., Hwang, H., Lee, N.-J. and McLeod, D.M. (2007), “Campaign ads, online messaging, and participation: extending the communication mediation model”, Journal of Communication, Vol. 57 No. 4, pp. 676-703.

Smith, A. and Anderson, M. (2018), “Social media use in 2018”, Pew Research Center, Internet and Technology, Washington, DC, available at: www.pewinternet.org/2018/03/01/social-media-use-in-2018/ (accessed August 9, 2018).

Stromer-Galley, J. (2014), Presidential Campaigning in the Internet Age, Oxford University Press, New York, NY.

Tolbert, C.J. and McNeal, R.S. (2003), “Unraveling the effects of the internet on political participation?”, Political Research Quarterly, Vol. 56 No. 2, pp. 175-185.

Wagner, K. and Gainous, J. (2009), “Electronic grassroots: does online campaigning work?”, Journal of Legislative Studies, Vol. 15 No. 4, pp. 502-520.

Wagner, K. and Gainous, J. (2013), “Digital uprising: the internet revolution in the middle east”, Journal of Information Technology and Politics, Vol. 10 No. 3, pp. 261-275.

Ward, S., Gibson, R. and Lusoli, W. (2003), “Online participation and mobilization in Britain: hype, hope and reality”, Parliamentary Affairs, Vol. 56 No. 4, pp. 652-688.

Xenos, M. and Moy, P. (2007), “Direct and differential effects of the internet on political and civic engagement”, Journal of Communication, Vol. 57 No. 4, pp. 704-718.

Xenos, M., Vromen, A. and Loader, B.D. (2014), “The great equalizer? Patterns of social media use and youth political engagement in three advanced democracies”, Information, Communication & Society, Vol. 17 No. 2, pp. 151-167.

Further reading

Bennett, W.L. and Segerberg, A. (2012), “The logic of connective action: digital media and the personalization of contentious politics”, Information, Communication & Society, Vol. 15 No. 5, pp. 739-768.

Bode, L. (2012), “Facebooking it to the polls: a study in online social networking and political behavior”, Journal of Information Technology & Politics, Vol. 9 No. 4, pp. 352-369.

Bode, L., Lassen, D.S., Kim, Y.M., Shah, D.V., Fowler, E.F., Ridout, T. and Franz, M. (2016), “Coherent campaigns? Campaign broadcast and social messaging”, Online Information Review, Vol. 40 No. 5, pp. 580-594.

Copeland, L. and Bimber, B. (2015), “Variation in the relationship between digital media use and political participation in US elections over time, 1996–2012: does Obama’s reelection change the picture?”, Journal of Information Technology & Politics, Vol. 12 No. 1, pp. 74-87.

Gainous, J., Wagner, K. and Marlowe, A. (2013), “Traditional cleavages or a new world: does online social networking bridge the political participation divide?”, International Journal of Politics Culture and Society, Vol. 26 No. 2, pp. 145-158.

Gainous, J., Wagner, K.M. and Gray, T. (2016), “Internet freedom and social media effects: democracy and citizen attitudes in Latin America”, Online Information Review, Vol. 40 No. 5, pp. 712-738.

Gray, T., Gainous, J. and Wagner, K.M. (2016), “Gender and the digital divide in Latin America”, Social Science Quarterly, Vol. 98 No. 1, pp. 326-340.

Mossberger, K., Tolbert, C.J. and Stansbury, M. (2003), Virtual Inequality: Beyond the Digital Divide, Georgetown University Press, Washington, DC.

Norris, P. (2002), Democratic Phoenix: Reinventing Political Activism, Cambridge University Press, New York, NY.

Shah, D.V. (1998), “Civic engagement, interpersonal trust, and television use: an individual-level assessment of social capital”, Political Psychology, Vol. 19 No. 3, pp. 469-496.

Shah, D.V., Cho, J., Eveland, W.P. Jr and Kwak, N. (2005), “Information and expression in a digital age: modeling internet effects on civic participation”, Communication Research, Vol. 32 No. 5, pp. 531-565.

Shah, D.V., Kwak, N. and Holbert, R.L. (2001), “‘Connecting’ and ‘disconnecting’ with civic life: patterns of internet use and the production of social capital”, Political Communication, Vol. 18 No. 2, pp. 141-162.

Corresponding author

Jason Gainous is the corresponding author and can be contacted at: jason.gainous@louisville.edu

About the authors

Jason Gainous is Professor of Political Science at the University of Louisville. He has published two coauthored books, one with Oxford University Press (Tweeting to Power: The Social Media Revolution in American Politics) and one with Rowman and Littlefield (Rebooting American Politics: The Internet Revolution). He has also published various articles in journals including American Politics Research, Journal of Information Technology & Politics, Online Information Review, Political Research Quarterly, Political Communication, Social Science Quarterly and Statistical Science among others.

Andrew Segal is Doctoral Student at the University of Texas. Before starting the program, he did his undergraduate work at the University of Louisville.

Kevin Wagner is Professor of Political Science at Florida Atlantic University. He has published two coauthored books, one with Oxford University Press (Tweeting to Power: The Social Media Revolution in American Politics) and one with Rowman and Littlefield (Rebooting American Politics: The Internet Revolution). He has also published various articles in journals including Journal of Information Technology & Politics, Online Information Review, Politics & Policy, Social Science Quarterly, among others.