Styles of Social Media Campaigning and Influence in a Hybrid Political Communication System

Abstract

Social media have the potential to influence power relations in political parties as they allow individual candidates to campaign more independently of the central party. In this paper, we scrutinize the relationship between individualization and digital social media in a study that combines the 2013 Norwegian Candidate Survey with candidates’ Twitter data. We ask, first, to what extent are social media used as an individualistic campaign tool? Second, does an individualized social media campaign style increase influence in the Twitter sphere? Third, what constitutes success on Twitter? We found two main styles of social media campaigning: a party-centered and an individualized style. Moreover, an individualized style did increase the possibility of being active on Twitter, but it had a negative effect on Twitter influence. The Twitter influentials are young, male, and relatively centrally placed in their parties. In a hybrid communication system, it appears that the candidates who gain influence in social media are those who are able to create a synergy between traditional media channels and social media.

Keywords

election campaign social media political parties candidates twitter

New digital media technology can influence power relations in party politics in several and sometimes contrary ways. For instance, digital media have been said to increase centralization processes, with increasing control of the campaign strategy transferred to the party leadership. This is essential in the modernization perspective of campaign change (e.g., Farrell and Webb 2000) and is understood as part of the general centralization processes in political parties (e.g., Katz and Mair 1995). This resonates well with a key development in established democracies in the twentieth century: the increased attention on the role of individuals in politics—often labeled the personalization of politics (Kriesi 2012). The waning of political cleavages and the alleged “decline of parties” (Dalton and Wattenberg 2000), as well as the dominant media logic in political campaign coverage, are factors driving such individualization processes (Poguntke and Webb 2005; Swanson and Mancini 1996). The role of individuals has particularly attracted attention in studies of election campaigns, and the focus has mostly been on the increasing importance of party leaders (Kriesi 2012; Poguntke and Webb 2005). In recent years, new digital media have also been identified as a channel with the potential to increase the focus on the personal side of politics; surely, social media are also another channel for party leaders to highlight themselves, as well as organize and centralize the party.

Election campaigns are not, however, only fought by party leaders, but also by the candidates standing for election. A focus on individual candidates represents what Balmas et al. (2014) called decentralizing personalization. Internet technology, especially the new online social media such as Facebook and Twitter, provides individual candidates with new opportunities to reach out to voters more independently of the central party (Karlsen 2011; Zittel 2009). With the adoption of social media, the campaigning media landscape has been transformed and described as a “hybrid media system” (Chadwick 2013). In such a media system, politicians and campaign teams target content to different audiences through a variety of media channels (Chadwick et al. 2016). Potentially, because social media allow decentralized and costless content production and diffusion, the hybrid media system may change the relations of power between actors involved in an election campaign.

In this article, we explore empirically whether the power balance between candidates and parties has changed as a result of social media use in election campaigns through a study of candidates running for parliament in the 2013 Norwegian campaign. In an electoral system based on proportional representation by list such as the Norwegian system, candidates may use social media in an election campaign with two nonexclusive goals: They may aim at mobilizing the electorate for their party, increasing their chances to be elected, and they may invest in building their reputational and political capital to increase their power, influence, and autonomy within their party. Both aims are partially dependent upon their ability to reach and influence audiences either as the result of their activity in social media or by accessing traditional media.

We ask three interrelated research questions aimed at assessing the extent to which social media are an avenue for candidates to increase their power, influence, and autonomy in relation to parties’ centralized apparatus: First, to what extent are social media used as an individualistic campaign tool? We explore the extent to which social media are considered important for different communicative aims—including conveying one’s personal side. Second, does an individualized social media campaign style increase influence in the Twitter sphere? We use a combination of survey and Twitter data to study the extent to which an individualized social media campaign style increases candidates’ activity and influence on Twitter. Third, what constitutes success on Twitter? To answer this question, we study the profile of those we call Twitter influentials—the politicians who are most visible on Twitter.

Our data comprise a combination of the 2013 Norwegian Candidate Survey and the candidates’ Twitter data. The Norwegian Candidate Survey is a survey of all candidates who ran in the 2013 national election. We combined the survey data with Twitter data gathered through the Twitter Application Programming Interface (API) and through a social media analytical tool developed by Crimson Hexagon. Norway is an interesting case, as its politics and campaigning are party centered, and there are limited structural incentives for politicians to highlight themselves during campaigns (Karlsen and Skogerbø 2015). Investigating individualization tendencies within such a setting can shed light on specific effects of social media on processes of individualization.

The results show that politicians use a variety of platforms in today’s hybrid political communication system and that social media are considered an essential part of candidates’ campaign media mix. Although a relatively small proportion report that they use social media to convey their personal side and to be visible to others in their own party, about half of the candidates using social media found them useful for this purpose. This is related to the candidates’ main communicative campaign style: a focus on their own candidacy. An individualized social media campaign style is positively related to activity on Twitter but negatively correlated to influence. The Twitter influentials, however, find social media useful for individual purposes, and their influence is rooted in a hybrid system of political communication.

The Role of Individuals in Party Politics

In the United States, the tendency toward individualization is evident among candidates running for Congress to the extent that the system is labeled “candidate-centered” (Agranof 1972; Brox and Shaw 2006; Plasser and Plasser 2002).¹ In the 1960s, American political parties changed their nomination process and introduced the primaries. One result of the reforms was that candidates set up campaign organizations and expertise independently of the party (Agranof 1972; Brox and Shaw 2006). Candidates hire a campaign manager and additional campaign staff. They identify their own campaign strategy, conduct their own polling, and organize volunteers in grassroots campaigns. Consequently, Plasser and Plasser (2002) distinguished between a U.S. candidate-centered style and a West European party-centered style of campaigning. In the West European party-centered model, the centralizing efforts of the party leadership are emphasized (Plasser and Plasser 2002). However, the level of individualized candidate campaigning can differentiate between and within the countries of Western Europe (Karlsen and Skogerbø 2015).

The concept of individualized campaigning refers to a situation where the candidates campaign independently of the party (Zittel and Gschwend 2008: 980). In this article, the attention is mainly on the “communicative focus” dimension of individualized campaigning, which refers to the focal point of the candidates’ communicative aim. On one hand, the communicative goal of the candidates might be to create as much attention as possible for themselves. On the other hand, it might be to create as much attention as possible for their parties.

Studies of political parties’ use of the Internet in the electoral arena have primarily focused on party Web sites and have revealed two main tendencies (Gibson and Ward 2009: 93): (1) standardization of information dissemination on the party Web sites and (2) conservatism in using the interactive possibilities offered by the new media. Nonetheless, the introduction of social media appears to have changed the focus toward mobilization and the innovative use of networks (Gibson 2015; Karlsen 2013). Moreover, social media has increased the scholarly interest in the individual candidates’ use of new technology in Europe as well (e.g., Graham et al. 2014; Karlsen 2013; Skovsgaard and van Dalen 2013; Spierings and Jacobs 2014). Studies of Twitter clearly dominate the literature as compared with studies of Facebook, primarily because Twitter data are more easily available.

Four main emerging areas of research can be identified based on the analysis of Twitter data. The first focuses on politicians’ reasons for using Twitter and on the demographic and political factors influencing Twitter adoption (Chi and Yang 2010, 2011). The second area of research involves the content analysis of tweets; it provides various classifications of politicians’ use of Twitter based on the tweet content (Glassman et al. 2010; Golbeck et al. 2010). The third area of research investigates the extent to which politicians use Twitter to interact with their electorate and how interactivity on Twitter may influence political communication by fostering dialogue or reinforcing one-way communication (Grant et al. 2010; Jackson and Lilleker 2011). The last area of research addresses the networks and media system, constituted by Twitter, and focuses on the networks of communication (Bruns 2012) emerging in election campaigns by collecting tweets based on given hashtags (Burgess and Bruns 2012; Larsson and Moe 2012) or by exploring the hyperlinks embedded in political tweets (Moe and Larsson 2013).

Recent studies of Facebook include studies of why parties use social media and whether social media communication entails permanent campaigning (Larsson and Kalsnes 2015). Several studies have also investigated the notion that social media will increase personalized communication and individualized campaigns (e.g., Karlsen 2011). Social media offer candidates who want to highlight themselves more opportunities to do so not only by reaching out to their own followers but also by reaching a secondary audience through the flow of messages in networks (Karlsen 2015; Vaccari and Valeriani 2015).

The existing literature shows that parties and politicians embrace the new opportunities offered by social media. But we do not know much about why they think it is important and whether social media are used differently by different candidates. In this context, we are particularly interested in whether candidates with a communicative aim of focusing on their own candidacy use social media for such purposes. However, other factors aside from communicative aims might influence social media use. First, social media use might differ based on context. The effect of social media is most likely related to the role that individual candidates already play in electoral politics. As mentioned above, the role of individual politicians differs between established democracies and varies based on the institutional setting, traditions, and culture (Karlsen and Skogerbø 2015; Plasser and Plasser 2002). Social media might be expected to increase tendencies toward individualized campaigning in systems where candidates are already more or less independent from parties (Karlsen and Skogerbø 2015). The extent of individualized campaigning also differs between candidates in party-centered systems. Candidates who are inclined to focus on themselves might recognize the potential of social media to promote themselves and use social media to a greater extent than others.

Social media should also be considered in relation to other communication channels and platforms. Media systems are arguably in the middle of a rather chaotic transition period induced by the digital media (Chadwick 2013: 4). This transition, characterized by interaction between older and newer media logics, has been labeled “a hybrid media system” (Chadwick 2013). In this hybrid political communication system, actors can use a wide range of media platforms to create, steer, and respond to a flow of communication (Chadwick 2013: 4). In this perspective, the question is not whether social media replace earlier communicative platforms but what role different social media play in relation to a wide range of new and old media platforms in the greater political communicative system.

Norway

The Norwegian political system is characterized by a parliamentary government, a stable multiparty system, and well-organized membership parties (Allern et al. 2016). Hallin and Mancini (2004) identified the Norwegian media system as a typical example of the democratic corporatist model, which is characterized by the historical coexistence of commercial media and media tied to organized social and political groups, as well as by a relatively active but legally limited role of the state (Hallin and Mancini 2004: 11). Internet access and use are comparatively very high, and Facebook in particular is widely used (Enjolras et al. 2013). In 2008, 31 percent of the population in Norway was on Facebook at least once a week, while in 2013, the proportion who used Facebook daily was 67 percent (88 percent for people under thirty).² In Norway, most voters use Facebook, and about 20 percent of the population used Twitter in 2012 (Enjolras et al. 2013).

Although Internet penetration in the population is high, other characteristics might act as countervailing forces with regard to the use of media technology in electoral politics. In Norway, the political parties control the nomination process. The process is decentralized as the nominations are made by representative conventions organized by the constituency branches of the party organization. Moreover, the electoral system is based on proportional representation and a list system; there is only a theoretical possibility of influencing the candidate order. Consequently, when candidates are nominated, their election depends solely on the party vote. Based on these characteristics, it is no surprise that campaigning in Norway is typically party centered, centralized, and nationwide, and that the candidates campaign as part of the greater party campaign organization (Karlsen and Skogerbø 2015).

Data and Measures

Our data comprise a combination of the 2013 Norwegian Candidate Survey and Twitter data. The Norwegian Candidate Survey is a survey of all candidates running for election for any of the eight parties that obtained representation in the 2009 parliamentary election. We distributed the survey in November 2013 as a Web survey using tools provided by Questback. We obtained a response rate of 42 percent, leaving us with 850 candidates. The responses were about evenly distributed among parties. Candidates from one party typically constitute between 12 and 15 percent of the total, but Progress Party (FrP) candidates are somewhat underrepresented and only constitute 8 percent of all candidates. With regard to gender, 46 percent are female. This echoes the gender difference between men and women running for election for these parties.

The Norwegian Candidate Survey contains several questions on campaigning in general and the use of social media in the campaign. The extent of general individualized campaigning is measured with the following question: What was the primary aim of your campaign? Where would you place yourself on a scale from 0 to 10, where 0 means “to attract as much attention as possible for me as a candidate” and 10 means “to attract as much attention as possible for my party?” Norwegian candidates are extremely party centered: Their mean score is 8.89 (SD = 1.86). In the analysis, we will invert the scale so that high values indicate individualized campaigning.

To study the importance of social media in general, we used the following question asked to all candidates: On a scale from 1 to 5, where 1 indicates not important at all and 5 indicates very important, how important were the following communication channels for you in your campaign? The channels listed are as follows: national television, regional television, national newspapers, regional newspapers, local newspapers, national radio, local radio, personal Web site, Facebook, and Twitter. For this question, we also report the results from the 2009 candidate survey.

To tap into different ways of using social media during campaigns, we used the following question: On a scale from 1 to 5, where 1 indicates not important at all and 5 indicates very important, how important were social media for you and your campaign with regard to . . .? (see Table 1 for items). These questions were only asked of candidates who reported that they used either Facebook or Twitter (N = 684).

Table 1.

Proportion of Candidates (Who Use Social Media) Who Consider Social Media Important in Achieving Goals.

	Not Important			Very Important		M	N (100%)
	1	2	3	4	5	M	N (100%)
Sharing links	1	2	11	32	54	4.4	684
Creating involvement	0	1	9	36	54	4.4	681
Reaching out	1	7	16	26	51	4.2	684
Direct communication	2	6	15	34	42	4.1	684
Mobilizing supporters	3	7	18	36	37	4.0	679
Conveying your side on news stories	5	11	25	36	22	3.6	678
Increasing visibility to others in the party	9	16	27	30	19	3.4	679
Organizing the campaign	13	17	22	30	20	3.3	681
Appearing modern	15	15	26	27	18	3.2	678
Publicity in traditional media	13	16	23	29	18	3.2	677
Showing personal side	14	20	28	26	12	3.0	677

Note. Q: On a scale from 1 to 5, where 1 indicates not important at all and 5 indicates very important, how important were social media for you and your campaign with regard to . . .?

We asked all candidates who reported that they used Twitter to give us their Twitter account names and their authorization to retrieve and store their Twitter data. The question used was as follows: Were any of the following activities part of your campaign? And if yes, how important were they? All candidates except those who stated that Twitter was not part of their campaign were asked about their Twitter information. Of the 352 respondents who were asked about their Twitter accounts, 172 candidates gave us information on their accounts and permission to collect their tweets.³

Twitter data were gathered using two types of data collection strategies. First, we gathered data for all candidates who gave us authorization to retrieve and store their Twitter data through the Twitter API, and collected all the tweets of the 172 candidates since they became active on Twitter. We collected number of tweets and number of entities (mentions, retweets by others, and hashtags). These Twitter data were incorporated into our survey data set. Second, we collected all tweets mentioning the 172 candidates; this was done using a social media analytical tool developed by Crimson Hexagon, which enabled us to collect data from Twitter Firehose API and collect all tweets that matched our search criteria. This analytical tool allowed us to use a machine-learning classification algorithm developed by Hopkins and King (2010) and adapted to text and social media analysis within the social sciences. Overall, we identified and collected 29,559 tweets mentioning the candidates active on Twitter during the short election campaign (May to September 2013). These tweets, using a classification algorithm, were classified into six categories⁴:

Political conversation: The tweet mentions the candidate and is part of a political conversation or an answer from a Twitter user to the candidate.

Nonpolitical conversation: The tweet mentions the candidate and is part of a nonpolitical conversation or an answer from a Twitter user to the candidate.

Political comment with @mention: The tweet mentions the candidate and contains a political comment addressed to the candidate.

Retweet (RT) political message/comment: The tweet mentions the candidate and is a retweet of a political comment or message tweeted by the candidate.

Nonpolitical comment with @mention: The tweet mentions the candidate and contains a nonpolitical comment addressed to the candidate.

RT nonpolitical message/comment: The tweet mentions the candidate and is a retweet of a nonpolitical comment or message tweeted by the candidate.

The measurement of influence on social media in general and on Twitter in particular has become a field of research in computer science (Cha et al. 2010; Suh et al. 2010). The most immediate gauge of influence on Twitter is the number of followers. Other measures of influence focus not only on the number of followers but also on the attention received by a Twitter user based on the different modalities according to which the audience may engage with a tweet, such as retweeting, replying, and mentioning. For example, Cha et al. (2010) compared three measures of influence: in-degree (number followers), retweets (number of retweets containing the user’s name), and mentions (the degree of engagement with others). They found that the number of followers—a measure of popularity—is not related to other influence measures based on the degree of engagement with an audience. Retweets are driven by the tweet’s value (content), whereas mentions are driven by the user’s name value (popularity). They concluded that in-degree alone (the number of followers) is not the most adequate metric for measuring a Twitter user’s influence. Influence through indirect communication and cascades depends on this active minority of followers, whereas the vast majority of passive followers do not affect the user’s influence. In short, a high number of followers may indicate popularity but does not guarantee influence, which is best measured by the number of retweets and mentions. In this study, we use the number of @mention by other users to measure influence.

Empirical Analysis

The empirical analysis has three parts corresponding to the three research questions. First, we study candidates’ use of social media in terms of the extent and what they use it for. Second, we investigate the relationship between an individualized social media style and Twitter activity and influence. Third, we explore the candidates on Twitter who are most visible—the so-called Twitter influentials—and study the extent to which they stand out from other candidates.

Individualized Social Media Campaign Style

Social media are a popular campaign channel for parliamentary candidates in Norway. In 2013, more than 80 percent of the candidates surveyed used social media in their campaigns. In addition, the significance of Facebook has increased considerably (see Figure 1). Facebook is considered far more important in 2013 than social media were in 2009. This also reflects that more candidates used social media in 2013 than in 2009. Candidates who used social media in 2009 considered it important then as well (Karlsen 2011).

Figure 1.

Important campaign communication channels for the candidates: 2009 and 2013.

Why do candidates consider social media important? Earlier, we discussed how candidates might use social media to enhance their own candidacy and show their personal side. Social media could also be used for several other communicative aims. In Table 1, we present the proportion of candidates who consider social media important for achieving eleven different types of goals.

Almost all candidates consider social media important for creating involvement in the campaign, sharing links, reaching out to as many voters as possible, direct communication with voters, and mobilizing party supporters. Although the importance of social media in conveying the candidates’ personal side was considered the lowest, almost 40 percent of the respondents considered social media important in showing their personal side. If we consider the overwhelming consensus on parties being the focal point of communication, this is quite a high number.

We used principal component analysis to investigate the relationships between the eleven communicative social media aims. This is a method used to simplify relationships between variables by analyzing correlations between them, with the aim of identifying underlying dimensions that might explain the correlations (Foster 2006). The results are shown in Table 2.

Table 2.

Two Social Media Campaign Styles: Principal Component Analysis.

	I	II
Creating involvement	0.769	0.136
Mobilizing supporters	0.728	0.096
Reaching out	0.710	0.131
Direct communication	0.665	0.246
Sharing links	0.561	0.217
Organizing the campaign	0.559	0.196
Increasing visibility to others in the party	0.029	0.786
Showing personal side	0.118	0.784
Conveying own version of politics and events	0.306	0.616
Appearing modern	0.206	0.560
Publicity in traditional media	0.355	0.478
Eigenvalue (initial)	3.95	1.37

Note. N = 645. Rotation method: varimax with Kaiser normalization.

Two dimensions were identified using this approach. The first one resembles a party communication dimension. The most important items are creating involvement and mobilizing supporters, which are typical collective campaign practices in party-centered systems. The second dimension resembles an individualized communicative dimension. The two most important items are “increasing visibility to others in the party” and “showing their personal side.” These are both communicative practices with the aim of enhancing the individual candidate.

In Table 3, we report the results of a multivariate analysis studying the effects of individualized campaigning on individualized social media style. Social media style is measured as an additive index of the two defining items of dimension II from the factor analysis. Earlier research has shown that top candidates have a higher tendency to focus on themselves in the campaign (Karlsen and Skogerbø 2015). Moreover, candidates in some parties are more likely to focus on themselves. Hence, we also included position on the election list and political ideology (as a proxy for party affiliation) in the model, in addition to gender and age. In step I of the regression, we included gender, age, and individual campaigning.

Table 3.

The Effect of Individualized Campaigning, Position on the List, and Left-Right Political Ideology on the Perceived Importance of Social Media for Individual Purposes.

	I	II
Constant	4.61 (0.01)	3.12 (0.01)
Gender	−0.21 (0.18)	−0.24 (0.19)
Age (in years)	−0.01* (0.10)	−0.01 (0.10)
Individual campaigning	0.26*** (0.32)	0.21*** (0.05)
Position on the list		0.05*** (0.04)
Left-right		0.11*** (0.53)
R ²	.06	.09

Note. N = 625. Dependent variable: individualized social media style index (0–8, 8 = individualized style; M = 4.4). Individual campaigning (0–10, 10 = focus on own candidacy). Position on the list has an inverted scale (0–27, 27 = first place). Left-right is the standard self-placement question, with a scale from 0 (left) to 10 (right). B = coefficients; standard error (SE) in parentheses.

p < .10. **p < .05. ***p < .01.

The results reported in Table 3 reveal that an individualized campaign style has a significant independent effect on individualized social media style. According to the model, the individual campaigning variable can move a candidate almost three points on the individualized social media style index. The effect holds even when we control for position on the list and left-right ideology. However, both the position on the list and left-right ideology has independent significant effects on individualized social media style. Hence, candidates who are placed high on the list, who are positioned on the right of the political spectrum, and who have an individualized communicative style are more likely to use social media for individualized purposes.

In the next section, we combine the survey data with Twitter data to investigate the relationship between individualized social media style, Twitter activity, and Twitter influence.

The Effect of Individualized Campaigning on Twitter Activity and Popularity

Candidates use social media to a great extent, including for individualized purposes. But to what extent is an individualized campaign style associated with a higher probability of being active on Twitter? To investigate this, we used the combined survey and Twitter data set and studied whether an individualized campaign style, both in general and on social media, has an effect on Twitter activity as measured by the number of tweets. In Table 4, we present the results of a stepwise regression analysis. The stepwise approach will help us identify the specific effects of the independent variables.

Table 4.

The Effect of Individualized Social Media Style, Individualized Campaigning, Position on the List, and Left-Right Political Ideology on the Total Number of Tweets.

	Model I	Model II	Model III	Model IV
Constant	2,237.66*** (428.67)	2,268.47*** (563.09)	1,312.43** (653.92)	1,889.38*** (690.43)
Gender (male)	403.41** (199.30)	447.50** (206.35)	495.30** (202.80)	472.89** (202.15)
Age (in years)	−31.46*** (8.39)	−30.14*** (8.49)	−27.21*** (8.38)	−29.23*** (8.54)
Individualized social media style	72.02 (49.20)	80.86 (50.96)	53.36 (50.91)	71.32 (51.36)
Individual campaigning		27.48 (51.10)	38.896 (50.21)	36.70 (50.37)
Position on the list		2.96 (18.70)	5.051 (18.33)	17.60 (19.03)
Left-right		−55.02 (41.17)	−48.07 (40.39)	−42.33 (39.95)
Twitter important			237.622*** (87.41)	222.34** (87.09)
Television important				119.08 (91.66)
National newspaper important				−34.70 (96.03)
Local newspapers important				−234.49** (100.62)
R² (adjusted)	.10	.10	.14	.16

Note. Only candidates who reported their Twitter accounts are included in the analysis (N = 154). Dependent variable: number of tweets by candidates (0–3,200). B = coefficients; standard error (SE) in parentheses.

p < .10. **p < .05. ***p < .01.

Individualized social media style increases the number of tweets by seventy-two for each point on the index. Although this is quite a sizable effect, it is not significant. However, significance testing in regression analysis is about the likelihood of the effect in the model being found in the population. Hence, as we have the population of candidates we are interested in, we should also consider the insignificant results, but we should do so with caution. The standard error is quite high, which indicates that although the b coefficient indicates a substantial increase for each point, there is much variation between candidates. The only variables with significant effects are gender, age, Twitter, and local newspaper importance. Men tweet more than women, and young candidates tweet more than older candidates. As expected, the ones who consider Twitter important tweet more, but gender and age are still significant when we include Twitter importance in the model.

Individualized social media style seems to increase Twitter activity. But, as discussed above, activity is not necessarily related to visibility and influence on Twitter. In this article, we operationalize influence as @mention of the candidate by other Twitter users during the campaign. In Table 5, we report the results of a stepwise multivariate regression analysis that investigates the effect of individualized social media campaign style on influence on Twitter. We also include the number of tweets in the model to investigate if the candidates’ activity is related to influence. As in the previous analysis, we also include age, gender, list placement, and political ideology as control variables.

Table 5.

The Effect of Individualized Social Media Style, Individualized Campaigning, Position on the List, and Left-Right Political Ideology on Twitter Influence (@Mentions).

	Model I	Model II	Model III	Model IV
Constant	354.74 (253.18)	−278.22 (334.75)	−258.36 (383.35)	−312.02 (416.20)
Gender (male)	−1.06 (4.96)	4.67 (4.99)	4.64 (5.02)	2.76 (5.22)
Age (in years)	154.70 (117.71)	103.26 (118.29)	101.63 (119.65)	115.12 (121.04)
Individualized social media style	−28.99 (29.06)	−43.79 (29.00)	−43.24 (29.55)	−50.77 (30.39)
Individualized campaigning		15.88 (28.87)	15.58 (29.10)	15.91 (29.66)
Number of tweets		0.17*** (0.05)	0.17*** (0.05)	0.17*** (0.05)
Position on the list		13.94 (10.56)	13.89 (10.60)	13.08 (11.22)
Left-right		−14.88 (23.37)	−14.98 (23.47)	−12.37 (23.57)
Twitter important			−5.57 (51.81)	−16.41 (52.33)
Television important				12.60 (54.19)
National newspaper important				59.14 (56.47)
Local newspapers important				4.11 (60.24)
R² adjusted	.00	.08	.08	.07

Note. Only candidates who reported their Twitter accounts are included in the analysis (N = 154). Dependent variable: number of @mention (0–6,234; M = 242.44, SD = 679.3). B = coefficients; standard error (SE) in parentheses.

p < .10. **p < .05. ***p < .01.

Twitter activity is the only variable with a significant effect. Candidates who tweet frequently are more likely to be influential. Interestingly, although individualized social media campaign style had a positive effect on Twitter activity, its effect on Twitter influence, although not significant, was negative. Higher placed candidates are more inclined to get attention, even when we control for Twitter activity. This indicates that top candidates have a greater chance of visibility on Twitter compared with lower placed candidates. But the large standard error indicates a lot of variation between candidates; hence, this effect is not significant.

To examine these factors more closely, in what follows, we will investigate whether the most popular candidates on Twitter have anything in common that sets them apart from other candidates.

The Twitter Influentials

As shown by Figure 2, popularity is very unequally distributed. A few candidates received most of the attention whereas most of the candidates are not very often addressed by other users on Twitter. We compare the eight most addressed candidates with the rest of the candidates to get insights about the reasons associated with their success.

Figure 2.

Distribution of attention (number of @mention by other users) among candidates on Twitter.

The eight most addressed candidates were mentioned 19,628 times during the campaign (May–September 2013) and received 66 percent of all the candidates’ @mention. The most addressed candidate had 6,130 @mention, while the eight most addressed candidates had 1,334 on average. In what follows, we study whether these candidates distinguish themselves from the other candidates. We first compare them with others in terms of the categories identified in the data section.

The results reported in Figure 3 reveal two interesting tendencies. First, the most successful politicians on Twitter seem to stick more to political issues than their colleagues do. Thirty-six percent of all tweets mentioning the top eight were part of a political conversation, compared with 25 percent of all tweets mentioning all candidates. Furthermore, 21 percent of all tweets mentioning the top seven were political comments with a mention, compared with only 12 percent of all tweets mentioning the other candidates. Second, the Twitter influentials seem to embrace the possibilities for interactivity to a greater extent than others, such as in political conversation. Therefore, the eight political Twitter influentials have a distinguishable Twitter style in terms of content. Do they also distinguish themselves from others in terms of gender, age, position on the list, and communicative patterns?

Figure 3.

Comparison of types of visibility and influence in the political Twitter sphere between all candidates and the top seven candidates on Twitter.

As indicated by the analysis above, being a man is a success criterion on Twitter, and six of the seven Twitter influentials are men. Their average age is thirty-three years eight months, which makes them considerably younger than the average candidate at forty-four, and the average candidate on Twitter at thirty-nine. The seven Twitter influentials come from four parties: two from the socialist left, two from the Greens, two from the Liberals, and two from the Conservative Party. Their average list position is 5.7, which is considerably higher than the mean for all candidates at 8.1 and the mean for candidates on Twitter at 7.3. However, their positions vary; four of them were placed on top, while the others were placed seventh, fourteenth, and fifteenth. This indicates that success on Twitter is compatible with a strong party position but is by no means a requisite.

When it comes to communicative patterns, the top seven, indeed, consider Twitter their most important communication channel. Moreover, they also rate national television, national newspapers, and national radio more important than the average candidate (see Table 6). This mostly reflects their relatively high national status indicated by their list placement. In terms of social media communicative aims, they resemble other candidates but consider social media more important in conveying their personal side than the average candidate. In addition, they stand out when it comes to communicating directly with voters (see Table 7). They find the opportunities to organize the campaign via social media of little importance compared with others.

Table 6.

The Importance of Different Communication Channels for All Candidates and the Twitter Influential.

	All Candidates	Twitter Influentials	Difference
Twitter	2.3	3.9	1.6
National newspapers	2.5	3.7	1.2
Facebook	3.6	3.6	0.0
Local newspapers	4.1	3.6	−0.5
Regional newspapers	3.6	3.6	0.0
National radio	2.4	3.4	1.0
National television	2.4	3.3	0.9
Personal Web site	2.2	2.7	0.5
Regional television	2.8	2.3	−0.5
Local radio	2.9	2.3	−0.6

Note. Mean on the scale from 1–5.

Table 7.

The Importance of Social Media Communicative Aims for All Candidates and the Twitter Influential.

	All Candidates	Twitter Influentials	Difference
Direct communication	4.1	4.4	0.4
Reaching out	4.2	4.3	0.1
Sharing links	4.4	4.1	−0.2
Creating involvement	4.4	4.1	−0.3
Appearing modern	3.2	3.9	0.7
Conveying your side on news stories	3.6	3.9	0.3
Publicity in traditional media	3.2	3.8	0.6
Showing personal side	3.0	3.4	0.4
Mobilizing supporters	4.0	3.3	−0.7
Visible to others in the party	3.4	3.0	−0.4
Organizing the campaign	3.3	2.3	−1.0

Note. Mean on the scale from 1–5.

Conclusion

Social media offer candidates new campaign communication channels, and successful candidates use them in tandem with other platforms in the emerging hybrid media system. This influences power relations in political parties. In this article, we have shown that in the context of the contemporary Norwegian campaign, social media are now one of the most important communication tools for candidates in their campaigning efforts, and even in this party-centered environment, candidates emphasize the possibilities to convey their personal side in these media. We found that candidates’ use of social media can be divided into two main dimensions: a party-centered and an individualized social media style. Candidates who had a communicative aim of focusing on their own candidacy were more inclined to have an individualized style on social media as well. Increasing a candidate’s visibility in his or her own party is an essential part of the individualized social media campaign style.

The individualized social media style increased the likelihood of being active on Twitter. However, in general, the relationship between an individualized social media campaign style and Twitter influence was negative. Based on these results, candidates who use social media to focus on their own candidacy are not the most successful and influential candidates on Twitter. Hence, the distinction between activity and influence is essential. The profile of the Twitter influentials modified this picture somewhat. They found Twitter useful for showing their personal side. Moreover, the influentials used the interactive opportunities to a greater extent than others, and their Twitter conversations were more about politics. The influentials are younger, male, and relatively centrally placed in their parties. However, Twitter influentials do not constitute the absolute top politicians, who mostly consist of party leaders, figuring on national television every day in the election campaign.⁵

These results indicate that Twitter do influence power relations in party politics as social media provide new avenues for candidates to communicate with their constituencies and with the general public. Even though the increase in individualized campaigning seems modest, our results indicate that the candidates who gain influence in social media are those who are able to create a synergy between traditional media channels and social media. Candidates are not created equals on Twitter and those who are influential appear to have communicative and political skills enabling them to harness both the affordances of social media as well as to generate attention and visibility in the traditional media.

As emphasized initially, in Norway, the electoral systems offer few incentives for individualized campaigning. However, the lack of incentives for individualized campaigning is not necessarily a lack of incentives for using social media platforms. Whether a candidate considers a communication platform worth using is most likely also based on communication needs, and groups that are reachable through the platform. Twitter is an elite medium used by journalists and political elites, and, hence, for Norwegian politicians, this is an essential platform to reach key actors in party politics as well as opinion leaders in the electorate (Karlsen 2015; Vaccari and Valeriani 2015).

It is nevertheless likely that in systems with higher levels of individualized campaigning, candidates will utilize social media to create platforms to focus on their own candidacy to a greater extent than what we have found in Norway. However, we have shown that an individualized social media style is related to increased activity, but not to increased influence. We believe that this finding is not related to the Norwegian context but to the characteristics of the hybrid media system. In hybrid systems, success to some extent depends on skills and ability that allow politicians to master the game of political communication in both traditional and new media platforms, generating synergies between these communication channels. Hence, this influence is based on candidates’ digital hybrid competence, and should be found in most European countries where candidates do not have a professional campaign organization built around them (as in the United States). In Norway, as in most other West European countries, Twitter communication is based on candidates’ competence as professional assistance is mostly reserved for party leaders. Hence, although more candidates are likely to embrace an individualized social media style in systems with higher levels of individualized campaigning, influence will depend on candidates’ digital and political competence.

Overall, our findings confirm that social media is an essential and integrated part of the emerging hybrid political communication system used by politicians and show that this hybrid communicative structure can be used to gain influence in party politics. This echoes how, in the hybrid political communication system, traditional possibilities interact with new possibilities (see Chadwick 2013), and consequently, although power is not disrupted, it is in some ways transformed.

Footnotes

Acknowledgements

We would like to thank the editors and two anonymous reviewers for very helpful comments and suggestions.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Rune Karlsen is Research Professor at the Institute for Social Research, Oslo, Norway. He is the author of After the Mass Party. Continuity and Change in Political Parties and Representation in Norway (with Elin Allern and Knut Heidar). His research interests include political communication, campaigning, political parties, political representation and electoral research.

Bernard Enjolras is Research Professor and Deputy Director at the Institute for Social Research in Oslo. He is Director of the Center for Research on Civil Society and Voluntary Sector and leads the Freedom of Speech Foundation project “Status of the freedom of speech in Norway”. He is the coordinator of the EU-FP7 project Third Sector Impact. His research interests are related to different aspects of civil society and the third sector, public policies and governance issues, civic and political engagement, social media and digitalization, terror and trust, and freedom of speech.

References

Agranof

Robert

, ed. 1972. The New Style in Election Campaigns. Boston: Holbrook Press.

Allern

Elin H.

Heidar

Knut

Karlsen

Rune

. 2016. After the Mass Party. Lanham: Lexington.

Balmas

Meital

Rahat

Gideon

Sheafer

Tamir

Shenhav

Shaul R.

2014. “Two Routes to Personalized Politics: Centralized and Decentralized Personalization.” Party Politics 20 (1): 37–51.

Brox

Brian J.

Shaw

Daron R.

2006. “Political Parties, American Campaigns, and Effects on Outcome.” In Handbook of Party Politics, ed. Katz

Richard S.

Crotty

William J.

, 146-159. London: Sage.

Bruns

Axel

. 2012. “How Long Is a Tweet? Mapping Dynamic Conversation Networks on Twitter Using Gawk and Gephi.” Information, Communication & Society 15 (9): 1323–51.

Burgess

Jean

Bruns

Axel

. 2012. “(Not) the Twitter Election: The Dynamics of the #ausvotes Conversation in Relation to the Australian Media Ecology.” Journalism Practice 6 (3): 384–402.

Cha

Meeyoung

Haddadi

Hamed

Benevenuto

Fabricio

Gummad

Krishna P.

2010. “Measuring User Influence on Twitter: The Million Follower Fallacy.” In Proceedings of the 4th International AAAI Conference on Weblogs and Social Media. Washington, DC, 23–26. http://www.icwsm.org/2010/papers.shtml

Chadwick

Andrew

. 2013. The Hybrid Media System. Oxford: Oxford University Press.

Chadwick

Andrew

Dennis

James

Smith

Amy

. 2016. “Politics in the Age of Hybrid Media.” In The Routledge Companion to Social Media and Politics, ed. Axel

Bruns

Enli

Gunn

Skogerbø

Eli

Larsson

Anders

Christensen

Christian

7-22. New York: Routledge.

10.

Chi

Feng

Yang.

Nathan

2010. Twitter in Congress: Outreach vs. Transparency. MPRA Paper Series, 23597. Munich: University Library of Munich.

11.

Chi

Feng

Yang

Nathan

. 2011. “Twitter Adoption in Congress.” Review of Network Economics 10 (1): 1–20.

12.

Dalton

Russell J.

Wattenberg

Martin P.

2000. Parties without Partisans: Political Change in Advanced Industrial Democracies. Oxford: Oxford University Press.

13.

Enjolras

Bernard

Karlsen

Rune

Steen-Johnsen

Kari

Wollebæk

Dag

. 2013. Liker–liker ikke. Oslo: Cappelen Damm.

14.

Farrell

David

Webb

Paul

. 2000. “Political Parties as Campaign Organizations.” In Parties without Partisans, ed. Dalton

Russell J.

Wattenberg

Martin P.

, 102-128. Oxford: Oxford University Press.

15.

Foster

Jeremy J.

2006. “Factor Analysis.” In The Sage Dictionary of Social Research Methods, ed. Jupp

Victor

, 115-116. London: Sage.

16.

Gibson

Rachel

. 2015. “Party Change, Social Media and the Rise of ‘Citizen-Initiated’ Campaigning.” Party Politics 21 (2): 183–97.

17.

Gibson

Rachel

Ward

Stephen

. 2009. “Parties in the Digital Age—A Review Article.” Representation 45:87–100.

18.

Glassman

Matthew Eric

Straus

Jacob R.

Shogan

Colleen J.

2010. “Social Networking and Constituent Communications: Member Use of Twitter during a Two-Month Period in the 111th Congress.” Journal of Communication Research 2 (2–3): 219–33.

19.

Golbeck

Jennifer

Grimes

Justin M.

Rogers

Anthony

. 2010. “Twitter Use by the U.S. Congress.” Journal of the American Society for Information Science and Technology 61 (8): 1612–21.

20.

Graham

Todd

Jackson

Dan

Broersma.

Marcel

2014. “New Platform, Old Habits? Candidates’ Use of Twitter During the 2010 British and Dutch General Election Campaigns.” New Media & Society. Published electronically August 12, 2014. doi:10.1177/1461444814546728.

21.

Grant

Will J.

Moon

Brenda

Busby

Janie

. 2010. “Digital Dialogue? Australian Politicians’ Use of the Social Network Tool Twitter.” Australian Journal of Political Science 45 (4): 579–604.

22.

Hallin

Daniel C.

Mancini

Paolo

. 2004. Comparing Media Systems: Three Models of Media and Politics. Cambridge: Cambridge University Press.

23.

Hopkins

Daniel

King

Gary

. 2010. “A Method of Automated Nonparametric Content Analysis for Social Science.” American Journal of Political Science 54 (1): 229–47.

24.

Jackson

Nigel

Lilleker

Darren

. 2011. “Microblogging, Constituency Service and Impression Management: UK MPs and the Use of Twitter.” The Journal of Legislative Studies 17 (1): 86–105.

25.

Karlsen

Rune

. 2011. “A Platform for Individualized Campaigning? Social Media and Parliamentary Candidates in the Party-Centered Norwegian Campaign.” Policy & Internet 3 (4): 1–25.

26.

Karlsen

Rune

. 2013. “Obama’s Online Success and European Party Organizations: Adoption and Adaptation of US Online Practices in the Norwegian Labor Party.” Journal of Information Technology & Politics 10:158–70.

27.

Karlsen

Rune

. 2015. “Followers are Opinion Leaders: The Role of People in the Flow of Political Communication on and beyond Social Networking Sites.” European Journal of Communication 30:301–18.

28.

Karlsen

Rune

Skogerbø

Eli

. 2015. “Candidate Campaigning in Parliamentary Systems: Individualized vs. Localized Campaigning.” Party Politics 21:428–39.

29.

Katz

Richard S.

Mair

Peter

. 1995. “Changing Models of Party Organization and Party Democracy: The Emergence of the Cartel Party.” Party Politics 1:5–28.

30.

Kriesi

Hanspeter

. 2012. “Personalization of National Election Campaigns.” Party Politics 18:825–44.

31.

Larsson

Anders

Kalsnes

Bente

. 2015. “‘Of Course We Are on Facebook’—Use and Non-Use of Social Media among Swedish and Norwegian Politicians.” European Journal of Communication 29 (6): 653–67.

32.

Larsson

Anders

Moe

Hallvard

. 2012. “Studying Political Microblogging: Twitter Users in the 2010 Swedish Election Campaign.” New Media & Society 15 (5): 729–47.

33.

Moe

Hallvard

Larsson

Anders

. 2013. “Untangling a Complex Media System: A Comparative Study of Twitter-Linking Practices during Three Scandinavian Election Campaigns.” Information, Communication & Society 16 (5): 775–94.

34.

Plasser

Fritz

Plasser

Gunda

. 2002. Global Political Campaigning. London: Praeger.

35.

Poguntke

Thomas

Webb

Paul

. 2005. The Presidentialization of Democracy: A Study in Comparative Politics. Oxford: Oxford University Press.

36.

Skovsgaard

Morten

Dalen

Arjen van

. 2013. “Dodging the Gatekeepers? Social Media in the Campaign Mix during the 2011 Danish Elections.” Information, Communication & Society 16 (5): 737–56.

37.

Spierings

Niels

Jacobs

Kristof

. 2014. “Getting Personal? The Impact of Social Media on Preferential Voting.” Political Behavior 36:215–34.

38.

Suh

Hong

Pirolli

Chi

E. H.

2010. Want to Be Retweeted? Large Scale Analytics on Factors Impacting Retweet in Twitter Network. IEEE Second International Conference on Social Computing (SocialCom).

39.

Swanson

David

Mancini

Paolo

. 1996. Politics, Media, and Modern Democracy. London: Praeger.

40.

Vaccari

Christian

Valeriani

Augusto

. 2015. “Follow the Leader! Direct and Indirect Flows of Political Communication during the 2013 General Election Campaign.” New Media & Society 17 (7): 1025–1042.

41.

Wattenberg

Martin P.

1991. The Rise of Candidate-Centered Politics. Cambridge, MA: Harvard University Press.

42.

Zittel

Thomas

. 2009. “Lost in Technology? Political Parties and Online Campaigns of Constituency Candidates in Germany’s Mixed Member Electoral System.” Journal of Information Technology & Politics 6:298–311.

43.

Zittel

Thomas

Gschwend

Thomas

. 2008. “Individualized Constituency Campaign in Mixed-Member Electoral Systems: Candidates in the 2005 German Elections.” West European Politics 31:978–1003.