Interactivity Is in the Eye of the Beholder: Function, Perception, Involvement, and Attitude toward the Web Site
Sally J. McMillan, University of Tennessee, Knoxville, TN
Much of the literature on computer-mediated communication assumes these new media forms are interactive. But what is interactivity? This study begins by briefly reviewing key literature that explores differing approaches to interactivity. While some scholars see interactivity as a function of the medium itself, others argue that interactivity resides in the perceptions of those who participate in communication. Research reported here compares functional and perceived interactivity. Relationships between interactivity, attitude toward the Web site, involvement in the subject matter, and demographic characteristics are also considered.
Review of the Literature
Rafaeli (1988, p. 11) defined interactivity as: “An expression of the extent that, in a given series of communication exchanges, any third (or later) transmission (or message) is related to the degree to which previous exchanges referred to even earlier transmissions.” Rafaeli conducted a number of studies (see for example Rafaeli, 1990; Rafaeli and Sudweeks, 1997) in which he examined interactivity as a process-related variable based on relatedness of sequential messages. Heeter (1989) suggested interactivity was a multi-dimensional concept based in the functions of the medium. Massey and Levy (1999) operationalized Heeter’s conceptual definition and examined Web sites for interactivity based on presence of functional features such as e-mail links, feedback forms, and chat rooms. Jensen (1998) defined interactivity as (p. 201): “A measure of a media’s potential ability to let the user exert an influence on the content and/or form of the mediated communication.”
But not all scholars believe that interactivity resides in the technical capacity of the medium. Morrison (1998) suggested that failure to view evolving media from the user’s perspective might be a blind spot in the study of interactivity. Newhagen (1998) noted “traditional concepts of audience have little descriptive power in the context of the Internet. Rather, the individual user comes to conceptual center stage.” While the capacity to carry out two-way communication and other technical aspects of a medium may help to facilitate interactivity, the uses that individuals make of evolving media may better explain the interactive process. Williams, Stover, and Grant (1994) suggested that although media technologies change rapidly, understanding individuals’ uses of those media is a key step in the theory-building process. Wu (1999) has also focused on the need to study interactivity from the perspective of perceptions of those who use interactive media.
McMillan and Downes (2000) interviewed individuals who teach, research, and create content in the evolving environment of computer-mediated communication. From those interviews, they identified key dimensions of interactivity that are based on individuals’ perceptions. The dominant dimensions to emerge from that study were direction of communication and control of the communication experience.
Direction of communication has been a key underlying concept in the work of scholars such as Bretz (1983), Rice (1984), and Rogers (1995) who examined role taking, feedback, and mutual discourse in computer-mediated communication. The implicit assumption in this work is that two-way communication is more interactive than one-way communication.
The concept of control is key to Jensen’s (1998) definition of interactivity. Finn (1998) also suggested that the sender/receiver ratio of control in content creation, presentation, and preservation is a key dimension of computer-based information systems. In general, as participants feel that they gain more control over the communication experience, they view that communication as more interactive.
Figure 1 presents a model of interactivity developed by McMillan (1999) that identifies four “types” of interactivity that may be found in computer-mediated environments. The model is based primarily on variation in direction of communication and control of the communication experience.
Figure 1. Four Models of Cyber-Interactivity
One-WayCommunication DirectionTwo-Way
/ Rich ContentOnline environments that provide in-depth searchable content such as databases. May also include customized information resources such as personalized newspapers. /
Virtual Community
Chat rooms, bulletin boards, and other environments where participants build shared communication. Sender and receiver roles become indistinguishable./ Packaged Content
Content created to attract an audience, promote a product or service, build a brand, or perform some persuasive but non-selling communication function. Includes e-mail news summaries and corporate Web sites. / Virtual Transaction
Environments that facilitate online transactions. Transactions may be for the sale of products, solicitations for membership or donations, providing online customer support, and so forth.
S = Sender, R = Receiver, P = Participant (sender/receiver roles are interchangeable)
McMillan and Downes (2000) identified direction of communication as the primary dimension related to how individuals perceive interactivity as a characteristic of the medium. But two other medium-related dimensions were also identified: time and place.
Some scholars have suggested that interactivity must occur in “real-time” (see for example Steur, 1992; Rice & Williams, 1984). Zack (1993) focused on the importance of simultaneous and continuous exchange of information in interactive communication. By contrast other observers (see for example Finn, 1998 and Rheingold, 1993) have suggested that one of the great appeals of cyber-interactivity is that participants can interact without the necessity for co-presence in time. The participants in the McMillan and Downes (2000) study seemed to suggest that as communication environments become more sensitive to the time demands of participants they exhibit greater interactivity. Rheingold (1993) suggested that interactive computer-mediated environments create a sense of place. And McMillan and Downes (2000) found that some, but not all, types of computer-mediated communication may create a kind of “online environment” that provides communicators with a sense of place.
McMillan and Downes (2000) identified control as the primary dimension related to how individuals perceive interactivity as a characteristic of the communication participants. But two other participant-related dimensions were also identified: activity, and purpose of communication.
Morrison (1998) found level of activity to be a key theme among individuals who shared their perceptions of interactive media. Rafaeli (1990) also recognized the importance of activity. He suggested interactivity is a series of active and reactive communications that are related. Respondents in the McMillan and Downes (2000) study suggested that computer-mediated communication that is informational in nature is more interactive than is communication that is perceived to have persuasive purposes. Schultz (1998) observed that newspaper-sponsored online forums are likely to be information-oriented as contrasted with online shopping forums that serve a persuasive function.
Packaged Content, as illustrated in Figure 1, is based on one-way communication from a sender to a receiver. The receiver has little control over the message and takes relatively simple actions to negotiate through the message. The perceived purpose of the communication is to persuade the receiver to the sender’s point of view. Timing is relatively “set.” Content availability is controlled by the schedule of the sender. These sites do not create a sense of place. A typical Packaged Content is the corporate Web site that does little more than present an electronic version of corporate brochures.
The second type, Rich Content, is also one-way communication. But control shifts. The sender creates messages, but the receiver exerts control over what he/she will view. The level of activity is greater than for Packaged Content. Rather than selecting from a menu of choices, the receiver initiates a series of actions with the computer that results in retrieval of specific information requested by the receiver. The perceived purpose of Rich Content is to inform rather than to persuade. Time is relatively set; the receiver cannot demand to see content until the sender has posted it. Individuals who use Rich Content sites are more likely to view them as information sources rather than as places where they go to interact. Typical examples include searchable databases and archived information.
The third type, Virtual Transaction, utilizes two-way communication. Because control remains primarily with the sender, the roles of sender and receiver are not completely eliminated. The sender may try to reduce activity required of the receiver by presenting a well-organized list of options from which the receiver can select. The perceived purpose is to persuade the consumer to make some type of transaction. Virtual Transactions are sensitive to time in two ways. First, the closer to “real time” that communication occurs, the more likely that a transaction will be completed. Second, use of tools such as e-mail allow both sender and receiver to communicate efficiently without requiring that they be in the same time or space. Successful Virtual Transaction sites do create a sense of place. Typical examples include commerce-oriented Web sites and online customer support facilities.
The fourth type, Virtual Community, utilizes two-way communication. In this type, senders and receivers switch roles so frequently as to be almost indistinguishable. All participants retain some control over the communication and all must exert effort to stay current on the exchange of actions and reactions. The perceived purpose of most Virtual Communities is to exchange information and build understanding among participants. Virtual Community allows for the flexibility of both real-time and time-shifted communication. Additionally, participants often view these Virtual Communities as “places” where they can gather. Typical examples include chat rooms, bulletin boards, and newsgroups.
A key question that grows from the literature is the relationship between interactive features of a Web site and perceptions of interactivity. Thus, the first hypothesis examines the relationship of the variables in the McMillan (1999) model and perceived interactivity:
H1The more features that add two-way communication, receiver control, a sense of place, time flexibility, activity, and information exchange to a Web site, the more interactive individuals will perceive the site to be.
For advertisers, a key question that grows from analysis of interactivity is whether interactivity has a relationship with attitude toward the Web site. The second hypothesis examines relationships between interactive functions at a Web site and attitude toward the site.
H2The more features that add two-way communication, receiver control, a sense of place, time flexibility, activity, and information exchange to a Web site, the more positive the attitude toward the Web site.
Wu (1999) found a positive relationship between attitude toward the Web site and perceived interactivity of the Web site. He argued that attitude toward the Web site is a critical measure of effectiveness that needs further exploration. Thus, hypothesis 3 explores relationships between perceived interactivity and attitude toward the Web site.
H3The higher the perceived interactivity of a site, the more positive the attitude toward the Web site.
McMillan’s (1999) study of interactivity found some evidence that people who are more involved with the subject of the Web site will find it to be more interactive. Ognianova (1998) also found involvement to be key factor in user’s perceptions of Web sites. Thus, this study further explores relationships between involvement and both interactivity and attitude toward the Web site:
RQ1What impact does a participant’s involvement in the subject matter of a Web site have on perceived interactivity of a Web site?
RQ2What impact does a participant’s involvement in the subject matter of a Web site have on attitude toward a Web site?
Finally, because our understanding of interactivity and attitude toward the Web site is still at an early stage of evolution, it is important to explore any demographic characteristics that may impact on either perceptions of interactivity or attitude toward the Web site:
RQ3What relationships exist between demographic characteristics of respondents and perceptions of interactivity of a Web site?
RQ4What relationships exist between demographic characteristics of respondents and attitude toward a Web site?
Method
Four Web sites were developed to test hypotheses 1 and 2. The sites were based on McMillan’s (1999) four models of cyber-interactivity. All four Web sites were about racewalking. This subject was chosen to test McMillan’s (1999) finding that involvement with the subject matter of a Web site might impact on perceptions of interactivity. While racewalking is an Olympic sport, it is not well recognized or understood outside of the community of track and field enthusiasts. Thus the subject offered the opportunity for further exploration of the role of involvement in perceived interactivity of sites.
All four Web sites used the same graphics, design, and structure developed by a graduate student with professional, computer-based design experience. All sites also included photographs of racewalkers taken at the 1996 Olympic Games in Atlanta by a sports photojournalist. Each site included four menu items: Technique, Schedule, Records, and Readings.
In all four sites, the Technique menu item led to a detailed tutorial on racewalking written by a track and field coach who is also a racewalking judge. One sub-menu provided detail on the walking rules. Another sub-menu led to training tips on the following topics: preliminaries, physiology, technique, pace/speed, and competition. The Schedule section provided both international and US schedules for major racewalking competitions for one year. Records were subdivided by gender and event. Finally, the Readings submenu led to information on two books about racewalking. The sites are no longer active, but html files are available from the author.
Each of the four test sites in this study was designed to match the corresponding model in the McMillan (1999) study. Table 1 summarizes key differences in the four sites based on dimensions identified in McMillan’s (1999) study. Changes were made to the sites to control for both of the key dimensions identified in Figure 1: direction of communication and control of the communication experience. Additional changes were made to address the remaining dimensions: time, place, activity, and perceived purpose. The Packaged Content site was the “base site” and additional interactive features were added to the remaining three sites based on the theoretical model outlined in Figure 1.
Table 1. Characteristics of Four Web Sites Examined in the Study
Features / Packaged Content / Rich Content / Virtual Transaction / Virtual CommunityTwo-way communication – Feedback forms / Yes / Yes
Control – Site map / Yes / Yes
Sense of place – Use of place-based metaphors / Yes / Yes
Time-sensitivity – Allow visitors to submit updates to records and schedules / Yes / Yes
Activity – Searchable records and schedules / Yes / Yes
Informative/Persuasive – Copy has an informative (as contrast with persuasive) tone / Yes / Yes
Two additional changes were made to further differentiate two-way models. The Virtual Transaction site linked to Amazon.com with encouragement for visitors to purchase reviewed books. The Virtual Community site added a chat room for real-time conversations.
Participants were drawn from two primary populations. The first were students in a major research university in the northeast. All were enrolled in computer-lab based classes. During a class session, they were asked to examine one of the four Web sites developed for this study. Students were randomly assigned to one of the four Web sites. A total of 80 students participated in the study.
The second group was recruited because of athletics activity – either involvement with one of three athletics and/or racewalking oriented listservs, or because these participants had been identified on a track and field Web site as active athletics leaders. This second group is referred to throughout the study as athletes. Each athlete received an e-mail invitation to participate in the study. The only incentive offered for participation was an abstract of study findings.
The recruitment message included a URL for the site to be examined (the URLs for the four sites were assigned randomly to recruits) and a unique ID number. The ID number could only be used once. The ID numbers were also compared with the recruiting list to ensure that each respondent had evaluated the appropriate Web site. A total of 98 athletes responded.
None of the participants were informed that four different sites existed. They all believed that they were simply evaluating a single Web site on the subject of racewalking. Participants were asked to view a site for 10-15 minutes and complete an online survey that was linked directly to the site. Analysis of participant ID numbers insured that each survey was appropriately matched with the site evaluated. A copy of the evaluation instrument is available from the author. The evaluation instrument was identical for all four sites.
Table 2 shows distribution of student and athlete evaluations. Cross tabulation of respondent type with site type revealed no significant difference in distribution (n = 178, df = 3, 2 = 4.64, p = .20).
Table 2. Distribution of Participants in Treatment Conditions
Site Examined / Students / Athletes / TotalPackaged Content / 27 / 20 / 47
Rich Content / 19 / 23 / 42
Virtual Transaction / 16 / 27 / 43
Virtual Community / 18 / 28 / 46
Total / 80 / 98 / 178
Perceived interactivity was measured using the scale developed by McMillan (1999). Respondents were asked to use a six-point Likert scale to indicate their level of agreement with seven statements. Table 3 shows each item in the scale. Alpha for the scale was .80.
Table 3. Interactivity Scale Items
StatementThis site facilitates two-way communication.
When I visit this site I get the sense that I am in a “place” in cyberspace.
This site seems to allow site visitors to communicate at times that are most convenient for them.
I feel that I have a great deal of control over my visiting experience at this site.
Visitors to this site need to take an active role in order to fully experience the site.
This site seems to be designed primarily to inform rather than persuade.
This site is interactive.
Attitude toward the Web site was measured with a scale based on traditional attitude toward the advertisement (Aad) measures commonly used in advertising, marketing, and consumer behavior research (e.g. Gardner, 1985; MacKenzie, Lutz & Belch, 1986; Mitchell & Olson, 1986). The seven-point semantic differential items were: bad/good, unpleasant/ pleasant, irritating/not irritating, boring/ interesting, dislike/like. Each of these items was reversed for statistical analysis. Alpha for the attitude toward the Web site scale was .92.