Blogging in the Shadow of Parties: Exploring Ideological Differences in Online Campaigning

Abstract

Blogging is an increasingly important practice in election campaigns, showing interesting variations across contexts. Recent research has shown that the adoption and use of blogs is strongly shaped by national institutional settings, that is, the different roles given to parties within political systems. However, intra-national differences in the practice of political blogging are yet to be explained. This article investigates the variation in usage of blogs in electoral campaigns in Sweden, a country characterized by strong political parties and a party-centered form of representative democracy. The central argument is that different parties utilize blogging in different ways. Just as blogging is shaped by how institutions support persons or parties, we propose that political blogging is shaped by party affiliation and ideological positions on individualism and collectivism. The empirical analysis, based on a survey among over 600 blogging politicians, confirms that ideological positions towards individualism and collectivism have a great impact on the uptake and usage of political blogs, portraying political blogging as a strongly ideologically situated practice of political communication.

Keywords:

• blogging
• political parties
• election campaigns
• political individualism

According to Castells (2001), societies based upon digital networks are qualitatively different from those in which power is centralized and communication flows are mainly linear. The creative and disruptive characteristics of networks are assumed to have profound consequences for organizations such as parties, which are now considered to derive their significance from a tension between network flows, on the one hand, and different strategies to maintain hierarchical control, on the other. This article sets out to explore this tension empirically, by examining the role of political blogging in election campaigning in Sweden.

There is little doubt that the Internet is capable of transforming the mode and content of mass communication, perhaps most notably in election campaigning. Presidential candidates in the United States, such as Bob Dole, Howard Dean, and, most recently, Barack Obama, have all made headlines for their creative use of the Internet to mobilize voters and generate support. Blogging, in particular, is recognized as a practice that allows candidates to engage in campaigns more directly and bypass parties to articulate individual opinions, thoughts, feelings, experiences, and identities (Albrecht, Lübcke, & Hartig-Perschke, 2007; Chadwick, 2007; Coleman, 2005; Wright, 2008). However, not all countries have nurtured the blogosphere in the same way. Empirical research has found that the adoption and use of political blogs is largely shaped by the institutional setting, that is, the different roles given to political parties (Anstead & Chadwick, 2009; Jackson, 2008; Zittel, 2003). In political systems where major party endorsement is still crucial to winning a seat, there is little incentive for candidates to openly champion divergent positions.

Sweden is an interesting setting in which to study the interplay and mutual shaping of these forces. This is where the shadow of party hierarchies is among the darkest in Europe. At the same time, Internet use is more developed here than in almost any other country. It is a country of partisans as well as pirates. And, perhaps, it is therefore a country in which we can find interesting intra-national differences. Thus far, most evidence suggests that it is resources (such as party size) and factors related to wider electoral dynamics (such as incumbency) that need to be taken into account when explaining variance within national settings (Gibson, 2004). More recently, however, a debate has started in the United States on a possible ideological bias in relation to online campaign communication (Benkler & Shaw, 2010; Karpf, 2008). Stemming from the observation that Democrats have demonstrated far greater ability in using the Internet to mobilize and engage supportive issue publics than Republicans, despite similar attempts to innovate online, researchers have raised the issue of whether or not the Web is a fundamentally liberal medium (cf. Karpf, 2009). From this point of departure, this article will explore how blogs are utilized by different parties in Sweden and lay out a more specific explanation for why ideology might matter. Just as blogging is shaped by how institutions support persons or parties, we propose, in short, that political blogging is shaped by party affiliation and ideological positions on individualism and collectivism.

The main empirical techniques used to map the development of online campaigning to date, link analysis and content analysis, are not the best choice of method when trying to grasp the intentions behind the practices of political blogging. Therefore, this article will draw upon a survey questionnaire of 600 blogging candidates in the run-up to the 2010 Swedish election. Approaching the role of ideology in this more direct way holds the potential to uncover unseen relationships and produce new knowledge. Few investigations have been conducted on the attitudes of political bloggers, and the studies that have been performed have focused on small samples of elite bloggers (cf. Coleman & Moss, 2008; Ekdale, Namkoong, Fung, & Perlmutter, 2010: Benkler & Shaw, 2010; Jackson, 2008; Trammell, Tarkowski, Hofmokl, & Sapp, 2006; Wright, 2008).

The article will examine the role of ideology in blogging with regard to three specific questions: (a) Does the blogosphere show an equal standing of left and right? (b) Do bloggers from the left and right approach political blogging differently? and (c) What is the role of ideology in explaining the network centrality of political blogs? In the concluding part of the article, the implications of the empirical evidence for understandings of online campaigning are discussed.

Theoretical Perspectives

Campaign Communication

According to most academic observers, the history of political campaigning includes three distinct phases (Norris, 2000). The first, the “premodern era,” preceding the introduction of television in political communication in the 1960s, was characterized by decentralized and staff-intensive campaign activities and high levels of party loyalty among voters (Gibson & Römmele, 2001, p. 34). In the wake of changes in the media landscape, campaigning evolved towards a second phase, the “modern era,” characterized by stronger centralization of campaign activities and a stronger focus on party leaders (Blumler & Kavanagh, 1999, p. 212; Gibson & Römmele, 2001; Karlsen, 2009). In recent decades, a new era of campaigning has arisen. The modernization of Western societies and the individualization of citizens have created voters who are less loyal to, identify less with, and less frequently organize within political parties. Formerly clear social structures firmly entwined with party choice were exchanged for more complex societies and mobile voters, increasing the need for individualized campaign communication. The development of individualized methods of campaign communication was supported by technological developments. Indeed, advancements in information and communication technologies (ICTs) made targeted campaign communication possible (Gibson & Römmele, 2001; Karlsen, 2009).

However, this so-called “postmodern era” of political campaigning is composed of a mixture of developments with uncertain outcomes. On the one hand, there are signs of a continuing professionalization of political campaigns, with power increasingly centralized in the hands of party elites. On the other hand, local campaign organizations seem to be experiencing a second spring (Norris, 2000), and local candidates have started to employ ICTs to run election campaigns, independent from their parties and the traditional media (Zittel & Gschwend, 2008). This ambiguity is highlighted by Zittel (2009), who offers two competing hypotheses regarding the Internet's impact on the structure of election campaigns—one “orthodox” and one “revisionist.” The former perceives the Internet as reinforcing the trend toward professionalized and centralized campaigns, allowing parties to target and mobilize groups of voters in more efficient and direct ways. Content is primarily about party program and party image, even though information on these topics may often be distributed “with a personal face” (2009, p. 305). According to the revisionist view, it is not all that certain that parties are in the driver's seat. Instead, it is argued that the Internet is a means for individual candidates to run candidate-centered campaigns, independent from their own parties, attracting as much attention as possible to themselves rather than to their parties.

Blogs are touted as having the potential to reinvigorate political communication by facilitating decentralization, individualization, and interactivity, in line with Zittel's revisionist view. At the same time, an unmistakable tension has arisen between orthodox and revisionist tendencies. According to Wright (2008, p. 53), “a battle is being fought.” On the one hand there are candidates trying to make use of the individualistic blog platform; on the other hand there are parties trying to carry out professionalized campaigns, controlling their messages and presenting uniform fronts. Blogging candidates are, thus, described as being implicated in a tug of war between two different cultures.

Politicians live in a world of certainty and tribal loyalty which is at odds with the blogging ethos of open-mindedness and knowledge-sharing. As long as politicians are expected to be never in doubt and ever faithful to catechismic party messages, their blogging efforts are always likely to look more like simulation than authentic self-expression. (Coleman, 2005, p. 276)

Cross-National and Intra-National Variations

In explanations of how blogging candidates handle the intricate situation described above, factors related to the institutional setting have been most prominent. One such factor is the electoral system. It is widely acknowledged that electoral systems allowing for a strong individual vote tend to have weaker party cohesion than countries relying more heavily on party votes, since competition within parties is strengthened (Katz, 2001). In a system where major party endorsement is crucial to winning a seat, there is every incentive to follow the party line. If the prospects for reelection depend more on the members themselves than the party they represent, members have greater incentive to give priority to their own opinions and preferred issues. However, the effect of individual votes on party cohesion is also said to depend on the rules for converting votes into seats. The key distinction here is between plurality systems and proportional representation. According to Rae (1971), the latter is, by nature, more party-oriented than the former, making the effect of individual voting less noticeable. Hence, the electoral system has been considered a decisive factor in understanding campaign strategies in different countries. When voters are given the option of voting for individual candidates and this choice makes a difference, candidates perceive an incentive to individualize their campaigns. As several scholars of online campaigning have successfully shown, this also seems to create incentives to use the Internet for this very purpose (Anstead & Chadwick, 2009; Ward & Cahill, 2007; Zittel, 2003).

In comparison to studies of institutional differences, relatively little attention has been given to intra-national variations in campaign communication and the fact that a candidate-versus-party emphasis within parties may affect the extent to which candidates distance themselves from their parties. This is a bothersome research gap considering that several empirical observations now point in such a direction (Hindman, 2008; Karpf, 2008; Benkler & Shaw, 2010) and also because previous research emphasizes party affiliation as one of the most important explanations for members’ support for different principles of representation (Esaiasson & Holmberg, 1996; Brothén & Gilljam, 2006; Gilljam, Karlsson, & Sundell, 2010a).

In a recent study, Hindman (2008) found that a left-right gap is indeed a feature of the contemporary online political landscape in the United States: Liberals dominate the audience for politics online. Through a blogosphere tracking tool, Karpf (2008) demonstrates the existence of a similar gap between left and right wing top-level bloggers in the United States. He shows that conservatives have continually failed to develop parallel organizations to those that have risen to prominence in the American left. Furthermore, by coding top-level political blogs, Benkler and Shaw (2010) reveal significant variations in content between the left and the right along several important dimensions: Sites on the left adopt more participatory technical platforms, are comprised of significantly fewer sole-authored sites, include user blogs, and more often use blogs as platforms for mobilization as well as discursive production. In discussing what would account for these differences, ideology is mentioned as one possible candidate, but more nuanced research is called for to understand the findings more fully.

On the ground that members carry a mindset of preexisting party goals and norms, previous research has, at times, explained different campaign strategies by ideological differences (Gibson, 2001; Kavanagh, 1996). In relation to different forms of representation, Holmberg (1974) specifically points out collectivism-individualism as the crucial dimension. For leftist parties, he argues, the guiding principle has been party solidarity, while right wing parties have been influenced by more liberal and individualistic traditions of representation. This dimension has also been important when explaining different approaches towards reform proposals that support more candidate-centered electoral systems (Möller, 1999; Nielsen, 2007). Again, parties’ positions and arguments have been proven to reflect their ideological views on collectivism and individualism.

Individualism-Collectivism and Political Blogging

How online campaigning is affected by attitudes towards individualism-collectivism is yet to be explored. However, the increasing inclination towards individualized online political communication leads us to think that a finer grained analysis may reveal that ideology will carry some weight. On the one hand, digital technologies enable loosely tied relations between individuals, causes, and organizations, while often giving individuals considerable autonomy in making choices about how, when, where, and with whom to affiliate and act (Segerberg & Bennett, 2011). On the other hand, digital technologies are enacted—made sense of, designed, and used—through the mediation of existing organizational and institutional arrangement with their own logic (Fountain, 2001). Parties and ideologies may thus stifle the effects of factors promoting change, and it is therefore reasonable to assume that parties and candidates that already embrace individualistic values are more willing than others to move in such a direction. The central argument of this article is thus that blogs are utilized in different ways by different parties. Just as blogging is shaped by how institutions support persons or parties, we propose that political blogging is shaped by party affiliation and ideological positions on individualism and collectivism.

In discussing individualism and collectivism, there is an important distinction to be made between actor-related individualization on the one hand and content-related individualization on the other (Lengauer & Winder, 2011; Van Aelst, Sheafer, & Stanyer, 2012; Vilegenthart, Boomgaarden, & Boumans, 2011). Political communication is individualized in an actor-related sense when the communicating actor is an individual as opposed to a group or a party. As such, blogs are in themselves individualistic media, and the practice of political blogging is an act of individualized communication. When it comes to content-related individualization, however, blogs can be individualized to varying degrees. As Lengauer and Winder (2011) argue, political positions presented on blogs can be of a collectivistic as well as an individualistic nature. Political communication defined as individualized in a content-related way may correspond to a collective position (i.e., a party program) but must be communicated as the position of an individual actor. It should be presented as “my position” (the individual's position) instead of “our position” (the party's position) (Lengauer & Winder, 2011, p. 13).

Blogging is thus an individualized form of political communication that could be set up to channel individualistic or collectivistic content. This flexible characteristic of blogs encourages us to move beyond an either-or perspective in order to analyze different degrees of individual and collective identification among bloggers from different parties. In relation to the actor-related aspect of individualization, we investigate the diffusion of political blogging among politicians from the left and the right of the ideological spectrum, studying the selective attention paid by candidates to new information technologies. Does the practice of blogging, like other trends of individualization in Swedish politics, have a clear left-right dimension? Next, addressing the content-related aspects of individualization and collectivism, we explore attitudes and motives related to blogging among candidates from the left and the right. This analysis is conducted in order to investigate possible differences between bloggers from different parties regarding the focus of blog content (focus on party versus focus on person) and usage of blogs as platforms for personal campaigning. Finally, we investigate to what extent ideology matters for the impact gained by candidates who blog. This analysis may shed light on what forms of blogging are supported within different party contexts in Swedish politics.

The Swedish Political Context

The Swedish political context is characterized by strong political parties with a central position in the representative democratic system. Parties have traditionally taken a central role in organizing political representation and electoral campaigns; all candidates are nominated by the political parties, and all representatives are organized in party groups. Thus, local and national party organizations are traditionally most influential in the organization of electoral campaigns in Sweden.

Although political parties in Sweden, as in most Western democracies, have experienced weakening public support, there are few signs that the power of political parties in political assemblies is weakening. During the last 40 years, Swedish parties have lost the lion's share of their members, voters have become increasingly mobile between elections (Holmberg & Oscarsson, 2004, p. 5), and party identification among citizens has strongly decayed (Oscarsson & Holmberg, 2009, p. 11). Still, the level of party loyalty among elected representatives has steadily increased during the same period. In 2006, almost half of Swedish local councilors said that they would choose to adhere to party views in a conflict between the (known) views of voters, their own views, and the views of their party (Gilljam et al., 2010a). The current political climate, thus, resembles the picture painted by Katz and Mair in the middle of the 1990s: “The parties are at once stronger, but also more remote; at once more in control, but also less powerful; at once more privileged, but also less legitimate” (Katz & Mair, 1994, p. 19).

The Party System

Of the countries that were democratized in the early 20th century, Sweden has had one of the most stable party systems (Pierre & Widfeldt, 1995). Besides the five parties that comprise the classic Swedish “five-party system,” only four other parties have, on any occasion, held seats in the national parliament. During the 2006–2010 parliamentary term, the national parliament included seven parties that are commonly divided into two blocs: a “left” bloc, containing the Left party, the Social Democrats, and the Greens, and a “right” bloc, containing the agrarian-based Centre Party, the Liberals, the Moderate party, and the Christian Democrats. Bloc identity has, at times, been unstable, but during this term the left and right blocs have been more organizationally and ideologically coherent than perhaps ever before. For the purpose of this article, it is also important to clarify that the ideological landscape differs between Sweden and, for instance, the United States, where most studies of ideology and political blogging stem from. Socialism is dominating on the left side, most so in the Left party and least so in the Green party. In Sweden the liberals are to be found on the right side of the spectrum, with a mix of conservatism in the Moderate party and the Christian Democrats. The highest level of individualism among Swedish politicians is thus found within the liberal right, while a more collectivist tradition is dominant within the socialist left. It is also worth noting that the right bloc in Swedish politics holds a majority of the seats in the national parliament and forms a coalition government.

Electoral System

In Sweden, a proportional voting system with closed party lists has been utilized in elections since the introduction of full democracy in the 1920s. In an attempt to revitalize Swedish democracy by strengthening candidates’ incentives for seeking personal mandates, a preferential vote was introduced through constitutional reform in the mid-1990s (Holmberg & Möller, 1999). The reform was a compromise between the Liberals and Moderates, who supported a stronger individual vote, on the one hand, and the Social Democrats and Left party, who strongly opposed individual voting, on the other (Möller, 1999, p. 271). The result was an optional preferential vote, allowing voters to make a preferential choice for one candidate from their party's list. If a candidate on the list received enough votes from his or her constituency to reach a preset limit of 5% of all preferential votes for that party, ¹ he or she was put at the top of the party list and got the first mandate won by the party in that constituency. The Swedish system for individual votes has been characterized as a weak system of preferential voting (Karvonen, 2004) that gives candidates increased, but still limited, incentives to conduct personal campaigns. Swedish voters’ preferential voting activity has been moderate and, over time, has decreased (Oscarsson & Holmberg, 2008). Their attitudes towards the preferential vote are polarized, with half of citizens supporting the existing system and the other half in favor of changing it (Nielsen, 2007, p. 24). As a whole, the addition of a preferential vote has been judged not to challenge the central position of political parties in Swedish democracy (Möller, 1999).

Party opinions on the preferential vote are also strongly polarized. While political representatives from the Left party and the Social Democrats resist a strengthened preferential voting system and the Greens are only moderately in favor of it, all of the parties in the right bloc are strongly in favor of such a reform (Gilljam, Karlsson, & Sundell, 2010b, p. 30). Skeptical politicians on the left have made several arguments for limiting the influence of preferential voting; it was thought to Americanize Swedish democracy, threaten the existence of political parties, give campaign financing a decisive role in elections, and undermine intra-party democracy as well as party cohesion (Holmberg & Oscarsson, 2004, p. 178). Increased opportunities for public control and accountability, as well as greater opportunity to strengthen the relationship between voters and representatives and increase activity among both voters and candidates in election campaigns, were arguments put forward by proponents of a strong preferential vote, especially politicians from the Moderate party and the Liberals (The Committee on the Constitution, 2003; Nielsen, 2007, pp. 9–16). All in all, this indicates a clear ideological divide within the Swedish system on the issue of preferential voting.

The Internet and Individual Campaigns

Taking the description above as a point of departure, the likelihood of finding individualized campaigning in Sweden should be fairly low, but varying between parties. Online campaigning would be likely to be collectivistic and party-centered on the left while more individualistic among the parties on the right, where support for an individualization of the electoral system is stronger. However, it has been argued that the diffusion of the Internet may have some significance when it comes to campaign strategies (Norris, 2000), and Sweden is a country where Internet access is widely diffused. Already in 2010, over 85% of the Swedish population had access to the Internet (Findahl, 2010). Blogging politicians are thus not only socialized into party cultures and ideologies with divergent positions on political individualism-collectivism, but they are also becoming increasingly engaged in an online culture in which individual-oriented communication is dominating. Hence, there are conflicting forces surrounding political campaign strategies of Swedish politicians, making this an interesting and critical case for investigation.

Methods and Measurements

This study is based on a Web survey sent to the authors of all blogs listed on the political blog-ranking site “Politometern” (www.politometern.se) for which e-mail addresses could be found. The survey was answered by 730 political bloggers. The 604 who were politicians from the seven parties represented in the Swedish national parliament during the parliamentary term 2006–2010 were included in the study. ² Included respondents either held political office at the time of the survey (the study includes members of the European Parliament, members of the national parliament, local councilors, and local board members) or were candidates in the 2010 elections. The survey was carried out during the 2 months following the 2010 Swedish election. Individual survey responses were linked to a network centrality index, calculating each blog's hierarchical position in the blogosphere based on the number of incoming links to that blog, and were compiled and presented on the Politometern Web site. The empirical material accumulated gives a broad overview of the Swedish political blogosphere and makes it possible to analyze bloggers’ background, strategies, and behavior in connection to their blogging as well as to their position in the Swedish blogosphere. This study is primarily based on quantitative analyses of bloggers’ survey responses and positions in the network centrality index. The variables used are presented in the following sections.

Dependent Variables

Two different sets of dependent variables are analyzed to address different research questions. First, we investigate respondents’ blogging strategies, addressing the division between individualism and collectivism. Second, we investigate the political impact of blogging.

Individualism-Collectivism

Three dimensions of individualism-collectivism are examined. First, we investigate bloggers’ attitudes toward preferential voting and separate supporters of a stronger preferential voting system (support stronger preferential vote) from other respondents. Second, respondents’ motives for using blogging as an instrument in personal election campaigns are investigated by dividing respondents into two categories: those who report that using blogging as an instrument in a personal campaign was a strong motive for blogging during the 2010 election campaign (strong personal campaign motif) and other bloggers. The third dimension of individualism-collectivism is measured by separating bloggers with an individual focus (blog focusing on person) from bloggers with a party focus (blog focusing on party). The variables employed to examine level of individualism and collectivism are used as both independent and dependent variables in different analyses.

Network Centrality

The political impact of blogging is a complex and elusive concept. To conduct a valid measurement of impact, we analyze a network centrality index based on the number of incoming links to each blog, creating a measurement of the position of each blog and blogger in the hierarchy of the Swedish political blogosphere. The data were gathered using the search engine Google blog search (GBS). GBS is a widely used search engine for finding content in the blogosphere that offers statistics on incoming links or “backlinks” to blogs (Thelwall & Hasler, 2007). GBS offers statistics on the number of blogs (defined as any site with an RSS or Atom feed that is searchable by the search engine) that, at a given point in time, have links to any content on a specific blog. This number is used to determine the network centrality ranking position of that blog. The blog with the most incoming links is ranked as number one, and thereafter blogs are listed in descending order. We transformed this ranking into an index on a 100-point scale where a higher number corresponds to a more central position in the Swedish political blogosphere. The data used for this study were the rankings of Swedish political blogs 2 weeks before the 2010 elections.

Independent Variables

The independent variables analyzed in this study can be divided into four categories: blogging activity and experience, party-related variables, party affiliation and ideological orientation, and social characteristics.

Blog-Related Variables

Level of experience and activity in the blogosphere are investigated using three variables. First, the length of time that respondents have blogged is studied using a variable that separates bloggers with at least 2 years of experience (long-time bloggers) from other bloggers (short-time bloggers). Second, we operationalize level of blogging activity with a variable that reflects the frequency with which bloggers write blog posts, dividing bloggers into two categories: those who update their blogs at least three to five times a week (frequent bloggers) and those who write posts less frequently. We also investigate level of activity in the blogosphere using two categories for respondents who read others’ blogs: those who follow 10 blogs or more (heavy blog readers) and those who follow fewer than 10 blogs.

Party-Related Variables

Level of party activity is operationally defined as the number of party meetings respondents reported having attended during the year preceding the survey (10 party meetings or more/less than 10 party meetings). We also analyze bloggers’ positions within their parties by separating incumbents, who held a political position in the last parliamentary term, from challengers, who were candidates in the 2010 election and had not held a political position during the last term.

Party Affiliation and Ideological Orientation

Bloggers’ party affiliation is represented through party dummy variables that are included in all analyses. The respondents’ ideological self-positioning was measured by two survey questions on a 11-point scale from 0 (most left) to 10 (most right). In order to create measurements of the ideological coherence between political bloggers and other politicians, these values were compared to other surveys using the same question for ideological self-positioning (see Gilljam et al., 2010b). Ideological similarity measures were created by comparing respondents self-positioning on the left-right scale to other politicians from the same party at the same political level (local, regional, or national ³ ). By measuring political affiliation as well as ideological self-positioning, we are able to relate the influence of party position as well as ideological orientation on the uptake, usage, and impact of political blogging. Also, opportunities are created for measuring the influence of ideological differences among bloggers from the same party.

Social Characteristics

Several social characteristics are analyzed: gender, education (specifically, whether or not respondents have a postsecondary education), age (specifically, whether respondents were born before or after 1970), and geographical location. The variable for geographical location separates respondents in the Stockholm metropolitan area from respondents in the rest of the country.

Empirical Analysis

Political Blogging Within the Left and Right

The distribution of political, social, and blogging variables reveals many similarities between the left and right blogosphere in the run-up to the 2010 Swedish election, but there are some interesting differences as well. Perhaps most important, there seems to be an ideological bias towards the right. The data show that bloggers from the right are somewhat overrepresented among respondents (58% vs. 42%) when compared to electoral turnout (53% ⁴ vs. 47%), despite the potential effect of being an incumbent and not as strategically motivated to innovate as the left (cf. Karpf, 2008).

Moreover, a general right orientation of the Swedish political blogosphere becomes evident when comparing the left-right self-positioning of blogging politicians to Swedish politicians in general. The measurement we used, an 11-point scale from 0 (most left-oriented) to 10 (most right-oriented), has been used in several studies of Swedish politicians and voters. It was thus possible to compare the bloggers’ ideological self-positioning to the positions of Swedish politicians from the same parties and on the same political level, based on data presented by Gilljam et al. (2010b, p. 19). The result is striking. A vast majority of bloggers position themselves to the right of the average value of political representatives from their own party at the same political level. ⁵ The pattern is consistent across parties, with the exception of the Left party. Also, a complementary analysis of this intra-party dimension was carried out, comparing the self-positioning among bloggers to the ideological position that the bloggers themselves assign their respective parties. Here as well, a majority of bloggers were found to be to the right of their parties. The mean value difference is statistically significant among all bloggers, within the left respective to the right party coalitions as well as within each political party (see Table 1). Thus, the blogosphere during the 2010 Swedish election cycle was ideological biased towards the right. Blogging more often attracted candidates to the right of the ideological spectrum, both between and within parties.

Table 1 Descriptive statistics for bloggers (percentages), by party and party coalition

	Left party	Social Democrats	Green party	Centre party	Liberal party	Moderate party	Christian Democrats	Left	Right
	(n = 44)	(n = 131)	(n = 77)	(n = 117)	(n = 97)	(n = 92)	(n = 46)	(n = 251)	(n = 351)	Difference	All
Blogging characteristics
Frequent blogger	36	30	18	27	23	40	29	28	29	+1	29
Long-time blogger (>2 years)	67	73	53	33	44	57	63	66	46	−20***	55
Follows 10 or more political blogs	12	24	15	16	18	21	17	18	18	0	18
Party activity and party position
Incumbent position (elected)	57	62	64	62	61	71	70	62	65	+3	64
Attended 10 or more party meetings last year	79	79	86	83	82	79	76	81	81	0	81
Ideological orientation: right of party	35	73	74	68	70	78	72	67	72	+5*	70
Social characteristics
Male	61	61	52	57	64	66	65	58	62	+4	61
High education	68	64	76	62	74	71	78	69	70	+1	70
Metropolitan area	11	20	21	12	26	21	13	19	18	−1	19
Age < 40 years	66	48	55	38	52	35	44	52	42	−10**	46
Note. Chi-square tests were performed on differences between left and right.
^† p < .1, *p < .05, **p < .01, ***p < .001.

Table 1 furthermore shows that highly active and experienced bloggers make up a large share of the respondents. A majority had blogged for more than 2 years before the 2010 Swedish elections, and almost a third wrote daily blog posts during the election campaigns. The respondents were also active as readers of political blogs: About a fifth followed 10 or more political blogs, and over 40% followed between four and 10 (data not presented in the table). However, the analyses reveal some party and party group differences regarding blogging habits. Relatively few political bloggers from the Greens (18%) and the Liberal party (23%) were frequent bloggers (posted daily during the last month of the campaign), compared to parties such as the Moderate party (40%) and the Left party (36%). Furthermore, a significantly larger share of bloggers on the left had blogged for a long time (more than 2 years): 66% compared to only 46% among the right. This sizeable difference between the two party groups, combined with the overrepresentation of bloggers from the right, indicates that a large number of political blogs were set up by right wing politicians soon before the campaign period. The parties as well as the party groups were similar, however, when it comes to the reading of other political blogs; about a fifth of the bloggers from each party group stated that they followed 10 or more political blogs regularly.

The analysis indicates that candidates from both the left and the right were highly involved in the party sphere, as well as the blogosphere. More than 80% of respondents on the left and the right had participated in 10 or more party meetings during the last year, and about two thirds of blogging politicians held incumbent positions. The remaining third were challenging candidates in the 2010 elections. Some party differences are apparent when it comes to incumbency, as 70% of the bloggers from the Moderate party were elected representatives prior to the 2010 election, compared to only 57% of the bloggers from the Left party. On the whole, however, the party groups were largely similar in this regard (62% incumbent among the left compared to 65% among the right).

Innovations such as blogs usually spread to an increasingly large group only after a period of time. When a phenomenon is new, it is limited to a relatively small group. From a diffusion perspective (Schmidt, 1986), we would therefore expect that some groups—including young, highly educated candidates in metropolitan areas—are particularly likely to use blogs. We find that blogging candidates are predominantly male, representing six out of 10 respondents. This difference roughly reflects the gender difference among local and regional Swedish politicians in general (42% women), but is somewhat higher than the gender differences among national parliamentarians (47% women) (Persson & Öhrvall, 2008; Wängnerud, 2010). The share of female political bloggers is slightly higher among the left-oriented than the right-oriented parties. As we would expect, the age structure among bloggers is different from that of politicians in general. Young politicians are overrepresented among bloggers, but there is nonetheless a fairly high number of bloggers over the age of 40. There is a significant difference in age structure between the two party groups, as left-oriented bloggers are more often under 40. No similar difference is apparent among left and right politicians in general (Persson & Öhrvall, 2008, p. 45). Finally, the geographical distribution of blogging candidates shows a strong presence of bloggers from the Stockholm metropolitan area. Every fifth blogger is located in the capital area. This centralization of the blogosphere near the capital roughly reflects the division of the Swedish population between Stockholm and the rest of Sweden. Hence, the political blogosphere has a clear foothold in the capital area but is not overrepresented there in relation to the population.

How and Why Do Candidates Blog?

Let us turn to the question of how and why political candidates blog. A key finding in our survey (see Table 2) is that blogging politicians on the right are much more individualized than those on the left, who are more strongly tied to the collective identities of their parties. The multivariate logistic regression analyses show that ideological self-positioning on the left-right scale significantly influences the likelihood that bloggers (a) support a stronger preferential voting system, (b) use their blog to support an individual election campaign, and (c) focus the content of their blog on themselves rather than their party (see Models 1, 3, and 5 in Table 2). This ideological effect is significant even when controls are made for a wide range of variables, including the blogger's party activity, political position, blogging habits, and social characteristics.

Table 2 Binary logistic regression models explaining individualism among political bloggers

	Support stronger personal vote		Personal campaign motive		Focus blog content on person
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Blogging characteristics
Frequent blogger	.757	.735	1.853**	2.125**	.917	.877
	(.218)	(.224)	(.206)	(.223)	(.187)	(.191)
Long-time blogger	1.032	1.185	.661 ^†	.913	1.531*	1.717*
	(.229)	(.244)	(.216)	(.235)	(.200)	(.212)
Heavy blog reader	1.635 ^†	1.791 ^†	.731	.906	.760	.791
	(.299)	(.305)	(.278)	(.294)	(.251)	(.255)
Party activity and party position
Active party member	1.247	1.247	6.110***	6.462***	.771	.728
	(.305)	(.313)	(.333)	(.350)	(.259)	(.263)
Incumbent position (elected)	.510**	.475	1.091	1.042	.947	.942
	(.250)	(.257)	(.232)	(.247)	(.213)	(.216)
Ideology and party affiliation
Ideological self-positioning (left-right)	1.611***	1.387***	1.384***	.999	1.080*	1.006
	(.049)	(.085)	(.043)	(.081)	(.036)	(.072)
Left party		.281 ^†		.127**		.644
		(.738)		(.722)		(.491)
Social Democrat		.685		.284**		.682
		(.397)		(.410)		(.352)
Centre party		1.458		3.716**		.922
		(.418)		(.406)		(.382)
Liberal party		2.468*		4.221**		1.961*
		(.428)		(.422)		(.393)
Moderate party		1.345		1.932		1.186
		(.530)		(.494)		(.469)
Christian Democrat		2.371		5.253***		.722
		(.550)		(.548)		(.493)
N	502	502	518	518	495	495
Cox and Snell R ^2	.278	.297	.224	.300	.033	.055
Nagelkerke R ^2	.371	.395	.300	.401	.045	.073
Note. All models include additional controls for respondents’ age, gender, level of education, and geographical location (metropolitan/rural area). Standard errors are presented within parentheses.
^† p < .1, *p < .05, **p < .01, ***p < .001.

When party affiliation is added to the model, the effects of ideological position either weaken (see Model 2) or become nonsignificant (see Models 4 and 6). In one of the models explaining bloggers’ attitudes towards strengthening the Swedish preferential voting system, we still find a significant effect of ideological self-positioning. This would mean that there are ideological differences among the bloggers within the same party that matter in response to this variable. Bloggers who take an ideological position further to the right are more likely to support a stronger preferential voting system than bloggers from the same party who position themselves further to the left. For the two other models, no suchideological intra-party variations are found. Instead, the results indicate that the variation among bloggers exclusively can be described as variation between parties.

Is this party effect of an ideological nature? The coefficients for each party indicate that it probably is. A pattern emerges showing that bloggers from the parties on the right are more individualistic than bloggers from the left parties. Even though all coefficients are not significant, the parties’ lineup is almost perfectly in accordance with their left-right orientations. The Social Democrats and the Left party are less individualistic than the reference category, the Greens, while the parties on the right are more individualistic than the Greens. The only exception to this rule is the coefficient for the Christian Democrats in the last model investigating the existence of a personal focus of blog content. The Christian Democratic bloggers are less individualized, in this regard, than the Green party bloggers but more individualistic than both Social Democratic and Left party bloggers. Bloggers from the Liberal party are by far the most individualized, being more than twice as likely as the Green party bloggers to support a stronger preferential vote, more than four times as likely to pursue blogging in order to support a personal election campaign, and almost twice as likely to focus on themselves rather than their party in their blogs.

All in all, these findings confirm that the blogosphere plays different roles within different parties. Bloggers from the parties on the right appear to embrace candidate-centered blogging, while the left gives the impression of having a more collective agenda. The variation is better explained by party affiliation than by the bloggers’ individual ideological positions, but the lineup nonetheless speaks of an ideological effect.

Table 2 uncovers a number of other interesting relations that deserve to be highlighted. Bloggers in incumbent positions do, just as often as challengers, have personal campaign motives for blogging and do as often have a personal blog focus. These results indicate that politicians’ blogging activities are not primarily guided by strategic concerns. It has been argued that “storming the castle is more fun” (Henke, 2008) and that challengers are in more need of alternative channels for campaigning (Karpf, 2008), but this does not seem to result in more individualistic and campaign-oriented blogging.

Considering that blogging candidates are situated within the blogosphere as well as the party sphere, the depth of their involvement in these spheres might influence their attitudes and strategies. Along the lines of cultural theory (Bale, 1997), less party involvement and more blog involvement could be expected to increase the probability of individualized campaigning, and vice versa. To some degree the evidence points in such a direction, but the results are rather complex. When it comes to involvement in the blogosphere, the data indicate that high activity in reading blogs is positively associated with support for a stronger personal vote, while activity in writing blog posts increases the likelihood of a personal campaign motive for blogging. Also, long-time bloggers are more likely to have a personal focus in their blogging. These analyses indicate that politicians with a high level of involvement in the blogosphere are more individualistic than other bloggers, a result that supports the often-made claim that blogging in itself is an individualized practice. However, in opposition to this argumentation, high party activity is strongly positively correlated with personal campaign motives. In fact, whether or not bloggers are active party members is the individual factor that has the most explanatory power when it comes to explaining personal campaign motives among political bloggers. On the whole, therefore, the data do not indicate that high involvement in the party sphere stands in the way of individualistic and campaign-oriented blogging.

Explaining the Network Centrality of Political Blogs

The analysis above reveals interesting differences in how candidates with different ideological orientations and party affiliations use their blogs. But what do these differences amount to when it comes to political impact? To investigate this question, we turn our attention to the network centrality index measuring the impact of Swedish political blogs through the number of incoming links to the blog (described in the methods section). The analyses are conducted through two linear regression models explaining centrality in the Swedish political blogosphere. In the first model all factors aside from party affiliation are included, and in the second model controls for party affiliation are added to the analysis.

One important result in Table 3 is that the level of experience and activity in blogging strongly influences the network centrality of political blogs. Activity and experience as a blogger are of far greater importance for gaining a central position in the political blogosphere than other political resources such as an incumbent position and active participation in party organizations. The results thus emphasize the potential of blogs to offer cheap and effective means of political communication to political underdogs that may help in “leveling the political playing field” (Gibson, Margolis, Resnick, & Ward, 2003, p. 50). Moreover, the results show that ideological positioning influences the level of network centrality. This relationship follows the same pattern as earlier analyses, reflecting, once again, the ideological bias of the Swedish political blogosphere to the right. The further to the right the bloggers position themselves, the higher level of impact they enjoy. But even though this may indicate a potential for change in a “revisionist” direction, pretty strong evidence points in the opposite direction.

Table 3 Linear regression models explaining the network centrality of political blogs

	Model 1	Model 2
Blogging characteristics
Frequent blogger	.202***	.218***
	(2.757)	(2.624)
Long-time blogger	.314***	.264***
	(2.659)	(2.592)
Heavy blog reader	.260***	.229***
	(3.243)	(3.074)
Party activity and political position
Active party member	−.048	−.030
	(3.557)	(3.362)
Incumbent position	.047	.058
(elected)	(2.85)	(2.697)
Individualism (attitudes and blogging practices)
Strong personal campaign motif	.043	.063
	(2.797)	(2.728)
Blog focusing on person	−.016	.004
	(2.600)	(2.486)
Support stronger personal vote	.053	.075
	(2.914)	(2.779)
Ideological orientation and party affiliation
Ideological self-positioning (left-right)	−.047	.286**
	(.557)	(.926)
Left party		.141*
		(5.966)
Social Democrat		.307***
		(4.299)
Centre party		−.142*
		(4.931)
Liberal party		.014
		(4.905)
Moderate party		−.206
		(6.051)
Christian Democrat		−.032
		(6.037)
N	353	353
R ^2	.392	.474
Adjusted R ^2	.369	.444
Note. Both models include additional controls for respondents’ age, gender, level of education, and geographical location (metropolitan/rural area). The Green party is used as the reference category in the second model. Standard errors are presented within parentheses.
^† p < .1, *p < .05, **p < .01, ***p < .001.

First, Table 3 shows that individualism among bloggers is unrelated to network centrality. If individualism had been found to be a predictor of blogging success, this would be a strong motivating factor for bloggers to run more candidate-centered campaigns in the future. However, in neither of the models, with or without controls for party affiliation, do significant relationships emerge.

Second, ideology only shows significance in the models where controls are made for party affiliation. This would mean that bloggers who stand to the right do not have more central positions within the political blogosphere in general, but only in relation to other bloggers from the same party. In line with previous studies, this result indicates that the Swedish political blogosphere is highly polarized, divided into a number of party-specific blogospheres. Readers usually seek out blogs that are written and populated by people from their own party (Adamic & Glance, 2005; Lawrence, Sides, & Farrell, 2010; Hargittai, Gallo, & Kane, 2008; Perlmutter, 2008; Farrell & Drezner, 2008; Nahon, Hemsley, Walker, & Hussain, 2011). Thus, bloggers with similar ideological positions but divergent party affiliations do not compete for the same audience. Instead, they compete with other bloggers from the same party, and only in this competition do bloggers with an ideological position further to the right seem to have a more central position.

Third, Table 3 shows that bloggers from the parties on the left on average have a more central position in the political blogosphere than do bloggers from the parties on the right. Most successful are bloggers from the Social Democrats and thereafter bloggers from the Left party. But identifying which political camp is dominating the political blogosphere is far from being clear cut. In terms of diffusion of political blogging, as well as the usage of this medium in an individualistic candidate-centered manner, the bloggers and parties on the right are clearly ahead. But when it comes to network centrality, the pattern is reversed. An exclusive investigation of the top blogs illustrates even more clearly the strong position of Social Democrats. Close to half (48.3%) of the bloggers who are represented among the top 100 political blogs in Sweden are Social Democrats, while less than a fifth (17.2%) are bloggers from the Moderate party. Although these parties won a roughly equal voting share in the 2010 national election (Social democrats: 30.66%; Moderate party: 30.06%), they have a strongly differentiated standing in the political blogosphere.

The key for understanding these different results is perhaps to differentiate between candidates’ blog incentives and parties’ incentives for responding to candidates’ blogs. In the party-centered system of Sweden, a challenger position does not seem to motivate candidates to utilize a blogging strategy that diverges from the party ideology. Politicians seem to partake in individualistic forms of blogging only when they are ideologically motivated to do so. But for a party in opposition, such as the Left party and the Social Democrats in the 2010 election campaign, there are strong strategic incentives to search for and promote new ideas, regardless of party ideology. Drawing on Karpf (2008), it is more probable that party networks that lost the last election are more actively searching for and promoting new ideas. Since the network centrality index primarily measures support of political blogs (through the number of links to the blogs), it makes sense that the bloggers of the oppositional left wing in general are higher ranked than the bloggers of the incumbent right. The divergent network centrality between the left and the right might thus reflect different incentives for supporting political blogging efforts between majority and oppositional parties. In this light, the gap between Social Democrats (the biggest losers in the 2006 election) and the Moderate party (the greatest winner in the 2006 election) becomes logical.

Conclusions

In the popular debate on online campaigning, deterministic perspectives are at the center. The new opportunity structure of the Internet is assumed to change everything or nothing at all. Institutional contexts make all the difference—or none. According to Wright (2011), this deterministic framework often distorts how researchers make sense of their empirical findings by creating undue expectations. In the face of all of the hyping of technology, there is danger of pessimistic conclusions when the empirical reality does not match an overly optimistic mindset. Considering the institutional context, with parties of unparalleled strength and an election system with limited incentives for candidate-centered campaigns, we should thus be careful not to have overly high expectations. It is important to note that a fairly large number of candidates blogged during the 2010 electoral campaign and that the blogging uptake has been quite broad. More than the usual number of pioneers have been blogging, and their blogs have often been used in candidate-centered campaigns. Furthermore, activity and experience as a blogger are of greater importance for gaining a central position in the blogosphere than other political capital such as an incumbent position and active participation in party organizations. Hence, the medium appears to be important.

The main conclusion to be drawn from this study is nonetheless that the uptake, usage, and impact of blogging among Swedish politicians, to a considerable degree, are mediated by ideologies and parties. First, blogs seem to be more attractive to politicians who stand to the right, both among and within parties. A possible explanation for this ideological asymmetry within the Swedish blogosphere lies in common understandings of blogging as an individualistic practice and, therefore, one that is more attractive to politicians with liberal ideological convictions.

Second, our findings strongly suggest that the blogosphere plays different roles within different parties. The parties on the right appear to embrace candidate-centered blogging, while the left gives the impression of having a more collective agenda. This duality is also logical, considering that left-oriented politicians are socialized into collectivistic party organizations that have traditionally opposed reforms for a stronger preferential voting system and that historically have been characterized by stronger party loyalty (Holmberg, 1974, pp. 163, 211; Esaiasson & Holmberg, 1996, p. 53; Gilljam et al., 2010b; Nielsen, 2007, p. 10). Within these parties, broad support for a stronger preferential vote, wide uptake of personal campaign strategies, and individually focused blogs would be nothing less than a landslide. The influence of preexisting party cultures and ideological positions confirms that political blogging is an ideologically situated practice of political communication.

Third, the network centrality of blogs varies according to ideological differences between candidates of the same party. Politicians who stand to the right have a more substantial impact within their party. However, the role that online campaigning plays in determining the outcomes of Swedish political competition is complex. In general, bloggers from the parties on the left, among which individualistic attitudes and blogging practices are scarcer, have a more substantial impact than political bloggers from the parties on the right. This party difference prevails in analyses where a broad set of statistical controls are included, and is given even more emphasis when analyzing bloggers on the top 100 list.

Thus, in relation to Zittel's two competing hypotheses regarding the Internet's impact on the structure of election campaigns, the “orthodox” and the “revisionist,” we end up with a mixture of tendencies. On the one hand, individualistic campaign strategies are fairly well diffused among bloggers, especially within the liberal right. In sheer numbers, the political blogosphere is also characterized by a clear ideological bias towards the right. On the other hand, the analysis shows that individualism is not related to the network centrality of bloggers and that left-right ideology only matters within parties. Across parties, there is instead a partisan advantage for the left wing parties. Though the bloggers from the right wing parties have attempted to move in a revisionist direction, they have not (yet) been able to mobilize and engage supportive publics to the same extent as the more orthodox bloggers from the left wing parties. This could be explained by the incentives guiding partisan uptake of candidate bloggers. Even though the challenger position did not seem to affect the activity of candidates, parties in opposition may very well have had stronger incentives to support blogging efforts within party networks.

In summary, this article demonstrates that ideology matters for the diffusion and usage of blogs. In Sweden, blogging clearly takes place in the shadow of parties. As long as the electoral system favors party over person, we cannot expect to see a general trend towards more individualized forms of campaigning. However, organizations and technologies—such as blogs—are always connected reciprocally (Fountain, 2001). Each one has effects on the other, and the results are often unpredictable. Therefore, a continuous spread and maturity of online campaigning may, in time, challenge existing party structures and drive Swedish candidates and parties from orthodoxy towards revisionism.

Notes

© Joachim Åström and Martin Karlsson

1. In national elections, the limit was reduced from 8% to 5% before the 2010 elections.

2. The survey was sent to 1,800 e-mail addresses and received 730 responses, amounting to a response frequency of 41%. There are strong indications that response frequency is underestimated, as several recipients of the survey reported that they were either no longer active as political bloggers or had never been political bloggers. It is not possible to calculate the response frequency among bloggers, as no data exist regarding the share of all Swedish political bloggers listed on the Politometern Web site who were incumbent politicians or candidates in the 2010 election.

3. Members of the European Parliament were excluded from this analysis due to a lack of corresponding data regarding other Swedish MEPs.

4. In these statistics, all other parties as well as all blank votes were excluded.

5. For further description of this measurement, see the methods section.