This article was downloaded by: [University of Tennessee, Knoxville]On: 15 September 2014, At: 10:38Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Information Technology & PoliticsPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/witp20

You’ve Got (No) Mail: How Parties and CandidatesRespond to E-mail Inquiries in Western DemocraciesCristian VaccariAccepted author version posted online: 11 Mar 2014.Published online: 12 May 2014.

To cite this article: Cristian Vaccari (2014) You’ve Got (No) Mail: How Parties and Candidates Respond to E-mail Inquiries in Western Democracies, Journal of Information Technology & Politics, 11:2, 245-258, DOI:10.1080/19331681.2014.899536

To link to this article: http://dx.doi.org/10.1080/19331681.2014.899536

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe 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 reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Journal of Information Technology & Politics, 11:245–258, 2014Copyright © Taylor & Francis Group, LLCISSN: 1933-1681 print/1933-169X onlineDOI: 10.1080/19331681.2014.899536

You’ve Got (No) Mail: How Parties and CandidatesRespond to E-mail Inquiries in Western Democracies

Cristian Vaccari

ABSTRACT. Although e-mail is one of the most popular components of users’ experiences of theInternet, its use by political actors in campaigns has rarely been studied. In this article, I explore politicalactors’ responsiveness to e-mails coming from citizens through a large-scale, longitudinal study of194 parties and candidates in Australia, France, Germany, Italy, Spain, the United Kingdom, and theUnited States between 2007 and 2013. In order to assess political actors’ e-mail responsiveness, twofictitious e-mails were sent to each of them: one requesting issue information, the other pledging tobe willing to volunteer. Results show that most parties and candidates fail to respond to both types ofe-mails, and that progressive parties tend to respond more than conservative ones.

KEYWORDS. Comparative research, e-mail, Internet politics, online campaigning, political parties,responsiveness

E-MAIL: THE GREAT UNKNOWN?

There is broad scholarly consensus on theidea that “parties have adapted more quickly tonew media technologies than to any previoustechnological advance” (Ward, 2008, pp. 1–2).Across the Western world, political actors haveestablished a diversified, multifaceted presenceacross a variety of digital platforms, including,among others, Web sites, blogs, social media,and mobile applications. However, in spite of allthese efforts by politicians, most voters in mostcampaigns do not seem to take notice. Apartfrom widely touted exceptions, such as BarackObama’s 2008 campaign and a handful of otherhighly publicized success stories, digital outletssuch as campaign Web sites and social media

Cristian Vaccari is lecturer in political science at Royal Holloway, University of London and at theUniversity of Bologna. He studies political communication in a comparative perspective, with a particularfocus on digital media. His latest book is titled Digital Politics in Western Democracies: A Comparative Study(Baltimore, MD: Johns Hopkins University Press, 2013). He is the principal investigator of a comparativeresearch project on social media and political inclusion (http://www.webpoleu.org).

Address correspondence to: Cristian Vaccari, Department of Politics and International Relations, RoyalHolloway, University of London, Egham, Surrey TW20 0EX, UK (E-mail: cristian.vaccari@rhul.ac.uk).

profiles fail to attract significant attention amongvoters. For instance, during the United States2010 midterm elections, the median candidatein competitive House districts (which on aver-age comprise about 700,000 voters) had onlyabout 1,800 Facebook “friends,” 350 Twitter“followers,” and 1,500 monthly visits to his/herWeb site (Nielsen & Vaccari, 2013). In theFrench 2012 presidential elections, the los-ing incumbent Nicolas Sarkozy had around630,000 supporters on Facebook, whereas thevictorious challenger François Hollande hadabout 120,000. These numbers pale by com-parison with their vote totals of 10.2 millionand 9.7 million in the first round, respectively,which increased to 18 million and 16.9 millionin the runoff. Parties’ and candidates’ online

245

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

246 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

tools are certainly useful to inform and engageparticular segments of the electorate, but theaudiences that politicians manage to attract onmost platforms are small compared to their votegoals and the publics that they reach via the massmedia.

One of the reasons might be that most of theelectorate either ignores some digital tools or isnot interested in politicians’ presence therein.In this respect, Rasmus Kleis Nielsen (2011)introduced a useful distinction between mun-dane online tools, which are routinely used bymost people as part of their everyday lives;emerging tools, which are used only by minori-ties at a given point in time; and specializedtools, which are only available to insiders or tothose who are willing to pay for them. Nielsennotes that, in spite of all the glamour that emerg-ing and specialized technologies often receive inpublic debates, “When it comes to mobilization,mundane Internet tools like e-mail and searchare more important than emerging tools (likesocial networking sites) or specialized tools (likecampaign websites)” (Nielsen, 2011, p. 756,emphasis in original). Mundane tools have abroader reach than other, fancier e-campaigningplatforms, and voter mobilization needs to reachaudiences as large as possible.

It is hard to think of an online tool thatis more “mundane” than e-mail. According toEurostat (2013), in 2012, 70% of citizens livingin the European Union had used the Internet atleast once a week in the last three months and65% had used the Internet to send or receivee-mails. Thus, virtually all Europeans who areonline use e-mail. Similarly, a Pew Internet& American Life survey shows that in 2011,92% of American adults with Internet accesshad used e-mail (Purcell, 2011). The near-universality of e-mail among citizens makes ita very promising tool through which politicianscan engage their constituents.

Research on digital politics, however, hasmostly neglected e-mail. Instead, most studieshave focused on Web sites (e.g., Gibson et al.,2003; Lilleker et al., 2011), blogs (Davis, 2009),social media (Vergeer, Hermans, & Sams 2013),and database technologies (Howard, 2006).

One rather isolated exception is Jackson’s(2004) study of e-mail use by British Members

of Parliament in 2002, which found that the vastmajority of MPs received relatively few e-mails,most of which were from non-constituents, andthat MPs claimed that they were coping rea-sonably well in dealing with them. Apart fromthis study, most available research on politicale-mails addresses how campaigns use e-mail tomobilize citizens rather than to respond to theirinputs, and has been conducted in one country—the United States—rather than relying on thecomparative method to test theories and derivefindings that can be generalized beyond a sin-gle country. For instance, Klotz (2007) notedthat U.S. Senate campaigns encourage citizensto send prewritten letters to media editors, whichresults in what he considers “plagiarized partici-pation” and puts unnecessary pressure on editorsto sort genuine communication from “Astroturf”letters. Klotz sees similar problems with “tella friend” buttons on Web sites, which allowcitizens to send campaign-crafted messages totheir online contacts under the false pretensethat they were written independently by thesender. Krueger (2006) found that, as a mobi-lization tool, e-mail reaches a more inclusiveaudience than offline channels, such as directmail, telephone calls, and door-to-door canvass-ing. Because the marginal costs of e-mail com-munication are practically null, political orga-nizations can afford to send messages to anycitizen who is on their list rather than care-fully choosing which voters to contact based ontheir political predispositions. However, Greenand Gerber (2004) studied the impact of e-mailmobilization through randomized field experi-ments and found negligible effects on both voterregistration and turnout. Similar findings wereobtained in another set of field experiments runby Nickerson (2007) and in a study based on asurvey of Internet users by Johnson and Krueger(2012). By contrast, Malhotra, Michelson, andValenzuela (2012) found that e-mails by offi-cial sources have a small but significant positiveeffect on turnout, while those coming from votermobilization organizations have no effect.

From citizens’ perspectives, e-mail is one ofthe tools through which voters can keep politi-cians accountable by interacting with them ina limited but practical way. First, the costs ofsending an e-mail to a politician or party are

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 247

lower than those of any other type of elite-citizencontact (such as face-to-face exchanges or hand-written letters), both in practical and psycho-logical terms. These affordances, however, alsoimply that e-mail presents politicians and theirorganizations with a greater flow of messagesthan they are used to managing through otherchannels, such as phone, mail, and personalcontacts. Effectively addressing this challenge,however, can be very important. As shown bya series of focus groups conducted by Stromer-Galley and Foot (2002), candidates’ failure torespond to e-mail inquiries can cause a breachof trust among voters:

The norms of interpersonal conversationare violated if a response is not received.The sender of the message feels slighted ifafter sending the message no response orfeedback is given back to the sender. . . .

The implication for candidates and theironline campaigns is that they risk giv-ing participants a negative impression bynot replying to e-mail messages. (Stromer-Galley & Foot, 2002)

By answering voter inquiries, politicians canshow that they are responsive to people’sdemands and thus strengthen their ties with thecitizenry in a way that is increasingly difficultto achieve through the packaged and insider-dominated mass media (Coleman & Blumler,2009, pp. 42–67; Entman, 1990). Strengtheningaccountability, in turn, is a potentially effectiveantidote to the crisis of representation (Dalton,2004) and the declining trust in political insti-tutions and authorities (Norris, 2011) that areaffecting Western democracies. Coleman (2005,p. 200) claims that representatives can recon-nect with citizens by substituting closenessfor distance, mutuality for detachment, coher-ence for exclusivity, and empathy for aloof-ness. Citizens’ trust in their representativesstems from “affective perceptions of attach-ment, affinity, and respect” (Coleman, 2005,p. 206), which depend more on how politi-cians relate to them than on policy choicesand legislative achievements. As suggested byColeman and Blumler, such empathy cannot be

built unless politicians become more accessi-ble, a goal that the affordances of digital mediacan help achieve: “Citizens regard mediated co-presence as grounds for rejecting forms of thinpolitical communication with which they oncehad to be satisfied. . . . As representatives canbe addressed at any time and in any placevia mobile phone, text message or e-mail, cit-izens experience greater communication equal-ity” (2009, p. 82). Whereas the multiplicationof message exchanges made possible by digi-tal media often results in fear of communicationoverload among politicians, “From citizens’ per-spective, the problem is one of non-responserather than overload” (Coleman, 2005, p. 208).As a young voter claimed in an interview, “Youcan e-mail [the MP]. But is he going to listen?”(Livingstone, 2004, p. 7). Indeed, “The technicalcapacity to facilitate citizen-to-representativedialogue makes failures of responsive commu-nication even more conspicuous than they hadbeen in the pre-digital era” (Coleman, 2009,p. 86). In sum, opportunities breed expectations,as citizens demand that political actors behavein a more responsive fashion on the Web than inmass media environments.

Based on these considerations, one mightexpect that political organizations such as par-ties and campaigns should make any effort torespond to the requests they receive via digi-tal media, particularly e-mail which is the mostmundane among them. Undecided voters inquir-ing about policy stances might be impressed bya swift and substantive response by a party orcandidate, even if staffers are responsible forreplying. Prospective volunteers might be turnedinto engaged activists by a timely e-mail thatprovides practical information and direction onhow they can help. Regardless of the content andscope of the original inquiry, a quick and preciseresponse makes any organization look efficient,competent, and caring for its constituents andpublics. Despite all these motivations for beingresponsive, however, some factors also work inthe opposite direction.

First, answering e-mails individually canplace a huge organizational burden on cam-paigns that need to allocate scarce volun-teers’ and staffers’ time to these activities.Moreover, these people must be trained so

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

248 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

that their responses are consistent with theoverall message and, in case more than oneperson is tasked with answering e-mails,their replies must not contradict one another.Although automated replies are often providedby campaigns, they can hardly fulfill the goals ofreconnecting elites with citizens and satisfyingusers’ requests.

Second, many things could go wrong whenstaffers and volunteers send individual mes-sages to strangers. For instance, responses couldmisrepresent the campaign message, dissatisfyrecipients, or reveal sensitive insider informa-tion that should not be made public. Giventhat e-mails can easily be forwarded, messagesthat were intended to be private may becomepublic and embarrass the organization. This con-cern exemplifies what Ward and Vedel (2006,p. 220) call the “fear factor” that often impedesinstitutional actors’ approach to innovations.

Third, political organizations may be less atease with incoming e-mails than with other dig-ital tools because they cannot control the timing,content, rules, and boundaries of the interaction.As Nielsen (2011) points out, when campaignsrespond to e-mail inquiries they cannot “man-age” the relationship through control of voterinformation, as they often do in other domains ofdigital electioneering (Howard, 2006; Vaccari,2010). Just as campaigns prefer broadcast-ing self-produced advertisements rather thananswering questions from journalists, so theyprefer e-mailing supporters (about whom theymight know enough to personalize the messagesthey send them) with news and requests ratherthan dealing with inquiries coming from peoplethey know little to nothing about.

RESEARCH QUESTIONS ANDMETHODOLOGY

This article aims to study political actors’e-mail responsiveness in a comparative per-spective. As emerged from the review of theliterature discussed above, most available stud-ies on the role of e-mail in campaigning havefocused solely on the United States, and nonehas employed the comparative method. This ten-dency, which is rather common in the literature

on digital politics, has two troublesome implica-tions. On the one hand, it might lead to unwar-ranted generalizations of U.S.-originated find-ings to other countries. This is not a desirableoutcome, because the American political systemoffers uniquely strong incentives for politicalactors to employ digital media, thus makingthe U.S. a deviant rather than modal case incomparative perspective (Anstead & Chadwick,2009). On the other hand, it is only by compar-ing different countries that we can understandhow digital politics is shaped by systemic factorssuch as technological development, the socio-political environment, and party-level character-istics (Nixon, Ward, & Gibson, 2003, p. 241).In light of the theoretical background and empir-ical evidence from previous research summa-rized above, this study aims to answer the fol-lowing questions across the realm of variousWestern democracies:

RQ1: To what extent and how quickly dopolitical actors answer e-mails fromcitizens?

RQ2: Do parties’ and candidates’ responserates and timing differ between e-mailsinquiring about issue information ande-mails requesting assistance on how toget engaged with the campaign?

RQ3: Which factors related to technologicaldevelopment, the socio-political envi-ronment, and party internal characteris-tics predict political actors’ responsive-ness to e-mails?

This final question incorporates a series ofcausal hypotheses derived from theory and evi-dence from previous research. Technologicaldevelopment in a country may affect the incen-tives that parties and candidates have to investin digital media, because it determines howlarge their potential online audiences are aswell as citizens’ expectations on the sophisti-cation of parties’ digital efforts. Regarding thesocio-political environment, the degree to whichcitizens participate in politics affects the demandfunction for parties’ online efforts. Party inter-nal characteristics such as resources, organiza-tion, and ideology may affect various aspects

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 249

of their online communication. The normaliza-tion theory (Margolis & Resnick, 2000) positsthat better resourced parties and candidates tendto be more visible and effective online, result-ing in uneven competitive patterns that mayalso be reflected in responsiveness to e-mails.Karpf (2012) has suggested that incumbencymay reduce incentives to adopt communica-tive innovations, which leads us to expect thatchallengers could be more responsive to citi-zen e-mails. Also, parties and campaigns thatengage with their supporters through inclusiveorganizational practices may be expected to bemore responsive than parties that are structuredin a more centralized and hierarchical fashion.Finally, ideology may affect political actors’approaches to the Web, including their respon-siveness to e-mails (Ward & Gibson, 2009).

To address these issues in comparative per-spective, my analysis will focus on Australia,France, Germany, Italy, Spain, the UnitedKingdom, and the United States, in a periodbetween 2007 and 2013. This sample offersmeaningful variance on many relevant charac-teristics and can thus be considered representa-tive of a broader realm of established democra-cies. Among the various types of case selectiontechniques, the diverse case selection strategyaims to achieve “maximum variance along rel-evant dimensions” within a defined population(Seawright & Gerring, 2008, p. 300). By includ-ing cases that differ as widely as possible insome theoretically relevant characteristics whilestill belonging to the same class, this methodallows incorporating various different causalconfigurations that may affect the dependentvariables. The sample of cases studied here canbe considered diverse because, although they allpertain to the realm of Western democracies,they differ from one another in many theoreti-cally relevant ways, including their technologi-cal development, political institutions, politicalculture, and political context.

In each of these countries, I focus on nationalelections—including European Parliament elec-tions for European Union members and U.S.presidential primaries—because they constitutea focal point in representative democracy andprovide one crucial mechanism that helps keepparties and representatives accountable. This

choice, however, implies that my findings cannotbe automatically generalized to “peacetime”periods between elections. I focus on politicalparties, their leaders, and presidential electioncandidates as the key actors involved in electioncampaigns. In order to identify which subjectsto study, before each election I compiled a list ofpolitical actors that could be considered relevantin the next campaign based on Sartori’s (1976)criteria of coalitional and blackmail potential.Parties (and their presidential candidates andnational leaders, to the extent that they had apersonal Web site) were thus included whenthey could be expected to become part of agoverning coalition or when they could con-ceivably affect electoral competition. Becausethese were prospective rather than retrospec-tive judgments, unfortunately they sometimesleft out parties that turned out to be rather suc-cessful or to use technology in an interestingfashion.1 That being said, retrospective evalua-tions confirm that all the major political actorswere included in the corpus. This procedure ledto a total of 304 contacts made with politicalactors between 2007 and 2013. A breakdown ofthe number of parties, leaders, and candidatesthat were contacted in each country for each yearis presented in the Appendix. Because multi-ple national elections occurred in all countries,some political actors were contacted in morethan one occasion.2 The total number of par-ties, party leaders, and presidential candidatesthat were contacted at least once is thus 194.Of these, 113 were parties and 81 were presiden-tial candidates or party leaders. Among all thepolitical actors studied here, 124 were contactedin one occasion, 31 in two occasions, 38 in threeoccasions, and one (Barack Obama, who ran inboth 2008 and 2012 and in both primary andgeneral elections) in four occasions.

This study will focus on two dependentvariables that measure whether political actorsresponded to a specific type of e-mail: onerequesting information on policy issues, theother pledging to volunteer for the recipientorganization. These types of e-mails correspondto the two main purposes of digital media forelection campaigning as identified by Gulatiand Williams (2007, pp. 2–3)—informing andengaging voters. E-mails were sent to each

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

250 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

political actor within the last two weeks beforeelection day. Fictitious e-mail addresses andsender names were used so that political orga-nizations would not be aware that their respon-siveness was being monitored, which could haveinfluenced their behavior.3 The e-mails werewritten in each country’s official language, weredirected to the address or contact form for gen-eral inquiries accessed through the official Website, and had substantially identical content. Theissue e-mail asked for the party or candidate’sposition on taxes, a policy on which each partymust take a stand and that is discussed in anyelection in every country. Thus, voter inquirieson taxes should be treated by all political actorsin all contexts with comparable priority lev-els, as opposed to policy domains, such as theenvironment, civil rights, or European Uniongovernance, which are more salient for someparties than others. The volunteer e-mail statedthat the sender was interested in helping theparty or candidate in carrying out any kind ofcampaign tasks and asked for instructions onhow to do so. Responses to these e-mails wereconsidered valid only if the researcher evalu-ated them as addressing the questions and beingwritten by a human being, while generic andautomated responses were discounted. The tim-ing of the responses since the e-mail had beensent was measured with three modalities: onebusiness day, two to four business days, and oneweek or more.

The first two research questions can beanswered by descriptive statistics on theresponse rates and timeliness for the two e-mailsthat were sent to the political actors included inthis study. The third research question addressesthe factors causing e-mail responsiveness andrequires a multivariate analysis. In order toperform it, the two response variables havebeen converted into dichotomies discriminatingbetween political actors that did and did notrespond to e-mails. These will constitute thedependent variables in two multivariate logisticregression models predicting responsiveness tothe issue question and volunteer pledge e-mails,respectively. In order to test the various hypothe-ses encompassed in RQ3, the model includesboth system-level and organization-levelvariables.

The system-level variables are as follows:

• Six dummy variables identifying the coun-tries where the political actors operate(France being the reference category) inorder to account for the role of any sys-temic characteristic not included in themodel

• Technological development, defined as thediffusion within the population of Internetaccess and broadband connections4

• The electoral system employed in the elec-tion, with majoritarian (first past the postplurality and run-off majority) systemscontrasted with proportional systems5

• Political participation, defined as the per-centage of voters who turned out in theelection (which also allows to distinguishbetween general elections, European elec-tions, and primaries, as voter turnout variessignificantly across these types of elec-tions)6

The organization-level variables are as fol-lows:

• The type of political actors analyzed,which differentiates between parties, pres-idential candidates, and party leaders (thelatter functioning as the reference category)

• Incumbency in national government (andin the majority supporting it in parliamen-tary systems)

• The percentage of votes achieved by eachparty, employed as a measure of itsresources7

• Organizational inclusiveness, measuredwith respect to leader and candidateselection, with normalized values rangingfrom the least (0) to the most (1) inclusivemethods8

• The ideological family to which the (candi-date’s or leader’s) party belongs,9 enteredas a set of dummy variables comparingeach ideological family with Christian-democrats and conservatives, which is thereference category

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 251

RESULTS

The first two research questions can beanswered by assessing how many parties andcandidates responded to the issue question andvolunteer pledge e-mails, and how long it tookfor them to do so. As Table 1 shows, polit-ical actors are rather unresponsive to e-mailinquiries. First, 6 in 10 e-mails went entirelyunanswered, which suggests that, in spite of theubiquity and mundane nature of e-mail as a com-munication tool, political organizations eitherlack or refuse to marshal the resources necessaryto deal with it in a way that matches citizens’expectations. Second, among the minority ofmessages that received a reply, about half hadto wait longer than one business day, which isnotable given that the requests were relativelysimple.

Even if they are somewhat related, the empir-ical correlation between responses to e-mailsasking issue questions and those pledging tovolunteer is not particularly strong: If we differ-entiate between political actors that respondedand those that did not, thus treating the twovariables as dichotomous, the φ coefficient mea-suring association between them is .497 (p =.000). Volunteer pledges are slightly more likelyto receive a response than e-mails asking infor-mation about policy issues (41% versus 35%)and to be answered within one business day(21% versus 16%). This finding is understand-able, since resource mobilization is a crucialcomponent of online campaigning, as most visi-tors of party Web sites are likely to be supportersof those parties (Bimber & Davis, 2003; Vaccari,2008). Political actors are more eager to respond

TABLE 1. Levels of Responsiveness to Issueand Volunteer E-Mails

Issue question(%)

Volunteer pledge(%)

No answer 65.1 58.9One week or more 3.9 3.3Two to four

business days15.5 17.1

One business day 15.5 20.7N 304 304

to messages that can result in new volunteers,donors, and message multipliers, than to thosethat may simply clarify an issue for a prospec-tive voter. However, the difference between theresponse rates to the issue question and volun-teer pledge e-mails is not statistically significant(φ = .064, p = .112 when the variables aretreated as dichotomous), suggesting that a clear-cut answer to RQ2 cannot be provided by thedata. The clearest finding is that the parties andcandidates in the seven countries included in thisstudy between 2007 and 2013 responded rarelyand rather slowly to e-mails. Given the widelyuneven levels of responsiveness that were foundand the fact that relatively few parties and can-didates replied to both e-mails, it is importantto understand which factors are related to suchresponsiveness. This point is addressed by RQ3,which I answered with logistic regression anal-yses predicting whether issue and volunteere-mails were responded to based on both sys-temic and organizational variables, as shown inTable 2.

The evidence presented in Table 2 suggeststhat system-level variables are less relevant thanorganization-level variables in predicting e-mailresponsiveness.10 The fact that there is no cor-relation between technological development ande-mail responsiveness is not surprising, giventhe mundane nature of e-mail as a communi-cation tool. The electoral system also does notcorrelate significantly with e-mail responsive-ness (although the coefficients are negative inboth models), and there is a small but statis-tically significant negative correlation betweenvoter turnout and responsiveness to both typesof e-mails. This might be a function of the factthat U.S. presidential primary candidates wereequally responsive to e-mails as other politicalactors included in this study, but the total turnoutof their races was substantially lower. Amongthe country dummy variables, Australian andGerman (with respect to issue e-mails) politi-cal actors were, all else being equal, more likelythan French ones to respond to e-mails. Sincethese variables were included solely to controlfor systemic factors that could not be accountedfor in the models, I will not speculate on thesefindings other than saying that institutional and

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

252 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

TABLE 2. Estimated Coefficients for Responsiveness to E-Mails with Issue Questions andVolunteer Pledges by Technological Development, Socio-Political Environment, and Party

Characteristics

Issue question Volunteer pledge

Coeff. s.e. Exp(b) Coeff. s.e. Exp(b)

Country (France = ref.)Australia 2.304∗ .760 10.013 2.000∗ .655 7.387Germany 2.098∗ .825 8.153 .859 .771 2.360Italy 1.480 1.002 4.391 1.720 .911 5.587Spain 1.419 .892 4.133 1.239 .846 3.453United Kingdom .477 .734 1.611 −.702 .750 .496United States −1.712 .917 .180 −.948 .738 .388Technological

development (%).018 .040 1.018 .019 .036 1.019

Majoritarian electoralsystem

−.618 .770 .539 −.262 .725 .770

Voter turnout (%) −.030∗ .013 .970 −.024∗ .012 .977

Type of political actor (party leader = ref.)Party 1.394∗∗ .504 4.029 1.851∗∗∗ .488 6.365Presidential candidate 3.060∗∗∗ .857 21.333 2.612∗∗∗ .745 13.625Incumbent in national

government−.560 .416 .571 −.352 .379 .704

Votes (%) .025∗ .012 1.025 .017 .012 1.017Organizational

inclusiveness (0–1).921 .725 2.512 .585 .669 1.796

Ideological family (Christian Democrat/conservative = ref.)Communist/radical left −.964 .642 .381 −.681 .532 .506Green 1.474∗ .615 4.368 1.542∗∗ .597 4.676Socialist/democratic 1.012∗ .472 2.750 .919∗ .459 2.506Left-Libertarian .183 .623 1.200 −.259 .579 .772Regionalist .548 .647 1.730 −.748 .659 .473Liberal −.217 .785 .805 .662 .691 1.939Radical right .203 .551 1.225 −.664 .531 .515Constant −2.700 2.322 .067 −2.666 2.142 .069

N 296 296Nagelkerke R2 .284 .232

Note: ∗∗∗p ≤ .001, ∗∗p ≤ .01, ∗p ≤ .05.

contextual factors should not be overlooked inthe study of parties’ uptake of digital media.

The results indicate that e-mail responsive-ness is mostly related to the characteristics ofpolitical organizations rather than the contextthey operate in. The variables that play thestrongest role are the difference between par-ties, presidential candidates and party leaders,ideology, and votes (employed as a proxy forresources). By contrast, incumbency and orga-nizational inclusiveness do not show any signif-icant correlation with the two types of respon-siveness predicted by the models, although thesigns of the coefficients are identical in thetwo models and are consistent with theoretical

expectations (negative for incumbency and pos-itive for inclusiveness).

Presidential candidates and parties are muchmore likely than party leaders to reply toboth types of e-mails. This difference can beattributed to the fact that presidential cam-paign committees and party organizations tendto be better structured than those overseeing theonline presence of party leaders in parliamen-tary systems, where elections are generally morecentered on parties. Especially, presidentialcandidates are less known than parties, partic-ularly during primaries, and more dependenton the resources that they gather during thecampaign, and thus have greater incentives to

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 253

respond to both issue inquiries and volunteerpledges.

The coefficients for ideological familieshighlight that both green and socialist anddemocratic parties are more responsive thanChristian-democratic and conservative oneswith respect to both types of e-mails. The prob-ability that both e-mails receive a reply is morethan four times as high for greens, and morethan twice as high for socialists and democrats,as it is for conservatives. Thus, among thetwo main political groupings that compete forgovernment leadership in most Western democ-racies, center–left forces appear to be moreresponsive to e-mails than their center–rightrivals. More broadly, progressive parties seemto be more responsive than all other parties, andextreme parties, both left-wing and right-wing,are among the least responsive, as shown bythe three negative coefficients out of four thatinvolved them.

Finally, vote percentages show a weak butsignificant correlation with responses to e-mailsfeaturing issue questions. Although this find-ing is consistent with the normalization theory,the correlation is not particularly strong, andthe coefficient related to the volunteer e-mail,albeit positive, is not significant. Larger polit-ical organizations are more likely to reply toe-mail inquiries about their issue stances, but,as can be inferred from the exp(b) coefficients,a 10% difference in the votes achieved by twoparties increases the probability that the largerone will respond by only 4% compared to thesmaller one, all else being equal. Moreover, thefact that no significant correlation was found forresponses to the volunteer pledge e-mail sug-gests that smaller political actors rationally pri-oritize addressing messages through which theycan recruit fresh volunteers, which may helpthem narrow the gap with larger competitors.

The findings presented in this article are lim-ited in various ways that should be acknowl-edged. First, I measured an outcome (responsesto e-mails or lack thereof) and employed statis-tical techniques to explain variance in it basedon various system- and organization-level fac-tors, but this approach can only shed light ongeneral patterns such as the ones discussed inthis paragraph. Equally valuable and necessaryare studies based on direct observation and

first-hand accounts of the functioning of polit-ical and campaign organizations, which canuniquely shed light on contextual factors thatare difficult to both measure and disentanglestatistically. In the case of e-mail, for instance,Karpf (2012), Kreiss (2012), and Vaccari (2008)discuss how organizations can scientifically testresponses to different messages in order to fine-tune their communication and obtain feedbackfrom their constituents, even if the latter are notaware of the process. Thus, e-mail may be aviable channel for two-way dialogue betweenpolitical actors and respondents in a passive andimplicit fashion rather than in the active andexplicit way that has been the focus of this study.Such use of e-mail may also be more consistentwith campaigns’ pressing need for time, control,and resource optimization, which often encour-ages them to avoid rather than embracing low-cost communication tools. Consistently, Nielsen(2009) has suggested that Internet activism oftenresults in “labors”—over-communication, mis-communication, and communicative overload—that increase the transaction costs for politi-cal organizations and thus impede rather thanenhance their functioning. In suggesting a ratherreluctant approach to e-mail responsivenessby most political actors, the findings of thisresearch reaffirm such considerations. Finally,even if this study employed regression analy-sis to assess multivariate correlations betweene-mail responsiveness and various independentvariables, the number of cases that could beincluded is rather limited in spite of the rela-tively long time-span of this research—a typicalcase of Lijphart’s (1971) “too few cases, toomany variables” conundrum. Although the mod-els yielded results that are both interesting andplausible, they should be treated with cautionand hopefully replicated with a larger number ofcases and across a broader realm of countries inorder to achieve more robust findings.

CONCLUSIONS

E-mail is ubiquitous in citizens’ everydaylives and is routinely used by political organi-zations for both internal and external communi-cation. As a channel of citizen-elite communi-cation, e-mail can enable politicians to bypass

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

254 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

the filter of the mass media and to show respon-siveness to voters’ demands. In spite of thesepremises, the evidence presented here suggeststhat most political actors across Western democ-racies in national elections between 2007 and2013 failed to respond to the majority of e-mailinquiries they received.

This dismal performance can be explainedby political organizations’ need to carefully pri-oritize their scarce resources. Answering largenumbers of personal messages that demand indi-vidual consideration requires time and man-power, which any campaign is hard-pressed toemploy as parsimoniously and effectively aspossible. Such imperative should be paramountin the final phases of election campaigns—which is when the data presented here werecollected—because the pressure and constraintson the available resources are highest then.In choosing whether or not to direct hardlyabundant assets to answering citizens’ e-mails,campaigns may employ two sets of tactical con-siderations.

First, parties and campaigns may valueknown over unknown constituents, using e-mailpreferably to follow up on personal conversa-tions in which they have already establishedtrust and rapport with supporters. For instance,in his ethnographic research on U.S. democraticcongressional campaigns, Nielsen notes that,once a volunteer had been recruited, interactionswere more likely to occur over e-mail than overthe phone or in person. As he notes, “The advan-tages are clear for both staffers and volunteers:an e-mail can spell out details about locationsand time more clearly than a rushed phone callfrom a busy campaign office, they can be copiedand pasted and sent quickly to many, and theydo not require that the receiver take an incom-ing call here and now or is physically present”(Nielsen, 2011, p. 763). Political organizationsmay thus be more responsive to e-mails fromcitizens with whom they have already estab-lished a personal connection than to e-mailsfrom strangers. Since the e-mails sent for thisresearch came from people that the addressedorganizations did not know, they may have beentreated with less attention.

Second, party and campaign staffers mayplace a premium on those interactions that canbe decisive for the completion of required tasks

and the meeting of upcoming deadlines overthose that can also be accomplished by othermeans. In trying to cope with scarce resources,chief among them time, they may reason that,if a voter is motivated enough to write to them,she or he may also be perseverant enough tofind the relevant issue information on the—equally accessible, or mundane in Nielsen’s(2011) definition—institutional Web site, and toparticipate in the campaign by clicking on the“get involved” buttons and creating an accounton the digital organizing hubs that are by nowcustomary on these platforms. From a hard-pressed campaign’s perspective, answering e-mails such as the ones that were sent for thisresearch may thus not be decisive to informand engage its online audiences. As a result,political actors may attribute lower priority toanswering these e-mails and may instead focuson those inquiries that involve tasks that cannotbe completed otherwise.

Besides showing that parties’ and candi-dates’ e-mail responsiveness is rather lowoverall, this research has also shed light onsome causal determinants of such phenomenon.First, political actors are slightly more likelyto respond, and to do so quickly, if the e-mail contains a volunteer pledge as opposedto an issue question. This might be a ratio-nal calculation—particularly for less resourcefulparties—because a voter who takes the time towrite an e-mail to a party or candidate ask-ing an issue question should already be in orclose to their tent given the effort she dedicatedto finding and contacting them online (Norris2003), and online campaigns have been knownto prioritize volunteer recruitment and sup-porter mobilization for these reasons (Bimber& Davis, 2003). However, the fact that responserates to volunteer pledges were only marginallyhigher—and not to an extent that was statisti-cally significant—than those to issue questionssuggests that political actors place limited valueon e-mail as a recruitment tool as well, or atleast not enough value to justify the investmentneeded to promptly respond to all messages.

Regarding organizational factors, parties andpresidential candidates respond more effi-ciently to incoming e-mails most likely becausethey tend to have larger staffs and morerobust routines than party leader campaigns in

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 255

parliamentary systems. The fact that larger par-ties answer issue questions more than smallerones suggests that offline inequalities may trans-late into political actors’ online responsiveness.However, the relative weakness of this corre-lation, and the fact that it was not significantfor volunteer pledges, also suggest that e-mailmay play a leveling role in political competition.Minor parties seem to be just as eager as majorones to harness its affordances for volunteermobilization.

Regarding ideology, the greater responsive-ness shown by progressive parties may berelated to their organizational structures andelite preferences, as well as contextual fac-tors that vary from country to country. Greenparties are rooted in grass-roots organizationalparadigms that value informal relationshipsamong members and identifiers over formalties and structures (Löfgren & Smith, 2003,p. 49). Accordingly, they tend to be at easewith the decentralizing affordances offered bythe Internet, of which engagement with support-ers over e-mail constitutes an important com-ponent, not least due to its mundane nature.By the same token, European socialist par-ties have been historically associated with themass-party model, focused on recruiting a vastmembership and enabling participation amongthem. Parties that have inherited this legacymay thus be expected to rely on the Internetmostly to disseminate information from thecenter to the periphery and to mobilize sup-porters when needed (Löfgren & Smith, 2003,pp. 45–46), functions which correspond to thetypes of e-mails these parties were more likelyto answer. These considerations do not apply tothe U.S. case, where party organizations havebeen much looser and campaign-focused onboth sides of the spectrum. However, contex-tual factors, such as the Democrats’ status asthe opposition party at the crucial time in whichonline campaigning was shaped and developedinto its present form (Karpf, 2012), have beensuggested as explanations for the Americanleft’s upper hand on the Internet over the lastdecade.

The findings presented in this study suggestvarious directions for future research. First, it

is an open question whether the patterns andcorrelations that were found during electioncampaigns are reproduced in “peacetime” peri-ods outside of elections. In these contexts, citi-zens’ demand for information and participationshould be feebler and political actors’ responsecapacity should thus be placed under lower pres-sure, but their interest in showing responsivenessshould also be weaker due to the lack of imme-diate electoral incentives. Second, the method-ology of this study could be usefully extendedto study e-mail responsiveness by other typesof political actors, such as social movements,interest groups, parliamentary representatives,and local elected officials. Third, social mediaopen many interesting opportunities for polit-ical actors to respond to individual voters’demands, as most citizens with Internet connec-tions have personal accounts on at least one ofthese platforms. Interactions between politiciansand citizens on social media occur in semi-public rather than private environments, so theiraudiences—and, thus, political implications—are potentially broader than those of privatee-mail exchanges. Finally, the causal patternssuggested here should be tested, and possiblyexpanded, by qualitative studies that focus onthe contextual conditions and real-world func-tioning of political organizations in dealing withe-mails and other types of interactions withcitizens.

This study has shown that, in spite of themundane nature of a technological artifact suchas e-mail, there is meaningful variation in itsuses among different types of political actors.As information technology allows parties andcandidates to engage with citizens in an increas-ingly complex and multilayered set of envi-ronments, it is important to empirically scruti-nize the uses of different tools and applicationsamong various types of subjects and organi-zations. Although digital politics appears tobe in a permanent state of flux, more endur-ing strategic considerations and internal con-straints shape the ways in which political actorsaddress these opportunities—or, in the case ofthe many e-mails sent for this research that wentunanswered and, perhaps, unnoticed, fail todo so.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

256 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

NOTES

1. For instance, the Pirate Party was not includedin the analysis of the 2009 German elections, but itwas included in subsequent analyses of the Australian,German, and Italian 2013 elections.

2. The fact that responses from some political actorswere analyzed more than once does not severely violatethe assumption, implicit in the regression models that willbe shown below, that cases must be fully independent.Although the same organization might respond (or failto do so) similarly to different e-mail inquiries in a shorttimeframe, it is plausible that its behavior might vary ina longer timeframe, for instance between a general anda European election, between two general elections, andbetween a primary and a general election.

3. From an ethical standpoint, concealing theresearch purpose of e-mail exchanges such as those stud-ied here can be justified because this activity did notviolate any individual rights of the subjects. First, thefocus was not on the individuals who responded to e-mailsor failed to do so, but on the organizations they workedfor. Second, no personal information, let alone sensitivedata, was solicited or collected through the e-mails sentand received for this research. Third, the content of theresponses was not recorded beyond assessing whetherthey addressed the question and had been written by ahuman being.

4. Technological development is measured as theaverage of the number of Internet users per 100 inhabi-tants and the number of broadband users per 100 inhab-itants. Both measures come from the InternationalTelecommunications Union (http://www.itu.int/ITU-D/ict). The index varies between 0 and 100. Cronbach’s α

measuring internal consistency of these indicators equals.773. Since at the time of this writing the latest avail-able data were from 2012, I used 2012 figures for thecountries that I studied in 2013 (Australia, Germany, andItaly).

5. This variable is sensitive to the fact that in Franceand the United Kingdom, general elections are held withmajoritarian systems, but European Parliament electionsemploy proportional systems.

6. As is well known, voting is the most commonform of political participation in Western democracies.Voter turnout data were obtained from official Web sitesof government institutions in each country. For U.S. elec-tions, turnout was calculated at the national level based onthe data available on http://elections.gmu.edu.

7. In parliamentary democracies where I analyzedparty leaders, I imputed to the leaders the votes achievedby their parties. This strategy is justified by the fact thatthis measure was employed as a proxy for resources, andleaders of major parties tend to be better resourced thanthose of minor ones. Although it would have been prefer-able to employ actual measures of financial resources (aswas done in Vaccari, 2013), these were not yet available

for some elections at the time of this writing. Many otherstudies of digital politics have employed votes as proxiesfor resources to overcome this limitation (see Strandberg,2008).

8. All parties and candidates were classified basedon the criteria that they (or, in the case of candidatesand leaders, their parties) employ to select their lead-ers and parliamentary candidates based on the categoriessuggested by Janda (1980). For leader selection meth-ods, Janda proposes eight categories, ranging in terms ofinclusiveness from selection by vote of party identifiersor supporters to selection by the leader’s predecessor. Forcandidate selection methods, the categories are nine, fromvote of local party supporters in an open primary to selec-tion by a national committee or party council (see Vaccari,2013). The variable included in the model averages thevalues for leader and candidate selection (Cronbach’s α =.571). Models run with two separate variables for leaderand candidate selection methods yielded substantiallyidentical results as the ones presented here.

9. Parties and candidates were classified based onthe taxonomy of party ideologies suggested by Vassalloand Wilcox (2006), to which I added the category ofregionalist parties.

10. Nagelkerke R2 coefficients for models includingonly the system-level independent variables are .110 forthe issue e-mail and .057 for the volunteer pledge.

REFERENCES

Anstead, N., & Chadwick, A. (2009). Parties, electioncampaigning and the Internet: Toward a comparativeinstitutional approach. In A. Chadwick & P. Howard(Eds.), The Routledge handbook of Internet politics (pp.56–71). Abingdon, UK: Routledge.

Bimber, B., & Davis, R. (2003). Campaigning on line:The Internet in US elections. Oxford, UK: OxfordUniversity Press.

Coleman, S. (2005). The lonely citizen: Indirect represen-tation in an age of networks. Political Communication,22(2), 197–214. doi: 10.1080/10584600590933197

Coleman, S., & Blumler, J. G. (2009). The Internet anddemocratic citizenship: Theory, practice and policy.Cambridge, UK: Cambridge University Press.

Dalton, R. J. (2004). Democratic challenges, democraticchoices: The erosion of political support. Oxford, UK:Oxford University Press.

Davis, R. (2009). Typing politics: The role of blogs inAmerican politics. Oxford, UK: Oxford UniversityPress.

Entman, R. M. (1990). Democracy without citizens: Mediaand the decay of American politics. New York, NY:Oxford University Press.

Eurostat. (2013). Information society. RetrievedDecember 16, 2013, from: http://epp.eurostat.ec.europa.eu/portal/page/portal/information_society/introduction.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

Vaccari 257

Gibson, R. K., Margolis, M., Resnick, D., & Ward, S. J.(2003). Election campaigning in the WWW in the USAand UK: A comparative analysis. Party Politics, 9(1),47–75. doi: 10.1177/135406880391004

Green, D. P., & Gerber, A. S. (2004). Get out the vote! Howto increase voter turnout. Washington, D.C.: BrookingsInstitution Press.

Gulati, G., & Williams, C. (2007). Closing the gap,raising the bar: Candidate Web site communica-tion in the 2006 campaign for Congress. SocialScience Computer Review, 25(4), 443–465. doi:10.1177/0894439307305624

Howard, P. N. (2006). New media campaigns and the man-aged citizen. Cambridge, UK: Cambridge UniversityPress.

Jackson, N. (2004). Email and political campaigning:the experience of MPs in Westminster. Journal ofSystemics, Cybernetics and Informatics, 2(5), 1–6.

Janda, K. (1980). Political parties: A Cross-national sur-vey. New York: The Free Press.

Johnson, K., & Krueger, B. S. (2012). Analyzing the capac-ity of unsolicited political email. In C. R. Livermore(Ed.), E-politics and organizational implications ofthe Internet: Power, influence, and social change (pp.220–244), Hershey, PA: IGI Global.

Karpf, D. (2012). The MoveOn effect: The unexpectedtransformation of American political advocacy. Oxford,UK: Oxford University Press.

Klotz, R. J. (2007). Internet campaigning for grassroots andAstroturf support. Social Science Computer Review,25(1), 3–12. doi: 10.1177/0894439306289105

Kreiss, D. (2012). Taking our country back: The craft-ing of networked politics from Howard Dean to BarackObama. Oxford, UK: Oxford University Press.

Krueger, B. S. (2006). A comparison of con-ventional and Internet political mobilization.American Politics Research, 34(6), 759–776. doi:10.1177/1532673X06290911

Lijphart, A. (1971). Comparative politics and the compara-tive method. American Political Science Review, 65(3),682–693.

Lilleker, D., Koc-Michalska, K., Schweitzer, E. J.,Jacunski, M., Jacskon, N. A., & Vedel, T. (2011).Informing, engaging, mobilizing or interacting:Searching for a European model of Web campaigning.European Journal of Communication, 26(3), 195–213.doi: 10.1177/0267323111416182

Livingstone, S. (2004). Media literacy and the chal-lenge of new information and communication tech-nologies. Communication Review, 7, 3–14. doi:10.1080/10714420490280152

Löfgren, K., & Smith, C. (2003). Political parties anddemocracy in the Information Age. In R. K. Gibson,P. Nixon, & S. J. Ward (Eds.), Political parties andthe Internet: Net gain? (pp. 39–52). London, UK:Routledge.

Malhotra, N., Michelson, M. R., & Valenzuela, A. A.(2012). Emails from official sources can increaseturnout. Quarterly Journal of Political Science, 7(3),321–332. doi: 10.1561/100.00011073

Margolis, M., & Resnick, D. (2000). Politics asusual: The cyberspace ‘revolution.’ London, UK:Sage.

Nickerson, D. W. (2007). Does email boost turnout?Quarterly Journal of Political Science, 2(4), 369–379.doi: 10.1561/100.00007032

Nielsen, R. K. (2009). The labors of Internet-assistedactivism: Overcommunication, miscommunication, andcommunicative overload. Journal of InformationTechnology & Politics, 6(3–4), 267–280. doi:10.1080/19331680903048840

Nielsen, R. K. (2011). Mundane Internet tools, mobilizingpractices, and the coproduction of citizenship in politi-cal campaigns. New Media & Society, 13(5), 755–771.doi: 10.1177/1461444810380863

Nielsen, R. K., & Vaccari, C. (2013). Do people ‘like’candidates on Facebook? Not really: Large-scale directcandidate-to-voter online communication as an outlierphenomenon. International Journal of Communication,7, 2333–2356. doi: 1932-8036/20130005

Nixon, P., Ward, S. J., & Gibson, R. K. (2003).Conclusions: The Net change. In R. K. Gibson, P.Nixon, & S. J. Ward (Eds.), Political parties andthe Internet: Net gain? (pp. 234–243). London, UK:Routledge.

Norris, P. (2003). Preaching to the converted? Pluralism,participation and party Websites. Party Politics, 9(1),21–45. doi: 10.1177/135406880391003

Norris, P. (2011). Democratic deficit: Critical citi-zens revisited. New York, NY: Cambridge UniversityPress.

Purcell, K. (2011). Search and email still top the list ofmost popular online activities. Retrieved October 16,2012, from: http://www.pewinternet.org/Reports/2011/Search-and-email.aspx.

Sartori, G. (1976). Parties and party systems: A frameworkfor analysis. Cambridge, UK: Cambridge UniversityPress.

Seawright, J., & Gerring, J. (2008). Case selection tech-niques in case study research: A menu of qualitativeand quantitative options. Political Research Quarterly,61(2), 294–308. doi: 10.1177/1065912907313077

Stranbderg, K. (2008). Online electoral competition indifferent settings: A comparative meta-analysis ofthe research on party websites and online elec-toral competition. Party Politics, 14(2), 223–44. doi:10.1177/1354068807085891

Stromer-Galley, J., & Foot, K. A. (2002). Citizen percep-tions of online interactivity and implications for polit-ical campaign communication. Journal of Computer-Mediated Communication, 8(1). doi: 10.1111/j.1083-6101.2002.tb00161.x

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014

258 JOURNAL OF INFORMATION TECHNOLOGY & POLITICS

Vaccari, C. (2008). From the air to the ground:The Internet in the 2004 US presidential cam-paign. New Media & Society, 10(4), 647–665. doi:10.1177/1461444808093735

Vaccari, C. (2010). “Technology is a commodity”: TheInternet in the 2008 United States presidential election.Journal of Information Technology & Politics, 7(4),318–339. doi: 10.1080/19331681003656664

Vaccari, C. (2013). Digital politics in Western democ-racies: A comparative study. Baltimore, MD: JohnsHopkins University Press.

Vassallo, F., & Wilcox, C. (2006). “Party as a car-rier of ideas.” In R. S. Katz & W. Crotty (Eds.),Handbook of party politics (pp. 413–421). London, UK:Sage.

Vergeer, M., Hermans, L., & Sams, S. (2013). Online socialnetworks and micro-blogging in political campaigning:

The exploration of a new campaign tool and a newcampaign style. Party Politics, 19(3),477–501. doi:10.1177/1354068811407580

Ward, S. J. (2008). Introduction: Parties and election cam-paigning online: A new era? In S. J. Ward, D. M. Owen,R. Davis, & D. Taras (Eds.), Making a difference: Acomparative view of the role of the Internet in elec-tion politics (pp. 1–13). Lanham, MD: Rowman &Littlefield.

Ward, S. J., & Vedel, T. (2006). Introduction: The potentialof the Internet revisited. Parliamentary Affairs, 59(2),210–225. doi: 10.1093/pa/gsl014

Ward, S. J., & Gibson, R. K. (2009). European politicalorganizations and the Internet: Mobilization, partici-pation, and change. In A. Chadwick & P. Howard(Eds.), The Routledge handbook of Internet politics (pp.25–39). London, UK: Routledge.

APPENDIXNumber of Parties and Candidates/Leaders Contacted in Each Year, 2007–2013

Year Total

2007 2008 2009 2010 2011 2012 2013 Contacts Actors

Country P C/L P C/L P C/L P C/L P C/L P C/L P C/L P C/L P C/L

Australia 13 6 16 4 9 4 38 14 20 11France 9 13 13 1 15 9 37 23 22 20Germany 14 7 10 0 24 7 10 7Italy 16 0 16 0 22 8 54 8 35 8Spain 9 3 10 0 11 2 30 5 11 4UK 8 0 10 3 18 3 10 3USA 3 19 5 16 8 35 5 28Total 22 19 28 22 61 8 26 7 11 2 20 25 41 12 209 95 113 81

Note: Each e-mail contact included both an issue question and a volunteer pledge; for Germany in 2009, the analyses wereconducted both during the European (seven parties) and federal elections (seven parties and seven party leaders); for theU.S. in 2008 and 2012, the analyses were conducted during both the presidential primaries and general elections (in 2008,15 candidates were contacted during the primaries, and four candidates and three parties during the general election; in2012, 11 candidates were contacted during the primaries, and five candidates and five parties during the general election).

Dow

nloa

ded

by [

Uni

vers

ity o

f T

enne

ssee

, Kno

xvill

e] a

t 10:

38 1

5 Se

ptem

ber

2014