reaching out or reaching in?

26
This article was downloaded by: [Texas A & M International University] On: 08 October 2014, At: 02:18 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Information, Communication & Society Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/rics20 REACHING OUT OR REACHING IN? Marc Hooghe a & Sara Vissers b a Department of Political Science , KU Leuven, Park Street 45, B-3000, Leuven, Belgium b Department of Political Science , KU Leuven, Park Street 45, B-3000, Leuven, Belgium E-mail: Published online: 13 Aug 2009. To cite this article: Marc Hooghe & Sara Vissers (2009) REACHING OUT OR REACHING IN?, Information, Communication & Society, 12:5, 691-714, DOI: 10.1080/13691180802483062 To link to this article: http://dx.doi.org/10.1080/13691180802483062 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub- licensing, systematic supply, or distribution in any form to anyone is expressly

Upload: sara

Post on 12-Feb-2017

225 views

Category:

Documents


0 download

TRANSCRIPT

This article was downloaded by: [Texas A & M International University]On: 08 October 2014, At: 02:18Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

Information, Communication &SocietyPublication details, including instructions for authorsand subscription information:http://www.tandfonline.com/loi/rics20

REACHING OUT OR REACHING IN?Marc Hooghe a & Sara Vissers ba Department of Political Science , KU Leuven, ParkStreet 45, B-3000, Leuven, Belgiumb Department of Political Science , KU Leuven, ParkStreet 45, B-3000, Leuven, Belgium E-mail:Published online: 13 Aug 2009.

To cite this article: Marc Hooghe & Sara Vissers (2009) REACHING OUT ORREACHING IN?, Information, Communication & Society, 12:5, 691-714, DOI:10.1080/13691180802483062

To link to this article: http://dx.doi.org/10.1080/13691180802483062

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all theinformation (the “Content”) contained in the publications on our platform.However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, orsuitability for any purpose of the Content. Any opinions and views expressedin this publication are the opinions and views of the authors, and are not theviews of or endorsed by Taylor & Francis. The accuracy of the Content shouldnot be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions,claims, proceedings, demands, costs, expenses, damages, and other liabilitieswhatsoever or howsoever caused arising directly or indirectly in connectionwith, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes.Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly

forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

Marc Hooghe & Sara Vissers

REACHING OUT OR REACHING IN?

The use of party websites during the 2006

electoral campaign in Belgium

While various authors have suggested that introduction of the Internet will lead tosubstantial changes in political communication patterns, others have argued that tra-ditional inequalities and patterns will only be reinforced. Websites of political partiesplay a specific role in this debate. On the one hand, they can be used by voters seekinginformation on various political parties. On the other hand, they might reach onlythose that are already ‘converted’ and have sufficient means to gather political infor-mation by other media. During the campaign for the 2006 local elections in Belgium,we conducted an online survey among visitors to party websites (n ¼ 3,224). Theanalysis shows that a substantial majority of online visitors is male and highly edu-cated. About 60 per cent of all respondents had already decided to vote for the partywhose website they had just visited, but another 40 per cent clearly compared parties’positions and candidates. Our tentative conclusion therefore is that during electoralcampaigns party websites will be used mainly by those who are already convincedabout the programme and the candidates of that specific party.

Keywords party websites; Internet; election campaigns; Belgium;digital divide

Introduction

All over the world, political parties are increasingly relying on new information andcommunication technologies (ICTs) to reach out to potential voters during electoralcampaigns. While just a few years ago, cyber campaigns were still situated at thefringes of the electoral process, it is expected that their impact on electoralcampaigns will continue to grow in the years ahead (Davis et al. 2008; Gibsonet al. 2004; Norris 2003; Pedersen & Saglie 2005; Small 2007; Strandberg 2006;Ward & Lusoli 2005). Competing claims have been made, however, about thepotential electoral impact of the Internet (Graber 2004). The more optimisticexpectation is that the Internet can function as a forum for debate and expression.

Information, Communication & Society Vol. 12, No. 5, August 2009, pp. 691–714

ISSN 1369-118X print/ISSN 1468-4462 online # 2009 Taylor & Francis

http://www.tandf.co.uk/journals DOI: 10.1080/13691180802483062

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

The Internet allows for a more intensive form of interaction than traditional media,and therefore it is expected that the Internet will lead to a more open and hencemore democratic exchange of political debate (Shah et al. 2001; Sinekopova2006). Other authors, however, are more sceptical about the alleged blessings ofthe Internet for political debate. It is expected that Internet users will not usenew information technologies to seek out information from varying sources,which indeed would lead to more pluralism in society and in information exchange.Rather, it is feared that new communication technologies will be used primarily toseek out like-minded people and like-minded information channels. Citizens areexpected to seek information that is in line with their pre-existing political interests,beliefs and orientation and to avoid information opposing or challenging their views(Norris 2001; Sunstein 2001). The Internet as an interactive medium inducesselectivity through filtering systems and search engines, which makes ‘chanceencounters’ less likely compared with conventional forms of mass media such asbroadcast television and (hard copy) newspapers. According to this pessimisticassumption, the Internet will lead to segmentation and group polarization (Sunstein2001). Citizens will no longer be confronted with information that is at odds withtheir own ideological preferences, as they will restrict their information-seekingbehaviour to like-minded news sources.

The research on the use of ICT during electoral campaigns thus far has beenconcentrated mostly on content analysis (Norris 2003; Ward & Gibson 2003).Partly, this focus is a result of the fact that online content is readily availablefor research and analysis. If we want to investigate the Sunstein thesis on thealleged polarizing effects of the Internet, however, content is not the idealtesting ground, as the thesis addresses information-seeking behaviour at theindividual level. It is to be expected that every political party will use itswebsite to promote its own programme and its own candidates (Lusoli 2005).To some extent, segmentation can be operationalized by measuring how manylinks these party websites contain to other competing parties. But even in thatcase, ICT will only function as a bridging mechanism if the website visitorsactually use these links to get connected to opinions that they would not getinto contact with otherwise. To determine the impact of party websites, it iscrucial, therefore, not to focus solely on content, but also on the behaviouralpatterns of website visitors, a step that hardly has been taken that often in thiskind of research (see, however, Boogers & Voerman 2003; Gibson et al. 2005;Norris 2003).

In this article, we want to investigate the empirical validity of the competingclaims about the impact of Internet on the electoral process, using the onlinepresence of Belgian political parties as a test case. More specifically, we want toascertain how Internet users process the information being offered on partywebsites. Do citizens ‘shop around’, investigating the information being offeredby various parties, even by parties that they would not read about otherwise? Ordo they specialize and seek information on just the single party they identify

6 9 2 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

with? To put it differently: can party websites be used to reach out to new groupsamong the population, or are they mainly a method to preach to the converted?

To ascertain whether political party websites ‘reach out’ or ‘reach in’, weconducted a web-based survey among the visitors of the websites of Belgian pol-itical parties during the campaign for the local elections of October 2006. Of the13 political parties in Belgium, 11 agreed to put a banner on their website,resulting in 3,224 complete responses. This ‘Internet and Political Parties’(I&PP) survey, therefore, provides us with an indication about the backgroundcharacteristics, the motivations and the information-seeking behaviour of visitorsto party websites during an electoral campaign.

In the remainder of this article, we first review the theoretical frameworkunderlying our research questions and the hypotheses that can be developedfrom this literature. Subsequently, we provide information on data andmethods before proceeding with the analysis of the motivations and character-istics of visitors to party websites.

New media and politics

The statement that new ICTs have a strong impact on political communicationpatterns will not come as a surprise: there is an intense ongoing academic debateabout the political consequences of Internet use (Bimber 2001; Graber 2004). Onthe one hand, we can distinguish supporters of the mobilization theory and, on theother hand, authors in favour of the reinforcement theory (Norris 2001). Themobilization theory propounds optimism towards the opportunities of theWorld Wide Web to affect citizens’ political activity. This theoretical approachassumes that the Internet has the potential to inform, mobilize and engage citizenswho are currently not involved with politics. Four main arguments are expressedin favour of the mobilization theory (Norris 2001; Strandberg 2006). The firstargument states that the Internet provides abundant opportunities for politicalengagement. Second, Internet lowers the costs of informing oneself and learningabout politics and public affairs (Bimber 2000). Third, the wealth of informationpresented on the Internet provides citizens with the opportunity to become betterinformed about politics as they can acquire all the resources necessary to expresstheir views and to participate in public life. And finally, the Internet is an interac-tive medium and it has the potential to strengthen the relation between citizens,the political world and intermediary organizations. Research suggests that theInternet is becoming an important communication tool for in particular, youngpeople, the more affluent and the higher educated (Boogers & Voerman 2003;Gibson et al. 2005; Norris 2003).

Other authors tend to follow the reinforcement theory (Davis 1999; Margolis& Resnick 2000). According to this thesis, the Internet does not have the capacityto engage and mobilize citizens in the realm of politics. Proponents of the

R E A C H I N G O U T O R R E A C H I N G I N ? 6 9 3

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

reinforcement thesis often claim they can rely on sound empirical research. First,Internet access is unevenly divided, both across and within societies. Moreover,socioeconomic inequalities can also be found in the use of the Internet: even ifall groups within societies would have access to Internet, it is clear there willbe profound social differences in the way the Internet will be used to retrieveinformation (Van Dijk 2005; Warschauer 2003). Second, it has been arguedthat online politics will only attract citizens who are already interested andengaged in offline politics. Thus, reinforcement based on motivation and interestis conceivable and perhaps even unavoidable: if citizens are not motivated to payattention to politics, they will refrain from doing so, no matter what kind ofinformation is being offered and no matter what kind of medium is being used.When we examine the motivation-based reinforcement theory more in depth,further arguments appear. The first and most important argument is the factthat the Internet, in comparison with more traditional mass media, is aninteractive medium. It requires some activity from its users who have to seekout information in a more active manner. Whereas in traditional forms of massmedia such as television broadcasts and hard copy newspapers citizens aremore likely to be exposed to political content without much effort, this is notthe case for Internet. Because one has to rely on search engines and all kinds offiltering systems the Internet induces selectivity. As Lazarsfeld et al. (1965,p. 42) have expressed it: ‘The more interested people are in the election, themore opinions they have on political issues, the more actively they participatein a campaign, and the more they expose themselves to campaigns propaganda.’Therefore, even if Internet access would be evenly divided among different socialgroups, it would not change their interest in politics.

Selective exposure and group polarization

The reinforcement thesis mainly claims that those who are already interested inpolitics will use the Internet for political purposes. We can, however, also takethe argument a step further. What interests us is not just the extent of politicalinterest and engagement, but also the ideological background of the informationsources that online users are exposed to. Research on the political use of massmedia suggests that people are more likely to expose themselves to contentcongruent with their views and opinions, a phenomenon that is called motivated(partisan) selectivity (Chaiken & Stangor 1987; Sears & Freedman 1967). Wewould like to broaden our understanding of the motivation-based reinforcementthesis by highlighting the phenomenon of group polarization.

Given the self-selective exposure of Internet information, many scholars andpoliticians fear that new information technologies will have an adverse effect ondemocratic debate. Sunstein (2001, p. 49) argues that new technologies, andmore particularly the Internet, are ‘dramatically increasing people’s ability tohear echoes of their own voices and to wall themselves off from others’. He

6 9 4 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

expects that with regard to entertainment and information, actors will show atendency to select sources that do not challenge their pre-existing viewpoints.This can lead to a fragmentation of opinion and to group polarization (Lee2007). Group polarization refers to the process ‘after deliberation, [in which]people are likely to move towards a more extreme point in the direction towhich the group’s members were originally inclined’ (Sunstein 2001, p. 65). Ifdifferent groups in society do not see and hear quite different point of views,mutual understanding might be difficult and it might be increasingly hard forsociety as a whole to tackle common problems (Mutz 2006; Nadel 2002).

Given these ongoing theoretical debates about the political and social impactof the Internet, our main research questions are the following:

1. Are political party websites primarily used by those who are already politi-cally interested or can they also reach a new audience currently not involvedin politics?

2. Are political party websites primarily visited by party members, partisans,citizens with the same party identification (motivated partisan selectivity) orcan we detect a pattern of differentiation? Do citizens surf to differentparty sites or do they stick to the political party they prefer in the first place?

Research on party websites

Previous research on the role of party websites in election campaigns has focusedprimarily on the content of party websites and the online presence and Internetuse among the parties and candidates (Davis et al. 2008; Strandberg 2008).Research on the use of party websites by the electorate has been less extensive.Norris (2003) conducted one of the earliest analyses on the use of party websitesfrom a bottom-up perspective. Her research showed that political party websitesare being accessed mainly by young people, by men, the well educated and themore affluent middle class. Political party websites were also much more successfulin the northern European countries than in southern or central Europe. Further-more, party websites tend to attract voters who are already highly interested inpolitics and engaged in politics in real life. Norris (2003) notes, however, thatyoung people are eager users of political information on the Internet, while thisage group is hardly being reached by more traditional forms of political communi-cation. The findings clearly support the reinforcement thesis, with one exception:websites are able to attract a younger audience, an age group that is almost usuallyabsent in party politics (Lusoli & Ward 2004).

Results of other survey research conducted in The Netherlands (Boogers &Voerman 2003), the UK (Lusoli et al. 2006; Ward et al. 2003), Denmark(Hansen et al. 2005) and Australia (Gibson & McAllister 2006) confirm theNorris findings. For the UK, too, survey results show a strong overrepresentationof males and highly educated citizens (Lusoli & Ward 2005). Survey results also

R E A C H I N G O U T O R R E A C H I N G I N ? 6 9 5

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

suggest that party website visitors are not at all interested in interactive forms ofcommunication, but that they mainly want to receive top-down information.Seeking information on the party platform and the candidates remains the mostimportant reason for visiting party websites. Despite the high expectations thatwebsites and mailing lists would enhance direct and interactive communicationbetween citizens and political decision-makers, party website visitors seem tobe less interested in the interactive features that are present there. Forty-twoper cent of the online party users in the Danish 2005 campaign reported thatthe most important reason for their visiting party sites was to clarify general pol-itical questions, followed by information on single issues and information about thecandidates (Hansen et al. 2005). Lusoli et al. (2006) even found that in 2005 Britishcitizens are less in favour of the more interactive features on their MP’s sites than insimilar research conducted in 2001.

Boogers and Voerman (2003) found that different categories of websitevisitors report different reasons to visit the website. Party supporters primarilyuse party sites to seek out information on the election campaign and the partyorganization. Visitors with another party affiliation or vote intention showmore interest for information on the candidates and the party platform,whereas voters who have not decided yet primarily seek out informationabout party policies.

Turning to the question of the electoral impact of visiting party websites, theempirical findings tend to point towards different directions. Some authorssuggest that online campaigns have the potential to influence voter turnout(Gibson & McAllister 2006; Ward et al. 2003), whereas others find little orno impact at all on voting behaviour (Hansen et al. 2005).

Most research conducted in this area employs general public opinion surveys.This source of information has the advantage that it is representative, but the down-side is that, given the relatively small number of visitors to party websites, theanalysis is performed on rather small subsamples. Especially for more sophisticatedmultivariate models, the number of website visitors in a general population surveyis usually too small. Therefore, Boogers and Voerman (2003) resorted to a differ-ent sampling technique by developing an online self-selection survey among partywebsite visitors. Respondents are recruited to participate in the survey at themoment they actually visit the party website. Given the specific character of ourresearch question, we decided to implement the Boogers and Voerman design inthe I&PP survey.

Data and methods

To answer our research questions, we need an accurate understanding of thebehavioural patterns and the background characteristics of respondents whoactually visit party websites. Our research population consists of the online

6 9 6 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

visitors of the Belgian political party websites during the electoral campaign for thelocal elections of 8 October 2006. It has to be remembered that the local, municipallevel remains very important in Belgian politics, and that therefore the national partyheadquarters and party leaders are heavily involved in this campaign (Geys 2006).This also implies that interest in the campaign is rather high among the population.It also needs to be remembered that voting is compulsory in Belgium for local elec-tions too. The result is that turnout for local elections is well over 90 per cent, whichis far higher than in most other European countries. It is safe to conclude, therefore,that in Belgium, local elections are just as salient as national elections, contrary towhat is more customary in other European countries.

Because of the local character of the elections, we decided to select not justthe national party websites, but – if possible – also the websites of the localchapters of the parties. At the time of the campaign, 13 political parties wererepresented in a national or regional parliament in Belgium. Eleven partiesagreed to participate in this study, and only two, Mouvement Reformateur (theFrench-speaking liberals) and Front National (French-speaking far-right party),refused to co-operate in the I&PP survey. This led to the inclusion of eightFlemish parties and three French-speaking parties (Table 1). Four politicalparties, i.e. SP.a, CD&V, VLD and Groen!, were able to diffuse the banner ontheir national as well as on (almost all) their local party sites. Vlaams Belangrefused to distribute the banner on the homepage of their websites, though thisparty agreed to distribute the banner via an electronic newsletter, where theyexplained that the survey examines the visitors of the website of VlaamsBelang. Groen! too diffused the banner simultaneously on their national andlocal websites and in their e-newsletter. Later on it turned out that e-newslettersare far more effective in attracting respondents than a mere banner on the website,with the result that a large part of the initial respondents were recruited throughthese e-mails from Groen! and Vlaams Belang. The software we employed in theI&PP survey, however, allowed us to detect the link through which respondentswere recruited, so it was possible to exclude the respondents that were invitedby means of an e-mail alert. For most analyses, therefore, we will restrictourselves to respondents that actually were recruited on the party websiteitself. It has to be remembered that this is a self-selected survey, since websitevisitors themselves decided whether they would join the survey or not. Inorder to have an additional control on the representativeness of the results, theI&PP survey will be compared with a recent representative general populationsurvey conducted by the Flemish regional government, and which concludedvarious questions on (political) Internet use (SCV survey). The SCV (Sociaal-culturele verschuivingen or social-cultural change) survey was conducted in2005 and it can be considered as representative for the adult population of theautonomous region of Flanders, Belgium (n ¼ 1,524).1

Given the fact that our research population consists of the online visitors topolitical party websites, we conducted – following the example of the Dutch

R E A C H I N G O U T O R R E A C H I N G I N ? 6 9 7

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

TABLE 1 Websites participating in the survey.

name language website Ideology

unique number of

visitors per day,

September 2006a

completed

surveys

Political party websites

SP.a Dutch s-p-a.be Socialist 4,604 369

Spirit Dutch spirit.be Progressive nationalist 1,060 136

CD&V Dutch cdenv.be Christian Democrat 1,600 354

N-VA Dutch n-va.be Conservative nationalists 2,003 274

VLD Dutch vld.be Liberals – 253

Vivant Dutch and French vivant.be Liberal – 30

Groen! Dutch groen.be Greens 993 299

PS French ps.be Socialist 6,000 203

CdH French lecdh.be Christian Democrat – 82

Ecolo French ecolo.be Greens 810 303

Not from political party websites

Extreme Right Dutch Mailinglist newsletter Extreme Right 724

Greens Greens

Vlaams Belang Dutch Website candidates Extreme Right 157

Various Dutch and French Search engines, blogs, portal sites 39

Total 3224

Note: Unique visitors are counted only once no matter how many times they visit the site (per day). Unique visitors are identified with their IP number.aWebmasters of these parties, interviews September 2006 (election campaign).

69

8IN

FO

RM

AT

ION

,C

OM

MU

NIC

AT

ION

&S

OC

IET

Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

Internet survey by Boogers and Voerman (2003) – a self-selected online survey.Therefore, we asked the webmasters of the political parties to place the surveybanner on (at least) the homepage of the national party website during the weeksprior to the local elections on 8 October 2006. The survey banner also ended upon websites of some local party chapters and candidates as well as on some blogsand portal sites. Via the survey button visitors of the websites were invited toparticipate in the survey.

The I&PP survey started on 1 September and ended on 9 October 2006.Since we wanted to focus on the election campaign, we closed the survey theday after the elections.2 During the research period 4,218 respondents filledin the survey, and among them 3,224 completed the questionnaire. Thedropout rate in our survey (starting the questionnaire but not completing it)was 23.6 per cent, which is in line with earlier online surveys with a lengthof 15–20 minutes (Best & Krueger 2004).

Using a self-selection online survey obviously has advantages and disadvan-tages. The major advantage is that this method allows us to reach our researchpopulation, i.e. the visitors of party websites. In practice, there is no other wayto reach this population in an efficient manner. If we add the number of uniquevisitors of all party websites (provided by the webmaster), a conservative esti-mate is that something like 5–8 per cent of the adult population of Belgium actu-ally visited a party website during the electoral campaign. If we want to dividethis sample further according to party preference, the website being visited andsocioeconomic background characteristics, this would have required a hugepopulation survey. For example, according to the webmaster of that party,some 40,000 people visited the Groen! party website. This implies that itwould require roughly 12,000 respondents in a general population survey toobtain information on a sufficient number of Groen! website visitors. It isclear that such a survey is out of the question, both for financial and practicalreasons. A self-selected web survey, therefore, seemed the only appropriatemanner to investigate the information-seeking behaviour of party website visitorsin a sufficiently detailed manner.

The downside of this survey method is of course that highly educated orstrongly motivated visitors of party websites probably will be overrepresentedin this kind of selection process. While we recruit the respondents exactly atthe moment they perform the activity we are interested in (i.e. visiting aparty website), essentially this remains a process of self-selection, as somepeople will click on our button and others will decide not to. Nevertheless, itappears that this distortion should not be overestimated. For example, in oursample 63 per cent of all respondents reported they had finished higher edu-cation. In the representative general population survey (SCV), on the otherhand, 55 per cent of those who reported political information-seeking behaviouron the Internet also had obtained higher education. The difference between bothsurvey sources, all in all, remains rather modest. Since we are fully aware of the

R E A C H I N G O U T O R R E A C H I N G I N ? 6 9 9

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

selection bias that is inherent in our research design, in the remainder of thisarticle we will compare our results with the results of this population surveywhenever this is relevant.

Who visits political party websites and why?

With the decline in public involvement, the dwindling membership base for pol-itical parties and the boom in the Internet, various claims have been formulatedabout the impact of this new medium on political participation. High expec-tations emerged on the Internet’s possibilities and its opportunities for politicalparticipation and engagement. Can party websites attract a new audience thatwas not reached by other political media? The data of the I&PP surveysuggest that this is only partly the case. The party websites are visited predomi-nantly by those who are already politically interested. A comparison between therespondents of the I&PP survey and the respondents of the SCV survey (2005)who use the Internet for political purposes can inform us about the selectivenessof online party visitors (Table 2). In the SCV survey, the political use of theInternet is measured by the following items: search for information aboutcurrent events and politics, seeking and downloading governmental information,downloading and using forms and applications on governmental sites and

TABLE 2 Survey sample and general political use of Internet.

respondents

I&PP survey

users of Internet

for political information (SCV)

Gender

Woman 24.3 39.1

Man 75.7 60.9

Age (years)

,30 32.5 31.4

30–44 31.9 32.5

45–59 23.5 27.8

.59 12.1 8.3

Education

Secondary or less 37.0 44.6

Tertiary 63.0 55.4

Source: I&PP survey (n ¼ 3,224); SCV general population survey (n ¼ 338) (comparison with

online respondents who use the Internet for political info and e-government at least once a

week).

Note: Level of education without students. Units are percentages.

7 0 0 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

contacting politicians and government officers. To make sure that our results areobtained not just by the self-selection process of our respondents, we willcompare the respondents in the I&PP survey with a subsample of the respondentsin the SCV survey who use the Internet for political information or for e-govern-ment purposes at least once a week. The figures in Table 2 show that just asamong more traditional political activities, higher-educated males are stronglyoverrepresented among party website visitors. Three out of four visitors of pol-itical party websites turned out to be male and 63 per cent of the visitors haveattained higher education. However, contrary to more conventional forms ofpolitical participation, political party websites are rather successful in attractinga younger audience. If we compare with the SCV dataset, we can observe that theI&PP survey recruited more men and more highly educated respondents, com-pared with the survey subsample of political Internet users. With regard to agewe do not observe any major differences between the two surveys. This compari-son with a representative general population survey allows us to assume that thedistortion, caused by our selection method, remains rather limited.

To examine whether Belgian political party websites were able to attractpeople that are usually not interested in politics, we asked the respondentsabout their level of political interest (0–10 scale) and their level of political par-ticipation during the last 12 months (both online as in real life). It is clear thatparty website visitors belong to the more politically interested and active groupsin society (Table 3). The average level of political interest is 8.4, and this is farabove the population average. The high level of political interest and activity ofthis selective group is also closely related to the proportion of party memberswithin this group. Almost 46 per cent of the online party visitors are a partymember, which is very high, compared with only 6 per cent of the generalpopulation in Belgium.

TABLE 3 Political participation in last 12 months.

online real life

Financial support for political party 5.4 18.9

Financial support for other political organization or purpose 3.4 10.6

Political discussion with friends, family, colleagues 12.6 67.7

Contact with politician 23.3 26.0

Participation in strike – 7.0

Participation in demonstration – 20.3

Active campaigning for political party 8.0 25.7

Participation in political meeting/demonstration – 36.6

Source: I&PP 2006, n ¼ 3,224, missing values ¼ 88. The questionnaire informed whether

respondents had participated in one of these acts during the past 12 months, either using the

Internet (‘online’) or without the use of Internet (‘real life’). Entries are percentages.

R E A C H I N G O U T O R R E A C H I N G I N ? 7 0 1

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

An important part of the literature examining the role of the Internet inelection campaigns focuses on the interactive opportunities created by theInternet. Earlier research, however, indicates that party websites offer fewpossibilities for interactive dialogue (Davis et al. 2008). The I&PP survey allowsus to ascertain whether website visitors are actually looking for this kind of inter-active communication. Table 4 suggests that online party visitors mainly visit partywebsites to gather information rather than to interact with the party, politicians,other members or voters. Sceptics, of course, could argue that website visitorsreport they do not want interactive communication for the simple reason thatthis is not being offered anyhow by the parties. Earlier studies, however, indicatedthat whenever this kind of forum was being offered on the websites, it was used byonly a very small number of website visitors (Hooghe & Stouthuysen 2001).

Political websites as a campaign tool

The Internet becomes an increasingly important campaign tool for political parties.When we look at the number of unique visitors per day for the different partywebsites, we observe almost a doubling of visitors during the campaign periodfor the local 2006 elections compared with the pre-electoral period. This is asubstantial increase when we look back at the number of visitors in the previouslocal elections (Hooghe & Stouthuysen 2001). The question here is what roleparty sites play in the overall election campaign. Do they add something extra tothe traditional campaign? Are political party websites able to attract voters whocannot be reached via the traditional media, i.e. broadcast television, newspapers,radio . . . or are they rather used in a complementary or even cumulative manner?To tackle these questions we included in the questionnaire a battery of media use

TABLE 4 Main reasons for visiting party websites.

Information party programme 53.8

Information election campaign 35.7

Information candidates 34.2

Information political party 30.2

E-mail political party 9.2

Join a political discussion 9.2

E-mail politician 3.1

Membership political party 2.3

Other reason 7.2

Source: I&PP 2006, n ¼ 3,224, missing values ¼ 70.

Note: Sum of first and second most important reasons provided to visit party websites (the

total sum therefore is 200 per cent). Entries are percentages.

7 0 2 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

questions. The results clearly indicate that party website visitors are also avidconsumers of all other media with regard to political information (Table 5). Thevisitors of party websites read newspapers more frequently than the generalpopulation, while television news is less important to them than the populationaverage. These differences can be explained largely by the high education level ofthe I&PP respondents.

We can safely conclude from all the data presented thus far that party web-sites cannot be considered as effective tools to reach an audience who would notbe reached by traditional media. Party website visitors are strongly interested inpolitics, they have a high degree of party identification and they also have suffi-cient other means of information. As such, this analysis is in line with thereinforcement thesis: party websites tend to reach those who are already cogni-tively involved in politics. We can conclude that websites of political parties arenot an effective tool to reach groups of the population that usually do not receivepolitical information. The comparison with the political Internet users in thepopulation survey demonstrates that this conclusion cannot be the result justof the selective nature of our sample, but most likely reflects actualinformation-seeking behaviour of this group.

Preaching to the converted?

With regard to background characteristics and media use, the respondents in theI&PP survey clearly are a very selective group. We were also interested,

TABLE 5 Media use and news consumption.

television newspaper

party

website

visitors

(I&PP)

political

Internet

users

(SCV)

general

population

(SCV)

party

visitors

(I&PP)

political

Internet

users

(SCV)

general

population

(SCV)

6–7 days a week 45.3 56.2 62.0 41.1 48.8 40.3

3–5 days a week 36.1 32.0 24.6 26.5 19.2 13.3

1–2 days a week 9.6 6.5 5.8 14.2 10.9 12.5

Less than once a

week

4.7 3.3 4.6 8.9 14.5 15.2

Never 4.3 2.1 3.0 9.4 6.5 18.5

Source: I&PP survey 2006 (n ¼ 3,224), with 53 missing values for television, 50 for newspapers.

Political Internet users from SCV survey 2005 (n ¼ 338). General population: all respondents

in SCV survey (n ¼ 1,524).

R E A C H I N G O U T O R R E A C H I N G I N ? 7 0 3

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

however, in determining for what purpose they visit the party website andwhether this contributes to the exchange of ideas during an electoral campaign.Do party websites attract party members and partisans or do they also attractsupporters of other parties and citizens who did not decide yet on their partypreference? To answer this question we asked about the intended voting behav-iour of the website visitors. First, we asked them whether they had alreadydecided on what party they would vote for in the 8 October elections, andwhen there was a positive answer, we also asked for what party they intendedto vote.

The combination of both questions allowed us to distinguish three differentgroups among the respondents:

1. respondents who already decided they will vote for the party they just visited;2. respondents who decided they will vote for another party;3. respondents who had not made a decision yet (or at least did not want to

divulge that decision in the survey).

We have already mentioned that in Flanders (Dutch-speaking political parties) allparliamentary parties participated in the research, whereas this was not the casefor the French part of the country. Furthermore, French language respondents tothis survey proved less willing to divulge information about themselves or abouttheir political preferences, probably for reasons of a stronger concern for privacyin that part of the country. Since it is important to have a representative sampleto conduct this kind of analysis, we limit this analysis of voting behaviour to theDutch-speaking respondents in the I&PP survey.3

The results (Table 6) show that 59 per cent of the respondents actuallywanted to vote for the party they just visited. Another 31 per cent wanted tovote for another party and another 10 per cent of all respondents had notdecided yet about whom they were going to vote for. It is interesting to notehere that the weekday visitors (i.e. respondents that filled in the questionnairebetween Monday morning and Friday evening) were more likely to vote for

TABLE 6 Visitors (Flemish) political party websites.

decided to vote

for the same party

decided to vote

for another party did not decide yet n

Weekday visitors 60.5 30.2 9.3 1,352

Weekend visitors 54.2 31.7 14.1 454

All visitors 58.9 30.6 10.5 1,806

Source: I&PP 2006, n ¼ 1,806 (until the election day, 8 October 2006), missing values ¼ 28.

Note: n: exclusive French-speaking political parties, newsletters, mailing lists and search

engines.

7 0 4 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

the party they just visited than the ones who visited the party websites during theweekend.

Our first answer to the question, ‘do party websites mainly preach to theconverted?’, therefore, has to be rather mixed: for 60 per cent they do, for40 per cent they reach another audience. Various caveats come to mind whenconsidering these percentages. It may be more likely that respondents willclick a survey button when they visit a party they identify with, rather thanwhen they visit an alternate party. It could also be that they more often visitthe website of their preferred party, and that they do so for longer periods oftime, thus making it more likely that they will fill in a questionnaire fromthat specific website. Nevertheless, the figures in Table 6 still suggest thatparty websites are being visited predominantly by those who are alreadyconvinced.

Rather than to look at frequencies, however, it is interesting to look at thecharacteristics of these different groups of website users. In order to conduct amultivariate analysis, we constructed three groups:

1. voters who decided to vote for the party they visited online;2. voters who decided to vote for a different party;3. voters who had not made a decision yet.

The dependent variable in the three analyses is whether one belongs to one ofthese three groups or not (0 ¼ does not belong; 1 ¼ belongs). Every respon-dent, therefore, belongs to only one of these groups. The three separate analysespredict the likelihood that one will be a member of one of the three groups(party supporters, other party, do not know yet).

We used the same independent variables in all three models (Table 7). Thepredictors used in the model are the day of the election campaign (0 ¼ 1September to 37 ¼ 9 October), sociodemographic characteristics, i.e. gender,age, education level and frequency of media use. To capture media use, weincluded a question on the frequency of watching news broadcasts on televisionand reading newspapers. Both variables are continuous variables with a scaleranging from never (¼1) to every day (¼6). With regard to the Internet weincluded general Internet use, i.e. the Internet use at home, at work or atschool, ranging from less than once a month (1) till every day (6). We also includedInternet skills measured by indexing seven items: use search engines, send e-mails,participate in chat rooms, sign petitions, phone via Internet, share files and make awebpage. We also included political interest (0 ¼ not at all interested in politics;10 ¼ very much interested in politics) and membership of a political party in themodel. The variable membership of a political party was provided with fouranswering options: (1) yes, active member (participated this year in one ormore activities), (2) yes, passive member (paid membership fee, but did notparticipate in any activity, (3) not a member and (4) not a member anymore

R E A C H I N G O U T O R R E A C H I N G I N ? 7 0 5

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

TABLE 7 Predicting voting behaviour for the party visited online.

n

will vote for

same party

will vote for

different party

not yet

decided yet

Day election campaign 1,691 1.010� 0.989� 0.998

Weekday/weekend

Weekday 1,273 1.186 1.145 0.948

Weekend (a) 418 1 1 1

Gender

Man 1,081 0.976 1.827��� 0.475���

Woman (a) 307 1 1 1

Age 1,691 0.999 0.999 1.004

Education

Tertiary 1,081 0.925 1.039 1.056

Secondary (a) 610 1 1 1

Student 307 0.683� 1.425� 1.142

No student (a) 1,384 1 1 1

Media use

TV 1,691 0.968 1.094� 0.921

Newspaper 1,691 1.012 1.026 0.951

Internet 1,691 1.073 0.876 1.033

Political interest 1,691 1.089� 1.142�� 0.812���

Internet skills 1,691 1.045 0.957 0.988

Membership political party

Active member 566 3.409��� 0.646�� 0.035���

Passive member 204 2.405��� 0.797 0.224���

No member anymore 98 0.707 1.460 1.066

No member 823 1 1 1

x2 208.122��� 66.222��� 360.578���

Df 14 14 14

Pseudo-r2 0.156 0.057 0.325

Source: I&PP survey 2006, n ¼ 1,834, missing values ¼ 143.

Note: The model includes only respondents who were recruited via a Flemish party website to

the questionnaire. The dependent variables with regard to voting behaviour are measured as

a dichotomy where 1 ¼ yes, and 0 ¼ no. Units are odds ratios (exp b). (a), reference category.

�p , 0.05.

��p , 0.01.

���p , 0.001.

7 0 6 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

(for more details about the variables we refer to Table 8). The effect of theindependent variables is presented as an odds ratio (exp b), expressing thelikelihood that a respondent will belong to a specific group (¼1) comparedwith a reference group.

The results of the analysis reported in Table 7 show a different pattern forthe three electoral groups, so that we can safely state that these are indeed threedistinct groups. First we turn to the largest group: respondents who are alreadycertain they will vote for the party they just visited. This is quite a powerfulmodel, with an explained variance of 0.16. Obviously, active party membersare much more likely to vote for the party they just visited. As the campaignadvances, there is a slight but significant increase in this group. Those whoscore high on political interest are also more likely to have a clear and fixed pre-ference, while we can observe that students are significantly less likely to belongto this category.

The model for the second group, i.e. the respondents who decided to votefor a different party, has a rather weak explanatory power of 0.06. Here, typi-cally, we find less active party members, and more men than women. Studentsare clearly overrepresented in this group, and we have to remember in thisrespect that most of them will go out voting for the first time. This groupalso tends to score high on political interest and watches slightly more televisionfor current events and political information. Members of this group tend to visitparty websites relatively early in the campaign.

Finally, we arrive at the strongest model to explain the membership of thegroup that has not decided yet on which party to vote for during the elections.This analysis has an explanatory power of 0.33, which is remarkably high. Firstof all, we can observe that the day of visiting the party websites does not havean effect on this variable. This implies that those who visit the party websiteearly on in the campaign, on 1 September, are just as likely to say they do notknow which party to vote for as those who visit the website just the day beforethe elections. Typical characteristics of this group are a relatively low level of pol-itical interest and the fact that one is not a party member. It is also striking to notethat women are more strongly represented in this group than among the overallgroup of respondents. To some extent, this might be linked to the fact that, ingeneral, in surveys women more often use the ‘do not know’ or ‘have notdecided yet’ option than men tend to do (Mondak & Anderson 2004).

If we compare all three models, one can observe that traditional backgroundcharacteristics like age or education level in general are not significant. This suggeststhat party website visitors, no matter what their voting intention is, are all citizenswith a higher education and are intensive Internet users, and they have, comparedwith the average citizen, more Internet skills. This explains why, among the I&PPrespondents, there is little further variance on these variables.

We can conclude from this analysis that those who are party members orthose who are strongly politically interested tend to use party websites in a

R E A C H I N G O U T O R R E A C H I N G I N ? 7 0 7

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

TABLE 8 Variables used in the analysis.

I&PP (2006)

name wording coding

medium

value (SD)

Day of the

election

campaign

Information of the I&PP log

files ranging from 0 to 37

(final day)

Ranging from 0 ¼ start of

the survey to

37 ¼ election day

19.777 (11.905)

Weekday/

weekend

Information of the I&PP log

files

0 ¼ weekday;

1 ¼ weekend

0.251 (0.434)

Gender Sex of respondent 0 ¼ male; 1 ¼ female 0.284 (0.455)

Age Year of birth ranging from

1901 till 1999

Age (in 2006) ranging

from 7 years to 105

years (sic)

38.134 (14.929)

Education What is the highest degree

you have achieved?

Ranging from (1) None to

(13) College degree

0 ¼ tertiary;

1 ¼ secondary or less

0.365 (0.481)

Student Which of these descriptions

best applies to your

situation (only one answer

possible)

0 ¼ student; 1 ¼ no

student

0.818 (0.386)

Television news How many times did you

watch TV news the last

week? Ranging from (1)

Never to (6) Daily

4.756 (1.464)

Newspaper How often did you read the

newspaper last week?

Ranging from (1) Never to

(6) Daily

4.431 (1.708)

Internet use How often do you use

Internet? Ranging from (1)

Never to (6) Every day

5.864 (0.452)

Political interest How strong is your interest

in politics (0 ¼ no interest

at all, 10 ¼ strongly

interested)

8.302 (1.878)

(Continued)

7 0 8 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

segmented manner. They tend to visit the party they are going to vote foranyhow. A substantial group, however, has not made a final decision on votingbehaviour yet. This group has lower levels of party identification and is lesspolitically interested.

Conclusion

With this article we hope to provide new insights into the theoreticalunderstanding of cyber politics and to provide new empirical material to the

TABLE 8 Continued.

I&PP (2006)

name wording coding

medium

value (SD)

Internet skills Did you ever perform one of

these activities on the

Internet: (1) use a search

engine; (2) send e-mail

with attachment; (3)

participate in chat rooms

or newsgroups; (4) sign

petition; (5) phone over

the Internet; (6) peer-to-

peer file sharing; (7)

create webpages

Sum scale: ranging from

0 ¼ no Internet-related

skills to 7 ¼ all

Internet-related skills

4.151 (1.734)

Membership

political party

Are you member of a

political party? (1) Yes,

active member (you have

participated in one or

more activities last year);

(2) yes, passive member

(you pay member fee, but

you do not participate in

activities); (3) no, (4)

member in the past (you

were once a member, but

now not anymore)

2.717 (1.359)

Note: Further information on the I&PP survey and the full questionnaire for that survey can be

found at www.kuleuven.be/citizenship.

R E A C H I N G O U T O R R E A C H I N G I N ? 7 0 9

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

discussion. This article approached the issue of what role party websites play inthe overall election campaign by focusing on the user dimension of onlinepolitics. Therefore, the article focused on the empirical question of who visitswebsites of political parties and for what purpose. Do party websites succeedin mobilizing new groups of voters, or do they just replicate or even reinforcepre-existing patterns of inequality and segmentation?

Exploration of the data of the I&PP survey suggests that political partywebsites are visited by a very specific segment of the electorate. In general,political party websites tend to attract higher-educated Internet users, menand citizens who are strongly politically interested. Almost half of the respon-dents of our web survey are even a member of a political party comparedwith 6 per cent among the general adult population of Belgium. Our findings,therefore, seem to support the reinforcement thesis, and our findings are herein line with some of the other available studies on politically motivated Internetusers (Boogers & Voerman 2003; Gibson et al. 2005). While these results mightbe an effect of the self-selection of the respondents, it has to be noted that ourresults do not differ all that much from similar questions asked in a generalpopulation survey.

Our second research question was whether party websites amount to‘preaching to the converted’, or whether they are used by voters to comparevarious parties. Here, our answer needs to be qualified. We can observe theoccurrence of two distinct groups: on the one hand, we have staunch party sup-porters who seek additional information on their own party; on the other hand,we have citizens who seek out information on other parties than their preferredone or who have not decided yet. Both groups have distinct characteristics: whilethe party supporters score extremely high on political interest and are predomi-nantly male, we can observe that the second group has lower levels of politicalinterest and there are more women present in this group. In this specific survey,60 per cent of all respondents belonged to the group of party supporters, and 40per cent to the group of pluralistic information seekers. Given the nature of ourprocess of data collection, these percentages should be seen as only a first indi-cation, and it could be a safe conclusion, therefore, that in reality both groupsare, speaking very roughly, equally represented. With regard to the Sunsteinthesis on polarization and segmentation, our only conclusion therefore can bethat the glass is half-full, or half-empty if one prefers. Party websites are animportant tool for intra-party communication, and to a large extent they areindeed used by party supporters and party members. Contrary to what somehave feared, however, they are also visited by a substantial number of Internetusers who do not even think about voting for that specific party, or who havenot decided yet.

What we hope to have demonstrated in this article is that perhaps there is nosingle answer to the general question whether the Internet presence of politicalparties leads to political segmentation. After all, segmentation is not a

7 1 0 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

characteristic that can be studied at the level of the website or the informationbeing offered. Segmentation is a characteristic of the information-seeking behav-iour of individual Internet users. Some websites might lend themselves moreeasily than others to segmented use, but in the end individuals have specific infor-mation strategies. Self-evidently, there are specific caveats if we want to studyindividual Internet use and, as we have seen, reaching representativeness is amajor concern. Nevertheless, it is clear that studies on the effect of party web-sites should not limit themselves to studying those party websites: individualusers, too, have to be taken into account, despite all the empirical problemsassociated with this step. What our study suggests is that there are distinctgroups of users, with distinct information-retrieval patterns. As such, we canalso expect that political consequences differ as a result of the specific interactionbetween the information being offered by the political parties and user charac-teristics. What is clear, however, is that we cannot find a single indication thatparty websites would be efficient tools to reach out to new audiences who arenot being reached by any other means.

Acknowledgements

We would like to thank the webmasters of the political parties in Belgium for theirhelp in collecting our data. We gratefully acknowledge the generous researchfunding by the Federal Science Agency of Belgium (InterMob Program).

Notes

1 The SCV survey was conducted on a face-to-face basis in 2005. Of the1,524 initial respondents, 656 used the Internet. To make the comparisonas close as possible, we limited ourselves to those who report that they usethe Internet at least once a week for political reasons and e-government(n ¼ 338).

2 It has to be noted that here we focus on traditional party websites and noton all kinds of more innovative ‘Web 2.0’ applications. These were,however, completely absent in this campaign (Hooghe & Vissers 2008).

3 Although we would have preferred to conduct this analysis on the entire(Belgian) sample, in stead of just on the Dutch-speaking part of it, thereis no reason to assume that there is a difference between Dutch- andFrench-speaking respondents in this analysis. A separate analysis on justthe French respondents shows broadly the same patterns, whereas anearlier analysis has shown that with regard to political Internet use,there are no significant differences between Dutch- and French-speakingcitizens of Belgium.

R E A C H I N G O U T O R R E A C H I N G I N ? 7 1 1

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

References

Best, S. & Krueger, B. (2004) Internet Data Collection, Sage, Thousand Oaks, CA.Bimber, B. (2000) ‘The study of information technology and civic engagement’,

Political Communication, vol. 17, no. 4, pp. 329–333.Bimber, B. (2001) ‘Information and political engagement in America: the search for

effects of information technology at the individual level’, Political ResearchQuarterly, vol. 54, no. 1, pp. 53–67.

Boogers, M. & Voerman, G. (2003) ‘Surfing citizens and floating voters: results ofan online survey of visitors to political web sites during the Dutch 2002general elections’, Information Polity, vol. 8, no. 1, pp. 17–27.

Chaiken, S. & Stangor, C. (1987) ‘Attitudes and attitude change’, Annual Review ofPsychology, vol. 38, pp. 575–630.

Davis, R. (1999) The Web of Politics: The Internet’s Impact on the American PoliticalSystem, Oxford University Press, New York.

Davis, R. et al. (2008) Making a Difference. A comparative View of the Role of the Internetin Election Politics, Lexington Books, Lanham, MD.

Geys, B. (2006) ‘District magnitude, social heterogeneity and local party systemfragmentation’, Party Politics, vol. 12, no. 2, pp. 281–297.

Gibson, R. & McAllister, I. (2006) ‘Does cyber-campaigning win votes? Onlinecommunication in the 2004 Australian election’, Journal of Elections, PublicOpinion and Parties, vol. 16, no. 3, pp. 243–263.

Gibson, R., Rommele, A. & Ward, S. (eds) (2004) Electronic Democracy. PoliticalOrganisations, Mobilisation and Participation Online, Routledge, London.

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

Graber, D. (2004) ‘Mediated politics and citizenship in the twenty-first century’,Annual Review of Psychology, vol. 55, pp. 545–571.

Hansen, K. M., Pedersen, K. & Wahl-Jørgensen, D. (2005) ‘Cyber-campaigning.The character, development and consequences of parties’ electioneeringin cyber space’, Paper presented at the Nordic Political Science AssociationConference, Reykjavik, August, 11–13, 2005.

Hooghe, M. & Stouthuysen, P. (2001) ‘Het politiek gebruik van internet naar aan-leiding van de gemeenteraadsverkiezingen van 8 oktober 2000’, Res Publica,vol. 43, no. 4, pp. 507–528.

Hooghe, M. & Vissers, S. (2008) ‘Websites as a campaign tool for Belgian politicalparties. A comparison between the 2000 and 2006 local election campaigns’,in Making a Difference: a Comparative View of the Role of the Internet in ElectionPolitics, eds R. Davis, D. Owen, D. Taras & S. Ward, Lexington Press,Lanham, MD, pp. 171–196.

Lazarsfeld, P., Gaudet, B. & Berelson, H. (1965) The People’s Choice. How the VoterMakes up his Mind in a Presidential Campaign, Columbia University Press,New York.

7 1 2 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

Lee, E.-J. (2007) ‘Deindividuation effects on group polarization in computer-mediatedcommunication’, Journal of Communication, vol. 57, no. 2, pp. 385–403.

Lusoli, W. (2005) ‘The internet and the European Parliament elections: theoreticalperspectives, empirical investigations and proposals for research’, InformationPolity, vol. 10, nos 3–4, pp. 153–163.

Lusoli, W. & Ward, S. (2004) ‘Digital rank-and-file: party activists’ perceptions anduse of the Internet’, British Journal of Politics and International Relations, vol. 6,no. 4, pp. 453–470.

Lusoli, W. & Ward, S. (2005) ‘Logging on or switching off? The public and the Inter-net at the 2005 general election’, in Spinning the Web. Online Campaigning inthe 2005 General Election, eds S. Coleman & S. Ward, Hansard Society,London, pp. 13–21.

Lusoli, W., Ward, S. & Gibson, R. (2006) ‘(Re)connecting politics? Parliament, thepublic and the Internet’, Parliamentary Affairs, vol. 59, no. 1, pp. 24–42.

Margolis, M. & Resnick, D. (2000) Politics as Usual: The Cyberspace ‘Revolution’,Sage, Thousand Oaks, CA.

Mondak, J. & Anderson, M. (2004) ‘The knowledge gap. A reexamination ofgender-based differences in political knowledge’, Journal of Politics, vol. 66,no. 2, pp. 492–512.

Mutz, D. (2006) Hearing the other side. Deliberative Versus Participatory Democracy,Cambridge University Press, Cambridge.

Nadel, M. (2002) ‘Customized news services and extremist enclaves in Repub-lic.com’, Stanford Law Review, vol. 54, no. 4, pp. 831–886.

Norris, P. (2001) Digital Divide. Civic Engagement, Information Poverty, and the InternetWorldwide, Cambridge University Press, Cambridge.

Norris, P. (2003) ‘Preaching to the converted? Pluralism, participation and partywebsites’, Party Politics, vol. 9, no. 1, pp. 21–45.

Pedersen, K. & Saglie, J. (2005) ‘New technology in ageing parties. Internet use inDanish and Norwegian parties’, Party Politics, vol. 11, no. 3, pp. 359–377.

Sears, D. & Freedman, J. (1967) ‘Selective exposure to information. A criticalreview’, Public Opinion Quarterly, vol. 31, no. 2, pp. 194–213.

Shah, D., Nojin, K. & Lance Holbert, R. (2001) ‘Connecting’ and ‘disconnecting’with civic life. Patterns of Internet use and the production of social capital,Political Communication, vol. 18, no. 2, pp. 141–162.

Sinekopova, G. (2006) ‘Building the public sphere. Bases and biases’, Journal ofCommunication, vol. 56, no. 3, pp. 505–522.

Small, T. (2007) ‘Canadian cyberparties. Reflection on internet-based campaigningand party systems’, Canadian Journal of Political Science, vol. 40, no. 3,pp. 639–657.

Strandberg, K. (2006) Parties, Candidates and Citizens On-line. Studies of Politics on theInternet, Abo Akademi University Press, Turku, Finland.

Strandberg, K. (2008) ‘Online electoral competition in different settings. A com-parative meta-analysis of the research on party websites and online electoralcompetition’, Party Politics, vol. 14, no. 2, pp. 223–244.

R E A C H I N G O U T O R R E A C H I N G I N ? 7 1 3

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4

Sunstein, C. (2001) Republic.com, Princeton University Press.Van Dijk, J. (2005) The Deepening Divide. Inequality in the Information Society, Sage,

Thousand Oaks, CA.Ward, S. & Gibson, R. (2003) ‘On-line and on message? Candidate websites in the

2001 general election’, British Journal of Politics and International Relations, vol.5, no. 2, pp. 188–205.

Ward, S. & Lusoli, W. (2005) ‘“From weird to wired”. MPs, the Internet and repre-sentative politics in the UK’, Journal of Legislative Studies, vol. 11, no. 1,pp. 57–81.

Ward, S., Gibson, R. & Nixon, P. (2003) ‘Parties and the Internet: an overview’, inPolitical Parties and the Internet. Net Gain?, eds R. Gibson, P. Nixon & S. Ward,Routledge, London, pp. 11–38.

Warschauer, M. (2003) Technology and Social Inclusion. Rethinking the Digital Divide,Massachusetts Institute of Technology, Cambridge.

Marc Hooghe is a Professor of Political Science in the Department of Political

Science at the Catholic University of Leuven (Belgium). He has published mainly

on political participation, social capital and inequality. Address: Department

of Political Science, KU Leuven, Park Street 45, B-3000 Leuven, Belgium.

[email: [email protected]]

Sara Vissers is a PhD student in Political Science in the Department of Political

Science at the Catholic University of Leuven (Belgium). Her main research

interest is inequality, Internet and political mobilization. Address: Department

of Political Science, KU Leuven, Park Street 45, B-3000 Leuven, Belgium.

[email: [email protected]]

7 1 4 I N F O R M A T I O N , C O M M U N I C A T I O N & S O C I E T Y

Dow

nloa

ded

by [

Tex

as A

& M

Int

erna

tiona

l Uni

vers

ity]

at 0

2:18

08

Oct

ober

201

4