Relationships between Blogs as eWOM and Interactivity, Perceived Interactivity,
and Parasocial Interaction

Kjerstin S. Thorson, Shelly Rodgers

University of Missouri-Columbia


The purpose of this research was to examine the effects of a political candidate’s blog-a form of eWOM (electronic Word-of-Mouth)-on attitudes toward the website, attitudes toward the political candidate, and intentions to vote. The results showed that interactivity in the form of a blog significantly influenced attitude toward the website, but not attitudes toward the candidate or voting intention. However, perceived interactivity influenced all three dependent variables, but did not interact with interactivity, suggesting that these are two separate constructs. The effects were mediated by parasocial interaction.


The Internet plays an ever-increasing role in the lives of millions of Americans-and in politics. As of June 2004, 63 percent of the adult American population-about 128 million people-reported going online (Pew 2004). Of that population, more than 40 percent said they used the Internet for political information during the recent campaign-an increase of more than 50 percent over those who reported doing so in 2000 (Pew 2004). A political candidate’s website is now recognized as a basic component of any successful campaign. Candidate websites serve not only as resources for distributing information and messages about the campaign, but also as powerful tools for fundraising and mobilization (Cornfield 2004).

Nearly 13 million people have made use of the Internet to engage with and participate directly in campaign activities (Rainie, Cornfield, and Horrigan 2005, p. i). This new emphasis on “e-politics” (Cornfield 2004) has encouraged innovation in candidates’ use of the web to communicate with and motivate their supporters-and to get them talking about politics. Of particular interest is a new online feature that came into use during the most recent campaign: the online campaign journal or blog. These blogs-defined as websites that contain online personal journals with reflections by the writers and the opportunity for visitors to comment-allow candidates and their staff to directly address and interact with website visitors.

Blogs, as new, ostensibly interactive elements designed to enhance the persuasive impact of campaign websites, provide an opportunity to explore whether (and how) providing an opportunity for visitors to publicly exchange ideas and opinions has an impact on important attitudinal and behavioral variables. But will the presence of such interactive features influence perceptions that the site is “interactive?” And, if so, do these perceptions influence individuals’ attitudes toward the website, the candidate, and their intention to vote for the candidate?

The purpose of this research was to examine the effects of interactivity in the form of eWOM and perceived interactivity on attitudes toward the website, attitudes toward the candidate, and voting intentions. This study also sought to understand the role of parasocial interaction-an illusion of intimacy with a media personality (Horton and Wohl 1982)-in determining the persuasive effects of interactivity on a candidate blog. This was accomplished with a two-factor between-subjects experiment using a convenience sample of undergraduate students at a large Midwestern university.

Literature Review

Electronic Word-of-Mouth

The advent of the Internet has brought about a word-of-mouth revolution (Dellarocas 2003). Individual networks of interpersonal influence have expanded rapidly. The tip from a neighbor concerning the best brand of dishwasher can be affirmed-or overridden-by a simple Internet search that reveals the published opinions of strangers. Online, millions of individuals engage in a mass-mediated exchange of personal information and opinions every day.

Advertisers have long known that persuasion doesn’t only occur from the top down. The people we talk to every day, our friends and acquaintances, are significant and influential sources of opinion and information about products, brand names, and vote choice. Indeed, “personal contacts seem to be most effective in causing changes in opinion and behavior” (Brooks 1957, p. 155).

Candidate blogs are an interesting case. They are both a form of advertising, containing content prepared by campaign staffers who hope to influence election outcomes, and a form of word-of-mouth, linking together the politically interested, spreading their thoughts to anyone willing to take the time to read. With blogs, “those touched by the candidate, and kept in touch with by the campaign, could now touch back-and outward to other people (Cornfield 2004, p. 32).

Drawing on Hennig-Thurau et al. (2004), we define electronic word-of-mouth (eWOM) as positive or negative statements made about a product, company, or media personality that are made widely available via the Internet. Although many eWOM studies are conceptualized as peer-to-peer communication, online communication via eWOM can occur at the peer-to-peer and on other levels as well. In the present study, we conceive of eWOM as communication that occurs predominantly at the candidate-to-consumer level. As political candidates post their thoughts and opinions in blogs on their individual websites, they attempt to persuade voters and potential voters to see their point of view and, hence, gain their vote. However, the presence of a blog can stimulate consumers to post their own ideas to the candidate, and in the process of reading those thoughts other consumers can be influenced by those postings. Thus, political blogs represent an interesting case study because the flow of communication represents reciprocity between politicians and consumers, consumers and politicians, and also consumers and other consumers.

Recent studies have shown that eWOM is sometimes more effective than marketer-generated communication. For instance, Bickhart and Schindler (2001) showed through a 12-week experiment that individuals who collected product information from consumer-generated (eWOM) forums such as chatrooms or bulletin boards expressed higher product category interest than individuals who collected product information from traditional, marketer-generated forms such as corporate websites. The authors argued that the exercise of collecting product information on eWOM forums did not increase knowledge about the products but, rather, increased an interest in wanting more information about the product or product categories. These findings suggest that eWOM may provide an important outlet in which political candidates-who have historically not wanted to have direct contact with constituents via their websites (Stromer-Galley 2000)-may be able to better influence constituents with regard to liking and voting preferences. That is, to the extent that political blogs are a form of eWOM we would expect that they would help to increase constituent interest in the political candidate, his/her viewpoints, etc.

In an examination of the factors preceding negative word-of-mouth, Lau and Ng (2001) identified a series of factors relevant to the process. Their analysis emphasized the importance of not only individual factors (including personality, attitudinal and product involvement variables), but also situational factors. That is, what is happening around an individual, such as the proximity of others, influences whether or not they will engage in word-of-mouth. In our study, we argue that the presence of interactive forums, such as the blog, on a political candidate’s website acts as a facilitating situational factor. Visitors to the website are able to discuss their thoughts and opinions at the time of receiving information from the candidate and his or her staff. This opportunity for peer-to-peer exchange is enabled-and encouraged-by interactivity, which we discuss next.


The rise of the Internet as a communications medium has led to extensive research addressing what, if anything, makes the Internet unique among mass media. One element of the online experience that has received much attention is interactivity. Although researchers agree that interactivity is a unique element of the Internet, no consensus has been reached on a definition. Some researchers approach interactivity from an interpersonal perspective, identifying how closely messages are related (e.g., Rafaeli 1988), while others define interactivity from a mechanical perspective, examining interactivity as a structural element of the medium (e.g., Coyle and Thorson 2001; Steuer 1992). Still others define interactivity as a multidimensional construct that contains elements of the interpersonal and mechanical, as well as other perspectives, such as function or contingency (e.g., Ha and James 1998; Macias 2003; Sundar, Kalyanaraman, and Brown 2003).

In an attempt to synthesize twenty years of research on interactivity, McMillan (2002) proposed that concepts of interactivity can be subdivided into three primary research traditions: human-to-human, human-to-document, and human-to-system interactions. She argued that each of these research traditions contain key characteristics that can usefully categorize three separate models of interactivity. McMillan (2002) emphasized that there are multiple forms of interactivity and suggested that researchers in this area should locate their work within one or more of these three traditions and draw upon the literature specific to that tradition.

For the purposes of this study, interactivity was examined within the traditions of user-to-user and user-to-document dimensions, as defined by McMillan (2002). From a user-to-user perspective, interactivity is conceptualized as drawing on the interpersonal communication perspective to examine mediated interactions between individuals (McMillan 2002). User-to-document interactivity emphasizes how individuals “interact with documents and the creators of those documents” (McMillan 2002, p. 169).

The context of the present study is that of a political candidate’s blog, which, in addition to offering a form of mediated interpersonal interaction (i.e., the user-to-user interactivity of posting a comment and communicating with the candidate and other visitors to the website), also offers the possibility for interpersonal interaction (i.e., the user-to-document interactivity of reading blog posts and perceiving an interaction with the blog author and anyone else who might read the very public posting). Hence, although it may be that individuals will take advantage of interactive components to communicate with the candidate, as recognized by the user-to-user perspective, we argue here that, from a user-to-document standpoint, it may also be the case that individuals will not use the interactive features and yet the mere presence of interactive elements may influence how individuals perceive and process the website and its contents.

Sundar, Kalyanaraman, and Brown (2003) conceptualized interactivity in terms of Rafaeli’s (1988) definition, which focused on a user-to-system perspective that utilizes content delivery to users in a contingent fashion. They conducted an experiment to investigate the effects of interactivity on impression formation of a political candidate’s website. The results of their experiment revealed that increased interactivity of a political website led to more positive impressions of the political candidate and levels of agreement with the candidate’s policy positions (Sundar, Kalyanaraman, and Brown 2003). Although the authors approached interactivity from a user-to-system perspective, which differs from the current approach, their findings at least demonstrate that degrees of interactivity can differentially impact individuals’ psychological processing of a candidate’s website, which leads to our first hypothesis:

H1: High (versus low) interactivity will yield more favorable attitudes toward the website, impressions of the candidate, and voting intentions.

Perceived Interactivity

One subset of the interactivity literature argues that the effects of interactivity are more appropriately understood not as an element of a medium, but as a perceptual variable constrained by what individuals perceive as interactive (McMillan and Hwang 2002; McMillan, Hwang, and Lee 2003; Newhagen, Cordes, and Levy 1995; Wu 1999).

As with interactivity, definitions of perceived interactivity provide no clear consensus. McMillan (2002) conceives of perceived interactivity as three overlapping dimensions: two-way communication, control of navigation/choices, and time to load/time to find. Newhagen, Cordes, and Levy (1995) defined perceived interactivity as “the psychological sense message senders have of their own and of the receivers’ interactivity” (p. 165). Burgoon et al. (2000) suggested that interactivity can be conceptualized based on the qualitative experiences that users equate with interactivity. The authors suggest three relevant dimensions: interaction involvement (“the degrees to which users perceive they are cognitively, affectively, and behaviorally engaged in the interaction”), mutuality (“the extent to which users perceive and create a sense of relational connection”), and individuation (“the extents to which users perceive they have a rich, detailed impression of the other’s identity”).

Drawing on the dimensions of Burgoon et al. (2000) and Kiousis (2002), this study defines perceived interactivity as the extent to which users perceive their experience as a simulation of interpersonal interaction and sense they are in the presence of a social other.

McMillan, Hwang, and Lee (2003) found that perceptual variables (involvement and perceived interactivity) were stronger predictors of attitude toward the website than structural variables (site features and creative strategies). A strong relationship was also found between involvement and perceived interactivity (McMillan, Hwang, and Lee 2003). Wu (1999) identified a positive relationship between perceived interactivity, defined as two-dimensional construct of navigation and responsiveness, and attitude toward the website. Jee and Lee (2002) found that perceived interactivity and attitude toward the website were highly correlated.

Applied to the present research, these findings suggest that participants who perceive the website as highly interactive will demonstrate more favorable attitudes toward the website compared to those who do not perceive the website as interactive. This discussion leads to our second hypothesis:

H2: High (versus low) perceived interactivity will yield more favorable attitudes toward the website, impressions of the candidate, and voting intentions.

Parasocial Interaction

Parasocial interaction (PSI) was originally defined by Horton and Wohl (1982) as the illusion of “face-to-face relationship” with a media personality (p. 188). The current study defines parasocial interaction as “interpersonal involvement of the media user with what he or she consumes” (Rubin, Perse, and Powell 1985, p. 156). The parasocial relationship is in many ways similar to interpersonal relations (Auter 1992). Most of the research in PSI has been conducted in the area of television and radio and some definitions of PSI explicitly incorporate a broadcast-specific perspective, e.g., a one-sided interpersonal relationship that television viewers establish with media characters (Rubin and McHugh 1987). However, PSI has also been applied to websites (Hoerner 1999) and related to the concept of interactivity (Auter 2002; McMillan 2002).

Scholars have suggested that increased personalization of mass media, that is, designing mass mediated communications that closely resemble interpersonal communication, may increase the credibility and persuasiveness of the message (Beninger 1987). Research on word-of-mouth suggests that interpersonal networks are sometimes powerful sources of influence (Brooks 1957). Rubin (2002) proposed that because parasocial interaction is indicative of active, involved media use, it can affect attitudes and behaviors. A study of PSI among audiences of talk radio found that parasocial relationships predicted planned, frequent exposure to programming, increased perceptions of the talk show host as a credible source of information, and increased the feeling that the talk show host influenced listeners’ attitudes and behaviors (Rubin and Step 2000). Thus, in relation to political campaign websites, increased levels of parasocial interaction may enhance candidates’ ability to influence the attitudes and behaviors of potential voters.

Another study showed that PSI levels increased depending on a manipulation of television program content (Auter 1992). Auter proposed that media messages can exhibit varying degrees of “parasociability” (Auter 1992, p. 175) depending on the degree of intimacy that is seemingly offered by the media personality. He manipulated the degree of intimacy by showing viewers a television program in which the main character “breaks through the fourth wall” by stepping out of character to address the audience directly. Auter argued that this manipulation, “by directly addressing the audience and adjusting to supposed responses” (Auter 1992, p. 177), increased the perceived interactivity level of the content and therefore parasocial interaction.

Auter’s experiment suggests that increased interactivity of a message will increase parasocial interaction, yet he did not specify a definition for interactivity or test whether his respondents perceived differences in levels of interactivity between the two experimental conditions. However, Auter’s findings, transferred to the Internet sphere, lead to the formulation of the following hypothesis:

H3: Parasocial interaction will mediate the effects of perceived interactivity on attitude toward the website and impressions of the candidate.



A total of 69 undergraduate students (52% female, 48% male) from a major Midwestern university were the participants. Participants were drawn from an introductory political science course and volunteered to participate for extra credit. Subjects were treated in accordance with IRB and campus-wide guidelines and policies on appropriate treatment of human subjects.

Independent Variables

The study had two independent variables: interactivity and perceived interactivity. Interactivity was manipulated and perceived interactivity was measured. Interactivity was manipulated by exposing participants to either the high or low condition. In the high interactivity condition, participants were provided with an opportunity to interact with the candidate’s website via a hyperlink that enabled individuals to post a message to the candidate’s blog, whereas the low interactivity condition did not provide an opportunity to interact with the candidate.

Perceived interactivity was measured using seven, five-point Likert scales that ranged from (1) strongly disagree to (5) strongly agree. The items, adapted from McMillan and Hwang (2002), were: Interacting with this website was like having a conversation with a representative of the campaign; I felt as if this website talked back to me while I was navigating; I perceive the website to be sensitive to my needs for information; The candidate’s website is interpersonal; I could communicate directly to the candidate if I wanted to; The candidate’s website is interactive; The website had the ability to respond to my specific questions quickly and efficiently (alpha=.84).

Mediating Variable

There was one mediating variable: parasocial interaction. A 14-statement scale for parasocial interaction was adapted from the 20-item PSI scale developed by Rubin, Perse, and Powell (1985) as well as Hoerner’s (1999) PSI scale for websites. The statements were measured using a five-point Likert scale ranging from (1) strongly disagree to (5) strongly agree. The items were: The website showed me what the candidate is like; When the candidate showed me how she feels about the issues, it helps me to make up my own mind about the issues; When I’m reading the website, I feel as if I am part of a close-knit group; I found myself comparing my ideas with what the candidate said; The candidate makes me feel comfortable, as if I were with a friend; I see the candidate as a natural, down-to-earth person; If the candidate were going to be on television, I would watch that program; Visiting this website helped me form opinions about the topics and issues presented at this site; When the candidate wrote about the issues, she seemed to understand the kinds of things I want to know; If there were a story about the candidate in the newspaper, I would read it; The candidate is interested in my opinions and comments; I would tell my friends about this candidate; I can trust the information I get from this candidate; I wanted to say something to this candidate.

Factor analysis followed by Varimax rotation revealed four distinct factors. Two factors contained only two measures and thus were excluded from the analysis. A third factor contained three measures, but yielded an insufficient reliability (alpha=.60) and was therefore excluded from the analysis. The remaining factor (alpha=.83) was comprised of six statements that included the interpersonal dimension of parasocial interaction: The website showed me what the candidate is like; The candidate made me feel comfortable, as if I were with a friend; When the candidate wrote about the issues, she seemed to understand the kinds of things I want to know; I would tell my friends about this candidate; I can trust the information I get from this candidate; I found myself comparing my ideas with what the candidate said. This factor, the interpersonal dimension of the parasocial relationship (PSI), was used in the subsequent analysis.

Dependent Variables

There were three dependent variables: attitude toward the website, impressions of the candidate, and voting intentions.

Attitude toward the website. Attitude toward the website was measured using Chen and Wells’ (1999) scale, which consisted of five, five-point Likert scales ranging from (1) strongly disagree to (5) strongly agree. The items were: This website makes it easy for me to build a relationship with this candidate; I would like to visit this website again in the future; I’m satisfied by the service provided by this website; I feel comfortable in surfing this website; I feel surfing this website is a good way for me to spend my time. One additional item, “Compared with other websites, I would rate this one as” was measured with a single, semantic differential that ranged from (1) one of the worst to (5) one of the best (alpha=.85).

Impression of the candidate. Impression of the candidate was measured with a seven-point semantic differential scale anchored by seven candidate traits: Inspiring/uninspiring; knowledgeable/unknowledgeable; immoral/moral; strong leader/weak leader; trustworthy/untrustworthy; not credible/credible; honest/dishonest (alpha=.88).

Voting intention. Voting intention was measured with a single five-point Likert scale that ranged from (1) strongly disagree to (5) strongly agree. The item was: “I would vote for this candidate.”

Stimulus Materials

Two websites were created, one for each level of interactivity. Both sites contained the same information about the candidate’s position on a series of issues: education, jobs and economy, healthcare, protecting America and supporting our troops, environment, and social security. Content was carefully crafted to reflect a moderate political position not readily identifiable as either conservative or liberal. This was done to control political partisanship to the extent possible. A small sample of actual candidate websites-both Republican and Democrat-were examined to ensure representative choices of wording and style.

High interactivity was manipulated by providing the candidate’s personal web journal, written in first person, and a “post your feedback to the candidate” button, which was an active hyperlink. The website in the low interactivity condition displayed the same text, written in third person, and no hyperlink or feedback button. Participants in the high interactivity group were instructed that they could post feedback to the candidate if they chose to do so. Nine out of the 35 participants (26%) in the high interactivity condition chose to post feedback. Posted feedback appeared immediately on the participant’s screen, formatted as a “comment” on the candidate’s blog post.


The experiment took place in a laboratory setting. Upon arrival, participants were randomly assigned to one of two conditions of a website: low interactivity or high interactivity. Participants in both conditions were instructed to spend time reading the website and, after seven minutes had expired, respond to a questionnaire about the website.

Manipulation Checks

A manipulation check (N=21) was conducted prior to the start of the study (during a pre-test) to confirm that the manipulation of interactivity (high vs. low) was successful. This was accomplished by using the measured items for perceived interactivity. Our justification is grounded in the logic that, if individuals perceive that high interactivity is more interactive than low interactivity, then we can conclude that the manipulation was a success. Our results showed that the mean differences for perceived interactivity between the low (M=20.00, SD=5.20) and high interactivity (M=21.71, SD=7.54) conditions were not statistically significant (presumably due to a small N). Thus, a second manipulation check was conducted as part of the actual experiment. The results showed that the high interactivity group had significantly higher perceived interactivity (M=20.8, SD=5.36) than did the low interactivity group (M=15.9, SD=5.87), (t(63)=-3.49, p<.001), which confirmed that the manipulation was successful.

Data Preparation

Prior to data analysis, perceived interactivity was transformed from a continuous- to nominal-level variable by conducting a median split, which assigned half of the subjects to high and half to low perceived interactivity (PI). The newly transformed variable was used in subsequent analyses of the data.

Data Analysis and Results

A general linear model (GLM) multivariate analysis (MANOVA) was conducted to examine Hypotheses 1 and 2. Hypothesis 1 predicted that high (versus low) interactivity would yield higher attitude toward the website, impressions of the candidate, and voting intentions. As shown in Table 1, the results of the MANOVA supported this hypothesis (=.86, F (3, 55) = 2.97, p < .05). There was a positive effect of interactivity, but only for attitude toward the site. Specifically, individuals in the high interactivity condition expressed more favorable attitudes toward the website than individuals in the low interactivity condition. However, interactivity did not significantly influence impressions of the candidate or voting intentions (see Table 2).

Table 1 Results of MANOVA
Influence of Interactivity and Perceived Interactivity on the Dependent Variables

Results of MANOVA Influence of Interactivity and Perceived Interactivity on the Dependent Variables

Table 2 Means and Standard Deviation of the Results of Hypotheses 1 and 2

Means and Standard Deviation of the Results of Hypotheses 1 and 2

Hypothesis 2 predicted that high (versus low) perceived interactivity would result in higher attitudes toward the website, impressions of the candidate, and voting intentions. The results of the MANOVA supported this hypothesis (=.46, F (3, 55) = 21.46, p < .001). Perceived interactivity significantly influenced all three dependent variables (see Table 2). The means and standard deviations can be found in Table 2.

Hypothesis 3 predicted that parasocial interaction (PSI) would mediate the effects of interactivity and perceived interactivity on attitude toward the website, impressions of the candidate, and voting intention. The procedures described by Baron and Kenny (1986) were used to test the mediation hypothesis. The following model was tested with a series of regressions: interactivity (I) and perceived interactivity (PI) were the independent variables; attitude toward the website, impressions of the candidate, and voting intentions were the dependent variables; and parasocial interaction (PSI) was the mediating variable. First, PSI was regressed on PI and I; second, three dependent variables were regressed on PI and I; last, PSI was regressed on the dependent variables and PI and I. The findings revealed statistically significant results at each of the three steps described above, thus indicating a complete mediation effect (Baron and Kenny 1986).


The purpose of this research was to test the effects of an interactive blog, a form of eWOM, on attitudes toward the website, attitudes toward the political candidate and intentions to vote. We were interested in the relationships between interactivity, perceived interactivity, and parasocial interaction in the context of a political candidate’s blog. Participants in the high interactivity condition reported significantly higher scores on perceived interactivity, demonstrating that perceptions of interactivity are influenced by the presence of a single interactive element on a website: in this case, a hyperlink that provided the opportunity to post public feedback to a political candidate’s blog. In addition, interactivity significantly affected attitude toward the website, although its influence on impressions of the candidate and voting intentions was not significant.

We also considered a second independent variable, a measure of perceived interactivity. Perceived interactivity had a significant effect on all three independent variables, providing confirmation of McMillan, Hwang, and Lee’s (2003) finding that, in terms of attitudinal outcomes, perceptual variables are perhaps more important than the structural features of interactivity present on a website.

As hypothesized, the relationship between interactivity, perceived interactivity, and the dependent variables was mediated by parasocial interaction. This finding strongly suggests not only that parasocial interaction is a phenomenon in Internet-mediated interactions, just as it is for television viewers and radio listeners, but that the formation of parasocial relationships is influenced by perceptions of interactivity.

These results strongly argue for persuasive effects resulting from providing an opportunity for website visitors to share their thoughts and opinions in a Blog forum. Scholars of eWOM have generally emphasized the impact on persuasion from hearing or reading information from peers concerning a product or brand name. This study approaches eWOM from the opposite direction, looking at how the relationship between the supporters and political candidates is affected by provision for a forum to speak one’s mind. The results suggest that even the presence of a blog and the perceived interactivity of that blog can enhance perceptions of the political candidate’s website, attitudes toward the political candidate, and voting preferences. Theoretical and practical implications are discussed next.

Theoretical Implications and Directions for Future Research

Our study clearly suggests that user-to-user interactivity or at least a feature providing the opportunity for such mediated interpersonal interaction can be an effective feature of candidate websites. This type of interactive feature is important for two distinct reasons: First, the presence of interactivity positively influenced evaluations of the website, and second, the presence of an interactive feature influenced individuals’ perceptions of interactivity and thus affected key attitudinal and behavioral variables-attitudes toward the candidate and voting intentions.

By examining perceptions of interactivity as distinct from the manipulation of interactive features, our study provides insight into the ongoing debate as to whether interactivity is a medium characteristic or individuals’ perceptions of interactivity that is the crucial measurement for studies on interactivity. Our results demonstrate that not all individuals perceive interactivity in the same way and that how interactivity is perceived may have a greater impact on attitudes than the mere presence of an interactive feature. Indeed, perceptions of interactivity significantly affected not only attitudes toward the website, but also positively influenced evaluations of and intention to vote for the candidate. Thus, our findings suggest that the experimental manipulation of interactive features alone may not be sufficient to capture the persuasive impact of ostensibly interactive websites. Measures of perceived interactivity are needed to provide a more complete picture of the effects of interactivity on attitudinal outcomes. In addition, future research is needed to begin to build an understanding of the relationship between interactive features and perceptions of interactivity.

We believe that as scholars continue to explore interactivity in the context of eWOM it will be important to consider in each study the perspective from which the construct is explored-user-to-user, user-to-document, or user-to-system-as advocated by McMillan (2002). In our study, we focused primarily on a user-to-user perspective, hypothesizing that interacting with the candidate’s blog, and thus the candidate and other website users, would have a positive effect on attitudes. However, the user-to-document approach is perhaps more central to our research. It provides a framework from which to explore those instances when interactive features, such as feedback or comment buttons on blogs, are not used (i.e., individuals choose not to interact), but still influence perceptions and attitudes merely through their presence. In our study, for example, only nine of the participants in the high interactivity condition chose to post comments to the blog, yet users in the high interactivity condition were still more likely to perceive the feedback mechanism as interactive, when compared to the low interactivity group. This suggests that the effects of interactivity and perceived interactivity on the dependent variables do not necessarily require actual engagement in a user-to-user interaction. Perceptions of the opportunity to interact may be sufficient for effects to occur, providing yet another reason for researchers to measure perceived interactivity in addition to experimentally manipulating interactive features. Future research should seek to better understand the comparative effects of engaging in an interaction versus merely perceiving the possibility for interaction. That is, we should consider the effects of “lurking” on a website that provides for user-to-user interaction and also the possible effects of engaging in a true two-way interaction (i.e., participant responds to candidate; then, the candidate or another site visitor responds back to the participant).

This concept is of particular relevance for scholars of eWOM. The results of this study suggest that providing an opportunity to comment on a candidate-sponsored blog has positive effects on site visitors-regardless of whether they make use of the interactive features. Therefore, forums for eWOM may have important effects not only by spreading opinions to interested visitors who read the thoughts of others, but also by changing the relationship between visitor and site sponsor through perceptions of interactivity and thus responsiveness.

Parasocial interaction proved to be an important variable in this study because it mediated the effects of interactivity and perceived interactivity on the dependent variables. Researchers have only begun to explore the psychological processes through which interactivity affects attitudes. In the case of the user-to-user or mediated interpersonal interactivity, it seems likely that parasocial interaction may play an important role in that process. In our study, we believe that providing an opportunity to interact with the candidate encouraged a sense of intimacy between participants and candidate, creating a facsimile of an interpersonal relationship. This sense of a relationship with the candidate, illusory though it may be, encouraged positive evaluations of not only the candidate’s website, but also evaluations of and intention to vote for the candidate. We believe that parasocial interaction is an additional positive effect of interactivity, one that warrants further exploration. In particular, what features of a website encourage parasocial interaction? How can website authors maximize the formation of parasocial relationships? Horton and Wohl’s (1982) original formulation of PSI posited that parasocial relationships deepen over time and repeated exposure; therefore it may be the case that only extended exposure to a candidate’s blog will result in the highest levels of parasocial interaction.

Practical Implications

While previous research has demonstrated that political campaigns resist investing time and resources in building and maintaining user-to-user or human-interactive features (Stromer-Galley 2000), the use of candidate blogs in the 2004 primary (Cornfield 2004) and general presidential campaigns suggests that this attitude may be shifting-and for good reason. Findings from this study suggest the candidate blog is an effective tool for augmenting the persuasive effects of a candidate website. In particular, the campaign blog, which provides an opportunity for interaction with the candidate or the candidate’s campaign staff, appears to encourage an increased sense of intimacy with the candidate. As such, it may be a powerful technique for building all-important relationships with potential voters.

The findings of this study are also applicable outside the domain of political communication. Both advertisers and companies operating e-commerce websites have been interested in the persuasive effects of interactivity. A number of corporations and Internet advertisers have recently begun to explore the possibility of using blogs on their websites in order to encourage relationship building with their consumers through forums for eWOM. Our results-particularly discovery of the mediating effects of parasocial interaction-provide a foundation for encouraging such endeavors. Increased interactivity in the user-to-user domain may encourage visitors to build new kinds of relationships with the sponsoring corporation. Based on our results, it seems likely that providing customers with the ability to share opinions and information through a blog affects corporate image not only by spreading opinions about the company, but also by positively impacting the relationship between the brand or company and the individual who perceives the website as interactive.

Limitations and Directions for Future Research

Findings of this study are limited in their application to the general population due to the use of a small convenience sample of undergraduate students. However, because the engagement of college students in political campaign activities is of particular interest to political candidates, the composition of this sample is perhaps of greater utility than it would be for other studies. Either way, more research is needed to explore politically active college students as a unique segment of the population and their use of and influence of political blogs.

Additional research is also needed to examine blog ownership/use/readership as a factor in the perceptions of the candidate blog. For example, if the respondent owns or operates a blog or live journal, will he/she differ in his/her expectations of interactivity and perceived interactivity when compared to bloggers who do not own/operate a blog or non-bloggers? Follow-up experiments can isolate the effects of this potential intervening variable.

Last, the factor analysis of the parasocial interaction scale demonstrated that further research is needed to determine the optimal set of measures for Internet-based PSI. It may be that not all of the dimensions relevant to PSI studies of television and radio viewing are relevant to the study of web-mediated PSI. Additional exploration and validation of PSI measures for the Internet will be an important next step in this line of research.


Electronic word-of-mouth (eWOM) in the form of a blog is a new form of advertising and political candidates appear to use this forum to share their viewpoints and engage their constituents. Even for consumers who simply read eWOM but don’t actually post comments can be influenced by those postings. Our study has shown that even the presence of an interactive feature, such as a hyperlink that enables users to post messages, influences how individuals perceive and process websites-in this case a political candidate’s website. However, it appears that Internet users have differing ideas about what constitutes interactivity. For those individuals who perceive a hyperlink as interactive, candidate evaluations and voting intentions seem to improve. Interactivity also seems to have a “bonus” effect that includes the creation of an illusory one-to-one relationship with the individual behind the website. We believe these findings shed new light on the effects of interactivity in the form of eWOM blogs on information processing.


Auter, Philip J. (1992), “TV that Talks Back: An Experimental Validation of a Parasocial Interaction Scale,” Journal of Broadcasting & Electronic Media, 36 (2), 173-181.

Baron, Reuben M. and David A. Kenny (1986), “The Mediator-Moderator Variable Distinction in Social Psychological Research: Conceptual, Strategic and Statistical Considerations,” Journal of Personality and Social Psychology, 51, 1173-1182.

Beninger, James R. (1987), “Personalization of Mass Media and the Growth of Pseudo Community,” Communication Research, 14 (3), 352-371.

Bickhart, Barbara, and Robert M. Schindler (2001), “Internet Forums as Influential Sources of Consumer Information,” Journal of Interactive Marketing, 15 (3), 31-40.

Brooks, Robert C. (1957), “Word-of-Mouth Advertising in Selling New Product,” Journal of Marketing, 22, 154-161.

Burgoon, Judee K., Joseph A. Bonito, Bjorn Bentsson, Artemio Ramirez, Jr., Norah E. Dunbar, and Nathan Miczo (2000), “Testing The Interactivity Model: Communication Processes, Partner Assessments, and the Quality of Collaborative Work,” Journal of Management Information Systems, 16 (3), 33-56.

Chen, Qimei., and William D. Wells (1999), “Attitude Toward the Site,” Journal of Advertising Research, 39 (5), 27-37.

Cornfield, Michael (2004), Politics Moves Online: Campaigning and the Internet, New York: Century Foundation.

Coyle, James R. and Esther Thorson (2001), “The Effects of Progressive Levels of Interactivity and Vividness in Web Marketing Sites,” Journal of Advertising, 30 (3), 65-77.

Dellarocas, Chrysanthos (2003), “The Digitization of Word of Mouth: Promise and Challenges of Online Feedback Mechanisms,” Management Science, 49 (10), 1407-1424.

Ha, Louisa, and E. Lincoln James (1998), “Interactivity Reexamined: A Baseline Analysis of Early Business Web Sites,” Journal of Broadcasting & Electronic Media, 42 (4), 457-474.

Hennig-Thurau, Thorsten., Kevin P. Gwinner, Gianfranco Walsh, and Dwayne D. Gremler (2004), “Electronic Word-Of-Mouth Via Consumer-Opinion Platforms: What Motivates Consumers to Articulate Themselves on the Internet?” Journal of Interactive Marketing, 18 (1), 38-52.

Hoerner, John (1999), “Scaling The Web: A Parasocial Interaction Scale for World Wide Web Sites,” in D. W. Schumann and E. Thorson, eds., Advertising and the World Wide Web, Mahwah, NJ: Lawrence Erlbaum Associates, 135-147.

Horton, Donald, and R. Richard Wohl (1982), “Mass Communication and Para-Social Interaction: Observation on Intimacy at a Distance,” in Inter/Media, G. Gumpert and R. Cathcart, eds., New York: Oxford, 188-211.

Jee, Joonhyung, and Wei Na Lee (2002), “Antecedents and Consequences of Perceived Interactivity: An Exploratory Study,” Journal of Interactive Advertising, 3 (1), <> (accessed on 12/92004).

Kiousis, Spiro (2002), “Interactivity: A Concept Explication,” New Media & Society, 4, 355-383.

Lau, Geok Theng, and Sophia Ng (2001), “Individual and Situational Factors Influencing Negative Word-Of-Mouth Behaviour,” Canadian Journal of Administrative Sciences, 18 (3), 163-178.

Macias, Wendy (2003), “A Beginning Look at the Effects of Interactivity, Product Involvement and Web Experience on Comprehension: Brand Web Sites as Interactive Advertising,” Journal of Current Issues and Research in Advertising, 25 (2), 31-44.

McMillan, Sally J. (2002), “Exploring Models of Interactivity from Multiple Research Traditions: Users, Documents, and Systems,” in Handbook of New Media: Social Shaping and Consequences of ICTs, L. A. Lievrouw and S. Livingstone, eds., London: Sage, 163-182.

— and Jang-Sun Hwang (2002), “Measures of Perceived Interactivity: An Exploration of the Role of Direction of Communication, User Control, and Time in Shaping Perceptions of Interactivity,” Journal of Advertising, 31 (3), 29-42.

—, —, and Guiohk Lee (2003), “Effects of Structural and Perceptual Factors on Attitudes toward the Website,” Journal of Advertising Research, 43, 400-409.

Newhagen, John E., John W. Cordes, and Mark R. Levy (1995), “Nightly at NBC.Com: Audience Scope and the Perception of Interactivity in Viewer Mail on the Internet,” Journal of Communication, 45, 164-175.

Pew Internet & American Life Project (2004), “The Internet and Democratic Debate.” <> (accessed on 12/9/2004).

Rafaeli, Sheizaf (1988), “Interactivity: From New Media to Communication,” in Advancing Communication Science: Merging Mass and Interpersonal Process, R.P. Hawkins, J.M. Wiemann, and S. Pingree, eds., Newbury Park, CA: Sage, 110-134.

Rainie, Lee, Michael Cornfield, and John Horrigan (2005), “The Internet and Campaign 2004,” Pew Internet and American Life Project, <> (accessed on 5/5/2005).

Rubin, Alan M. (2000), “Impact of Motivation, Attraction, and Parasocial Interaction of Talk Radio Listening,” Journal of Broadcasting & Electronic Media, 44 (4), 635-655.

— (2002), “The Uses-and-Gratifications Perspective of Media Effects,” in Media Effects: Advances in Theory and Research, J. Bryant and D. Zillmann, eds.,Mahwah, NJ: Elbaum, 525-548.

—, Elizabeth M. Perse, and Robert A. Powell (1985), “Loneliness, Parasocial Interaction, and Local Television News Viewing,” Human Communication Research, 12, 155-180.

Rubin, Rebecca B. and Michael McHugh (1987), “Development of Parasocial Interaction Relationships,” Journal of Broadcasting & Electronic Media, 31, 279-282.

Steuer, Jonathan (1992), “Defining Virtual Reality: Dimensions Determining Telepresence,” Journal of Communication, 42, 73-93.

Stromer-Galley, Jennifer (2000), “On-Line Interaction and Why Candidates Avoid It,” Journal of Communication, 50, 111-132.

Sundar, S. Shyam, Sriram Kalyanaraman, and Justin Brown (2003), “Explicating Web Site Interactivity: Impression Formation Effects in Political Campaign Sites,” Communication Research, 30 (1), 30-59.

Wu, Guohua (1999), “Perceived Interactivity and Attitude toward the Website,” Paper presented at the American Academy of Advertising Annual Conference, Albuquerque, NM.

About the Authors

Kjerstin Thorson is an M.A. candidate at the University of Missouri School of Journalism. Her research explores the impact of new media on political evaluations and participation in politics.

Shelly Rodgers (Ph.D.) is an Assistant Professor in the Department of Advertising at the University of Missouri-Columbia’s School of Journalism. Her research examines information processing of Internet health care advertising, marketing and communication, and the impact of new technologies on psychosocial well-being.