Skip to main content Accessibility help
×
×
Home

Information:

  • Access
  • Cited by 4
  • Cited by
    This article has been cited by the following publications. This list is generated based on data provided by CrossRef.

    Kruikemeier, Sanne Van Noort, Guda and Vliegenthart, Rens 2016. The Effect of Website Interactivity on Political Involvement. Journal of Media Psychology, Vol. 28, Issue. 3, p. 136.

    Sandberg, Linn A. C and Öhberg, Patrik 2017. The role of gender in online campaigning: Swedish candidates’ motives and use of social media during the European election 2014. Journal of Information Technology & Politics, Vol. 14, Issue. 4, p. 314.

    Furner, Christopher P. Racherla, Pradeep Babb, Jeffry and Zinko, Robert 2018. Optimizing Current Practices in E-Services and Mobile Applications. p. 114.

    Dai, Yue and Walther, Joseph B 2018. Vicariously Experiencing Parasocial Intimacy with Public Figures Through Observations of Interactions on Social Media. Human Communication Research, Vol. 44, Issue. 3, p. 322.

    ×

Actions:

      • Send article to Kindle

        To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Interactivity between Candidates and Citizens on a Social Networking Site: Effects on Perceptions and Vote Intentions
        Available formats
        ×

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Interactivity between Candidates and Citizens on a Social Networking Site: Effects on Perceptions and Vote Intentions
        Available formats
        ×

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Interactivity between Candidates and Citizens on a Social Networking Site: Effects on Perceptions and Vote Intentions
        Available formats
        ×
Export citation

Abstract

Voters and political candidates increasingly use social networking sites (SNSs) such as Facebook. This study uses data from an online posttest-only experiment (N = 183) in analyzing how exposure to supportive or challenging user comments on a fictional candidate's Facebook page influenced participants’ perceptions of and willingness to vote for the candidate, as well as whether candidate replies to each type of user comments affected these outcomes. Participants who viewed a page with supportive comments and “likes” reported more favorable perceptions of and greater support for the candidate, relative to participants who viewed a page with challenging comments. Thus, the appearance of interactivity between a candidate and other users on the candidate's Facebook page can shape the responses of those viewing the page. However, exposure to candidate replies to either supportive or challenging comments did not lead to significantly more favorable perceptions or a greater likelihood of voting for the candidate.

INTRODUCTION

The use of online SNSs among the U.S. public has increased dramatically in recent years. As of December 2012 (shortly after the present study was conducted), fully 67% of online U.S. adults reported using Facebook, the most popular SNS (Pew Research Center's Internet & American Life Project, 2014). Thus, it is no surprise that candidates for public office increasingly use SNSs, including Facebook, in their campaigns. In 2008, presidential candidate Barack Obama made extensive use of Facebook and other SNSs in his victorious campaign (see, e.g., Cogburn and Espinoza-Vasquez, 2011; Haynes and Pitts, 2009; Johnson and Perlmutter, 2010). At present, it is typical for candidates at all levels of government to create their own Facebook pages.

A number of recent studies have investigated the role of SNSs, such as Facebook in political campaigns. Some of this research analyzes the content of messages from and about candidates on SNSs (e.g., Fernandes et al., 2010; Sweetser and Lariscy, 2008; Woolley et al., 2010). Other studies examine relationships between citizens’ use of SNSs and their political attitudes and/or behaviors, including political cynicism (Hanson et al., 2010), political involvement and self-efficacy (Kushin and Yamamoto, 2010), and political participation (Towner, 2013; Zhang et al., 2009). Still other research tests whether message features shape the effects of candidate communication through SNSs on a range of outcomes, including vote intentions (Lee and Jang, 2013; Lee and Oh, 2012; Lee and Shin, 2012b).

To date, however, relatively little research has directly tested the possible effects of observing interactivity between citizens and candidates through SNSs (for exceptions, discussed below, see Lee and Shin, 2012a; Utz 2009). Thus, existing research largely neglects one of the most distinctive features of SNSs, as compared to other forms of web campaigning: the extent to which they allow for public conversations between campaigns and potential voters. For example, Facebook allows users to post comments on other users’ profile pages as well as to “like” posts by other users. Thus, citizens can and do provide publicly visible feedback on candidates’ Facebook pages. Candidates, in turn, may choose to reply to user comments with their own comments. To provide what is, to our knowledge, the first attempt to explore the effects of witnessing such interactions on Facebook, this study develops a theoretical framework based on the warranting principle, which posits that people place particular weight on information generated by those other than the target of that information in forming impressions online (Walther and Parks, 2002). It then analyzes data from an online posttest-only experiment (N = 183) to test how exposure to supportive or challenging user comments on a fictional candidate's Facebook page––and candidate replies to each type of comment––influenced participants’ perceptions of and willingness to vote for the candidate. By addressing one of the key features of SNSs, the potential for public interaction between candidates and users, the study seeks to provide a better foundation for further research on their use in the campaign context. In doing so, it also speaks to the potential efficacy of user-generated comments and candidate-generated replies in political campaigning through social media.

SOCIAL NETWORKING SITES, INTERACTIVITY, AND CANDIDATE EVALUATIONS

For more than half a century, observers have proposed the potential for “bandwagon” effects on public opinion, in which processes such as “social facilitation” (Allport, 1924) lead citizens “to behave as (they) see others behave” (Allport, 1940, 250). More recent accounts have drawn on the concept of “low-information rationality” to argue that voters tend to rely on information shortcuts, including interpersonal cues, to evaluate political candidates (e.g., Popkin, 1991). Consistent with this, a sizable body of research demonstrates interpersonal influence on candidate impressions and vote choice in offline settings (e.g., Huckfeldt and Sprague, 1995; Mutz, 2006).

A growing body of the literature also demonstrates that SNSs can influence how individuals form impressions of others (for an overview, see Wilson et al., 2012). For example, information on a person's Facebook profile, such as the number of “friends,” can influence how users evaluate that person (e.g., Tong et al., 2008). Of particular relevance for the purposes at hand, a few recent studies suggest that observing interactivity between political candidates and other users through SNSs can influence candidate impressions. Building on previous findings that the presence of interactive features on candidate websites influenced outcomes such as information recall (Warnick et al., 2006) and information efficacy (Tedesco, 2007), Utz (2009) showed that voters exposed to a profile (on Hyves, a Dutch SNS) for a fictional candidate reported more favorable perceptions when the candidate was presented as responding to user comments than when the candidate was not presented as doing so. In a similar vein, Lee and Shin (2012a) demonstrated that exposure to a candidate's page on Twitter (another SNS) induced a greater sense of conversation with the candidate when it featured high interactivity (through active responses to followers’ questions) than when it featured low interactivity. This greater sense of “social presence,” in turn, led to more favorable evaluations and vote intentions regarding the candidate.

The warranting principle, developed in research on computer-mediated communication, suggests that observing comments from SNSs users other than the candidate may be especially powerful in shaping candidate impressions. This principle posits that individuals typically rely on information with greater warranting value in forming impressions online and that the warranting value of information online is “derived from the receiver's perception about the extent to which the content of that information is immune to manipulation by the person to whom it refers” (Walther and Parks, 2002, 552). Thus, individuals should tend to see information generated by sources other than the target of that information as relatively immune to manipulation and, thus, more credible whereas they should tend to see information generated by the target as more self-interested and, thus, less credible (Walther et al., 2009). Though initially developed in the context of other forms of computer-mediated communication, recent research has demonstrated that the warranting principle applies in the context of SNSs. Several studies show that information provided about a person by other SNS users can influence impressions of that person (Utz, 2010; Walther et al., 2008). Furthermore, Walther et al. (2009) found that “other-generated” (friends’) comments on an individual's Facebook page influenced viewers’ impressions of that individual more than did (the individual's own) “self-generated” comments.

Thus far, research has not applied the warranting principle in developing hypotheses about how other- and self-generated comments on SNSs may influence impressions of political candidates. With this in mind, the goals of the present study are to explore whether, and if so how, (1) supportive and challenging “other-generated” comments on a candidate's Facebook page influence perceptions of and willingness to vote for that candidate, and (2) candidate “self-generated” replies to such comments influence the same outcomes. The first set of hypotheses builds on the premise, derived from the warranting principle, that other-generated information on a SNS profile should provide credible, and thus influential, cues for forming impressions (Walther and Parks, 2002; Walther et al., 2008; 2009). In the case at hand, observing supportive comments and “likes” should lead to more favorable candidate evaluations, whereas observing challenging comments should lead to more negative evaluations.

  1. H1A: Compared to individuals exposed to a candidate Facebook profile with no comments or likes, those exposed to a candidate Facebook profile with supportive comments and likes from other users will view the candidate more positively and be more likely to vote for the candidate.

  2. H1B: Compared to individuals exposed to a candidate Facebook profile with no comments or likes, those exposed to a candidate Facebook profile with challenging comments (and no likes) from other users will view the candidate more negatively and be less likely to vote for the candidate.

  3. H1C: Compared to individuals exposed to a candidate Facebook profile with supportive comments and likes, those exposed to a candidate Facebook profile with challenging comments (and no likes) will view the candidate more negatively and be less likely to vote for the candidate.

The findings from previous studies of candidate responses to user interactivity on SNS also suggest that observing candidate-generated replies to user comments may lead to more favorable impressions of the candidate (Lee and Shin, 2012a; Utz, 2009). Thus, the present study tests the following hypotheses:

  1. H2: Exposure to candidate replies to supportive comments on Facebook will lead to more positive perceptions of the candidate and a greater likelihood of voting for the candidate.

  2. H3: Exposure to candidate replies to challenging comments on Facebook will lead to more positive perceptions of the candidate and a greater likelihood of voting for the candidate.

On the other hand, the warranting principle suggests that candidates’ self-generated comments may do less to shape impressions than will other-generated comments given that users may see the former as more open to manipulation, and hence less credible, than the latter (Walther et al., 2009).

METHODS

This study used data from an online between-subjects, posttest-only experiment conducted from October 19 to November 5, 2012. Participants were Delaware residents who were Internet users and had been recruited previously for a research panel through two telephone surveys (N = 1,215). All panelists for whom a valid e-mail address was available (N = 660) were e-mailed a link to a Qualtrics survey that included the experiment. Of the 183 participants who completed the experiment, 47% were women and 53% were men; 4% self-identified as African American; and 1% self-identified as Hispanic. In terms of party, 43% identified as Democrats, 25% identified as Republicans, and 32% identified as “Independent or other.” Median education level was college graduate (73% had a four-year degree or higher), and median age was 61 (42% were 65 or older). 1 The sample roughly mirrored the state population from which it was drawn in terms of gender and partisanship, though it was less representative in terms of race, ethnicity, education, and age. 2

The nature of the sample raises potential concerns about the external validity of the results; in particular, the median age of the sample does not reflect the median age of Facebook users, who are disproportionately young (Pew Research Internet Project, 2014). However, 63% of the participants reported that they used Facebook, only slightly less than the percentage (67%) among a nationally representative December 2012 survey of Internet users (Pew Research Internet Project, 2014). When the analyses were restricted to Facebook users only, results substantively similar to the ones described below emerged at p < 0.10 (except for two marginal effects, noted below). 3

Instructions at the beginning of the survey told participants, “For this study, you will be asked to look at the Facebook page of a candidate for the U.S. House of Representatives. You will then be asked a series of questions.” 4 The next page of the survey presented a (static) screen capture image of a Facebook page for a fictional candidate (see Appendix A). The design of the page was modeled on real candidate pages. The candidate's name, Dennis Felton, was selected by randomly choosing a first name from a list of 2012 political candidates and then randomly choosing a last name from the same list (the same name was used across all conditions, for all participants). 5 At the top of the page was photograph of the candidate (the “profile picture”) against a background image of the U.S. flag (the “wallpaper”). 6 Below this were two posts from the candidate, each of which was typical of real candidate posts and neither of which included any partisan or policy content. The first post included a photograph of the candidate speaking to voters at a restaurant along with text discussing the event. The second post included a photograph of a campaign t-shirt along with text soliciting a donation. The page included no information about the candidate's district or party identification. 7

The page image was edited to manipulate the presence and tone of comments from fictional Facebook users, as well as the presence of candidate replies to those comments (see Appendix B for wording). The comments and replies were modeled on ones from real candidate Facebook pages; the replies thanked the commenter, promoted the candidate in general terms, directed the commenter to contact the candidate for more information, and solicited the commenter's vote. 8 Participants were randomly assigned to view one of five versions of the page:

  • A page that included no comments on the candidate posts and served as a baseline (N = 29; hereafter, No Comments).

  • A page that included a supportive comment for each candidate post, as well as 8 “likes” for the first post and 3 for the second (N = 29; Supportive Comments). 9

  • A page that included a challenging comment for each candidate post, with no “likes” for either post (N = 30; Challenging Comments).

  • A page identical to the one that included supportive comments except that it also included a candidate reply to each comment (N = 32; Supportive Comments + Candidate Replies).

  • A page identical to the one that included challenging comments except that it also included a candidate reply to each comment; replies here were identical to replies for the supportive comments (N = 33; Challenging Comments + Candidate Replies).

There was also a sixth condition (N = 30), in which participants saw a website for the candidate (following typical candidate website format) that was identical in substance to the Facebook page with no comments. Results for this condition did not differ significantly from the results for the No Comments condition. Given that the study at hand focuses on the effects of exposure to Facebook interactivity, rather than of exposure to Facebook versus a website, the analyses presented here exclude the website condition.

After viewing the page, respondents answered questions about a variety of topics. Based on a series of factor analyses and reliability analyses, responses to the items of interest were used to construct three measures, each coded to range from 0 to 3 (see Appendix C for wording). The first was an 8-item index measuring perceptions of candidate traits (coded so that higher values indicated more positive perceptions; M = 1.28; SD = 0.73; α = 0.96). The second was a four-item index measuring perceptions of candidate motives (coded similarly; M = 1.69; SD = 0.56; α = 0.76). The third was a single-item measure for likelihood of voting for the candidate (coded so that higher values indicated greater likelihood of doing so; M = 1.05; SD = 0.79). 10

RESULTS

The analyses focused on comparing means for each of these three measures across the experimental conditions (see Table 1). 11 A series of t-tests were used to test the hypotheses.

Table 1 Candidate Evaluations and Vote Intentions, by Experimental Condition

Note. Table entries are means; standard errors are in parentheses.

H1A predicted that participants in the Supportive Comments condition would view the candidate more positively and be more likely to vote for the candidate, relative to participants in the No Comments condition. The contrast across these two conditions was not statistically significant for perceptions of candidate traits (t = 0.95; p = n.s.) or candidate motives (t = 1.22; p = n.s.) but was significant and in the expected direction for vote intentions (t = 2.66; p < 0.01). 12 Thus, supportive comments (versus no comments) exerted a discernible positive effect on likelihood of voting for the candidate but not on perceptions of the candidate.

H1B, in turn, predicted that participants in the Challenging Comments condition would view the candidate more negatively and be less likely to vote for the candidate, relative to participants in the No Comments condition. A significant difference across the two conditions did not emerge for perceptions of candidate traits (t = 1.32; p = n.s.) or vote intentions (t = −0.52; p = n.s.), though a marginal difference emerged in the expected direction for perceptions of candidate motives (t = 1.78; p = 0.07; this difference was not significant at p < 0.10 when the sample was restricted to Facebook users only). In short, challenging comments (versus no comments) may have exerted a negative impact on perceptions of candidate motives but did not discernibly influence perceptions of candidate traits or likelihood of voting for the candidate.

Whereas H1A and H1B focused on comparisons to the No Comments baseline, H1C predicted that participants in the Supportive Comments condition would view the candidate more positively and be more likely to vote for the candidate than participants in the Challenging Comments condition. Here, the differences across conditions for perceptions of candidate traits (t = 2.30; p < 0.05), perceptions of candidate motives (t = 3.02; p < 0.01), and vote intentions (t = 2.18; p < 0.05) were all significant and in the expected direction. Furthermore, in each case the difference was almost half a point on a 0–3 scale: 0.46 for traits, 0.44 for motives, and 0.45 for vote intentions. Thus, exposure to supportive comments versus challenging comments exerted substantial effects on perceptions of and likelihood of voting for the candidate.

H2 focused on whether the inclusion of candidate replies to supportive comments would alter perceptions of and willingness to vote for the candidate. Here, the key comparison was between the Supportive Comments and Supportive Comments + Candidate Replies conditions. Perceptions of candidate traits did not differ significantly across these two conditions (t = −0.72; p = n.s.). However, participants in the Supportive Comments + Candidate Replies condition reported marginally more negative perceptions of candidate motives (t = −1.88; p = 0.06; this difference was not significant at p < 0.10 when the sample was restricted to Facebook users only) and a significantly lower likelihood of voting for the candidate (t = −2.67; p < 0.01) than did those in the Supportive Comments condition. Thus, the candidate replies produced no discernible improvement in candidate perceptions while actually reducing candidate support.

Finally, H3 focused on whether the inclusion of candidate replies to challenging comments produced any effects. This time, the key comparison was between the Challenging Comments and Challenging Comments + Candidate Replies conditions. No significant differences between these conditions emerged for perceptions of candidate traits (t = 0.59; p = n.s.), perceptions of candidate motives (t = −0.56; p = n.s.), or vote intentions (t = −0.95; p = n.s.). Put simply, replies to challenging comments neither hurt nor helped the candidate to any discernible degree.

CONCLUSION

The appearance of interactivity between a candidate and other users on that candidate's Facebook page can shape perceptions and vote intentions held by those viewing the page. The key finding here is that experimental participants who viewed a page with supportive comments and likes from other users reported more favorable perceptions of and greater support for the candidate, relative to participants who viewed a page with challenging comments. These other users were not Facebook “friends” of the participants; to the participants’ knowledge, they might not have existed at all (in fact, they did not). Nevertheless, their comments swayed participants. This outcome is consistent with expectations derived from the warranting principle (Walther and Parks, 2002; Walther et al., 2008, 2009), which suggests that people will tend to see other-generated online comments as relatively difficult to manipulate and, thus, credible. As such, it reinforces but also extends other recent findings that observing interactivity between candidates and citizens on SNSs can influence potential voters (Lee and Shin, 2012a; Utz, 2009). At a broader level, this finding resonates with arguments that voters rely on information shortcuts such as social cues to evaluate candidates (Popkin, 1991) and that interpersonal influence can shape candidate evaluations and vote behavior (Huckfeldt and Sprague, 1995; Mutz, 2006).

The contrast between participants who viewed a page without comments and those who viewed a page with supportive comments/likes was not as marked (with a significant difference only for vote intentions). Nor was the contrast between participants who viewed a page without comments and those who viewed a page with challenging comments (with only a marginal difference on perceptions of candidate motives). Thus, it is not clear whether supportive comments/likes or challenging comments contributed more to the effects observed. In every case except one, the differences across conditions here were in the hypothesized directions. 13 Thus, the failure to find statistically significant differences may partly reflect the relatively small cell size per conditions. Future experimental tests with greater statistical power could be useful in isolating the extent to which supporting and/or challenging comments uniquely contribute to the effects observed.

As for candidate replies to either supportive or challenging comments, they never led to significantly more favorable perceptions or a greater likelihood of voting for the candidate. Indeed, the only clear effect of replies to supportive comments was to reduce the likelihood of voting for the candidate. The results provide little basis for speculating that increased statistical power would have yielded evidence that candidate replies led to more favorable candidate perceptions or a greater likelihood of voting for the candidate. The direction of the effect for candidate replies to supportive comments was negative––i.e., in the “wrong” direction––for all three dependent variables (significantly so for vote intentions, and marginally so for perceptions of motives). The direction of the effect for candidate replies to challenging comments was similarly negative (though not significantly so) for perceptions of candidate motives and vote intentions. The effect of candidate replies to challenging comments on perceptions of candidate traits was positive, but fell far short of statistical significance.

The results for candidate replies are consistent with the argument, derived from the warranting principle, that self-generated comments will influence impressions less than do other-generated comments (Walther et al., 2009). At first glance, these same results seem to clash with previous findings that candidates elicited more favorable responses by interacting with users through SNSs (Lee and Shin, 2012a; Utz, 2009). However, the present study differed from earlier research in terms of the population under study (U.S. residents versus residents of other nations), the SNS under study (Facebook versus other SNSs), and the nature of the comments and replies (broad in the present study versus more specific in the other studies). The results here suggest that U.S. candidates will not necessarily benefit by making general replies to user comments on Facebook, but the applicability of the findings beyond this set of conditions may be limited.

There are also other limitations to the present study. To begin with, it focused on a fictional candidate rather than a real one from the participants’ own district. In addition, it used constructed stimuli, though these stimuli were modeled on real-world content. Finally, it relied on a sample that was not fully representative of the general public or Facebook users, particularly in terms of age, though this sample was much more diverse than a typical student sample (and the results are largely unaltered when focusing only on Facebook users in the sample). Thus, future research could test the boundaries of the effects observed here by studying a broader array of user comments and candidate replies, across a wider range of political contexts, among additional samples.

Additional research could also investigate the conditions under and processes through which interactivity on SNSs influences candidate perceptions and vote intentions. In terms of the former, research could examine whether, and if so how, individual-level factors such as political knowledge, partisanship, and previous SNS use moderate such effects. For example, partisanship may play an important role in shaping how users respond to comments by and about candidates in the presence of partisan cues (as opposed to when these cues are absent, as in the present study). In terms of the latter, research could test whether factors such as the depth and nature of cognitive processing mediate these effects of interactivity. The present study did not capture how closely participants read the comments in the conditions that included them (though the significant differences between the group that read the supportive comments and the group that read the challenging comments indicate that at least some participants read these comments). Thus, lack of attention could help to explain the absence of effects in some cases. At the same time, it is not necessarily obvious that Facebook users would process comments on candidate pages more closely (or differently in other ways) in real-world settings.

Keeping in mind its limitations, the present study's findings carry potentially important political implications. At the most basic level, they suggest that it can matter for candidates whether other users say positive or negative things about them on their Facebook pages. Extending this point, the results may also carry strategic lessons: for example, candidates may face incentives to delete negative user comments while encouraging positive user comments. By contrast, candidate efforts at replying to either positive or negative user comments may pay little, if any political dividends. Consistent with the logic that user-generated information can carry more weight than information attributed to candidates, media reports allege that some politicians and campaigns have even used fake “sock puppet” and “troll” accounts to generate positive feedback on their own SNS pages and to attack opponents on the opponents’ own SNS pages (e.g., Dolan, 2011; Stueff, 2011). Candidates have long used carefully orchestrated social cues, from endorsements to photo opportunities to stage-managed public events, in their efforts to persuade voters that they are riding a wave of popular support. The increasing use of SNSs by voters provides candidates and other actors with new tools for projecting images of popularity or unpopularity in ways that may carry electoral consequences.

SUPPLEMENTARY MATERIALS

For supplementary material for this article, please visit Cambridge Journals Online: http://dx.doi.org/10.1017/XPS.2014.29.

1 Of the participants, 7% were under 35 years of age, 7% were between 35 and 44, 14% were between 45 and 54, 30% were between 55 and 64, 34% were between 65 and 74, and 8% were 75 or older.

2 Based on 2013 U.S. Census estimates, 51% of state residents were women, 22% were African American, 9% were Hispanic, 29% had a college degree, and 16% were 65 or older. In a 2013 survey of a probability sample of state residents (conducted by the authors), 40% of respondents identified as Democrats, 22% as Republicans, and 37% as independent or other.

3 The full results of these analyses are included in the supplemental appendix.

4 Participants were not explicitly told whether the candidate was real or fictional. We chose to use a fictional candidate because the study was conducted during the month preceding a general election, and we wished to avoid (1) influencing respondents’ attitudes toward any actual candidate, or (2) capturing responses that reflected attitudes toward any actual candidate. Omission of the full purpose of the research constitutes very low-level deception; furthermore, there was very little risk of harm to any of the participants, and the possible harms of using a real candidate would have been greater than the risk of harm from the procedure used. Respondents were not debriefed following the study.

5 The list was taken from the Federal Election Commission website (http://www.fec.gov/data/CandidateSummary.do?format=html&election_yr=2012).

6 The photograph of the fictional candidate was actually a photograph of a real politician from another state (former U.S. Representative Earl Pomeroy of North Dakota, who was defeated in 2010) whom participants would have been extremely unlikely to recognize.

7 None of the dependent variables (perceptions of candidate traits, perceptions of candidate motives, or vote intentions) differed significantly across participant party identification, suggesting that participants did not infer the candidate's party affiliation from the page.

8 Randomly generated names (created using the same procedure as the one used to select the candidate name) and stock images were used for the commenters. These names and images were consistent across all conditions that included comments.

9 For the sake of realism and parsimony, supportive comments and likes were manipulated together, rather than separately. Thus, the analyses cannot disentangle the relative contributions of the supportive comments and likes to the effects described below. However, the present study is more interested in capturing the effects of positive versus negative user feedback than in distinguishing between the effects of different forms of positive user feedback.

10 The wording for this question was as follows: “If you lived in Dennis Felton's Congressional district, how likely would you be to do each of the following––not likely at all, not too likely, somewhat likely, or very likely? . . . Vote for him.”

11 Randomization checks on party identification and the demographic variables revealed no significant differences across conditions. See the supplemental appendix for results.

12 All significance levels were based on two-tailed tests.

13 The one exception was the comparison between the No Comments and Challenging Comments conditions for vote intentions.

References

Allport, Floyd H. 1924. Social Psychology. Boston: Houghton Mifflin.
Allport, Floyd H. 1940. “Polls and the Science of Public Opinion.” Public Opinion Quarterly 4 (2): 249257.
Cogburn, Derrick and Espinoza-Vasquez, Fatima K.. 2011. “From Networked Nominee to Networked Nation: Examining the Impact of Web 2.0 and Social Media on Political Participation and Civic Engagement in the 2008 Obama Campaign.” Journal of Political Marketing 10 (1–2): 189213.
Dolan, Eric W. 2011. Sen. Scott Brown Aide Caught Using Twitter ‘Sock Puppet’ to Mock Opponent. Raw Story, (August 24). (http://www.rawstory.com/rs/2011/08/24/sen-scott-brown-aide-caught-using-fake-twitter-account-to-mock-opponent/).
Fernandes, Juliana, Giurcanu, Magda, Bowers, Keven W. and Neely, Jeffrey C.. 2010. “The Writing on the Wall: A Content Analysis of College Students’ Facebook Groups for the 2008 Presidential Election.” Mass Communication and Society 13 (5): 653675.
Hanson, Gary, Haridakis, Paul M. and Wagstaff, Audrey. 2010. “The 2008 Presidential Campaign: Political Cynicism in the Age of Facebook, MySpace, and YouTube.” Mass Communication and Society 13 (5): 584607.
Haynes, Audrey A. and Pitts, Brian. 2009. “Making an Impression: New Media in the 2008 Presidential Nomination Campaigns.” PS: Political Science and Politics 42 (1): 5358.
Huckfeldt, R. Robert and Sprague, John. 1995. Citizens, Politics and Social Communication: Information and Influence in an Election Campaign. New York: Cambridge University Press.
Johnson, Thomas J. and Perlmutter, David D.. 2010. “Introduction: The Facebook Election.” Mass Communication and Society 13 (5): 554559.
Kushin, Matthew J. and Yamamoto, Masahiro. 2010. “Did Social Media Really Matter? College Students’ Use of Online Media and Political Decision Making in the 2008 Election.” Mass Communication and Society 13 (5): 608630.
Lee, Eun-Je and Jang, Jeong-woo. 2013. “Not So Imaginary Personal Contact with Public Figures on Social Network Sites: How Affiliative Tendency Moderates Its Effects.” Communication Research 40 (1): 2751.
Lee, Eun-Ju and Oh, Soo Y.. 2012. “To Personalize or Depersonalize? When and How Politicians’ Personalized Tweets Affect the Public's Reactions.” Journal of Communication 62 (6): 932949.
Lee, Eun-Ju and Shin, Soo Y.. 2012a. “Are They Talking to Me? Cognitive and Affective Effects of Interactivity in Politicians’ Twitter Communication.” Cyberpsychology, Behavior, and Social Networking 15 (10): 515520.
Lee, Eun-Ju and Shin, Soo Y.. 2012b. “When the Medium is the Message: How Transportability Moderates the Effects of Politicians’ Twitter Communication.” Communication Research 41: 10881110. doi: 10.1177/0093650212466407.
Mutz, Diana C. 2006. Hearing the Other Side: Deliberative versus Participatory Democracy. New York: Cambridge University Press.
Pew Research Center's Internet & American Life Project. 2014. Social Networking Use. (http://www.pewresearch.org/data-trend/media-and-technology/social-networking-use/).
Popkin, Samuel L. 1991. The Reasoning Voter: Communication and Persuasion in Presidential Campaigns. Chicago: University of Chicago Press.
Stueff, Jack. 2011. Sarah Palin Has Secret ‘Lou Sarah’ Facebook Account to Praise Other Sarah Palin Facebook Account. Wonkette, (February 22). (http://wonkette.com/438825/is-sarah-palin-commenting-on-her-own-facebook-fan-page).
Sweetser, Kaye D. and Lariscy, Ruthann W.. 2008. “Candidates Make Good Friends: An Analysis of Candidates’ Uses of Facebook.” International Journal of Strategic Communication 2 (3): 175198.
Tedesco, John C. 2007. “Examining Internet Interactivity Effects on Young Adult Political Information Efficacy.” American Behavioral Scientist 50 (9): 11831194.
Tong, Stephanie T., Heide, Brandon V. D., Langwell, Lindsey and Walther, Joseph B.. 2008. “Too Much of a Good Thing? The Relationship Between Number of Friends and Interpersonal Impressions on Facebook.” Journal of Computer-Mediated Communication 13 (3): 531549.
Towner, Terri L. 2013. “All Political Participation Is Socially Networked? New Media and the 2012 Election.” Social Science Computer Review 31 (5): 527541.
Utz, Sonja. 2009. “The (Potential) Benefits of Campaigning via Social Networking Sites.” Journal of Computer-Mediated Communication 14 (2): 221243.
Utz, Sonja. 2010. “Show Me Your Friends and I Will Tell You What Type of Person You Are: How One's Profile, Number of Friends, and Type of Friends Influence Impression Formation on Social Network Sites.” Journal of Computer-Mediated Communication 15 (2): 314335.
Walther, Joseph B., Heide, Brandon V. D., Hamel, Lauren M. and Shulman, Hillary C.. 2009. “Self-Generated Versus Other-Generated Statements and Impressions in Computer-Mediated Communication: A Test of Warranting Theory Using Facebook.” Communication Research 36 (2): 229253.
Walther, Joseph B., Heide, Brandon V. D., Kim, Sang-Yeon, Westerman, David and Tong, Stephanie T.. 2008. “The Role of Friends’ Appearance and Behavior on Evaluations of Individuals on Facebook: Are We Known by the Company We Keep?Human Communication Research 34 (1): 2849.
Walther, Joseph B. and Parks, Malcolm R.. 2002. “Cues Filtered Out, Cues Filtered In: Computer-Mediated Communication and Relationships.” In Handbook of Interpersonal Communication, 3rd ed. Thousand Oaks: Sage Publications.
Warnick, Barbara, Xenos, Michael, Endres, Danielle and Gastil, John. 2006. “Effects of Campaign-to-User and Text-Based Interactivity in Political Candidate Campaign Web Sites.” Journal of Computer-Mediated Communication 10 (3).
Wilson, Robert E., Gosling, Samuel D. and Graham, Lindsay T.. 2012. “A Review of Facebook Research in the Social Sciences.” Perspectives on Psychological Science 7 (3): 203220.
Woolley, Julia K., Limperos, Anthony M. and Oliver, Mary B.. 2010. “The 2008 Presidential Election, 2.0: A Content Analysis of User-Generated Political Facebook Groups.” Mass Communication and Society 13 (5): 631652.
Zhang, Weiwu, Johnson, Thomas J., Seitzer, Trent and Bichard, Shannon L.. 2010. “The Revolution Will Be Networked: The Influence of Social Networking Sites on Political Attitudes and Behavior.” Social Science Computer Review 28 (1): 7592.