interactivity and its facets revisited

31 downloads 141 Views 204KB Size Report
professor of marketing and director of the Office of Scale Re- search, College of ..... of Internet-based interactivity would be a search engine's results when a ..... To test Hypotheses 1 to 5, an SEM (structural equation model- ing) analysis can be ...
INTERACTIVITY AND ITS FACETS REVISITED Theory and Empirical Test Grace J. Johnson, Gordon C. Bruner II, and Anand Kumar ABSTRACT: Interactivity is reviewed with the goal of addressing several important shortcomings in existing work: a need for generalizability of definitions across all situations and technologies; an explicit incorporation of the nonverbal facet of communication; and a reconsideration of control as an intrinsic aspect of all manners of interactivity. Based on this review, four facets of interactivity are proposed and tested experimentally. Three of the facets—responsiveness, nonverbal information, and speed of response—were found to be significantly related to perceived interactivity. It is important to note that support was also found for modeling interactivity as a second-order formative construct. Implications of the findings for interactivity theory, as well as for managers developing advertising strategies using interactive media, are discussed.

Internet-based advertising has been growing by leaps and bounds, with its year-on-year growth “vastly outstripping any other media channel” according to the Internet Advertising Bureau (Pilkington 2004). The evidence, however, is that the full potential of the Internet as an interactive medium is not being exploited; advertising continues to be a largely one-way monologue (Crain 2004). Experts estimate that despite the opportunity for interactivity provided by the Internet, the ratio of talking by advertisers to listening to what consumers have to say is of the order of 50:1 (Peters and Hessan 2003). The reason for this imbalance is that in spite of the ubiquitous nature of the term, advertisers have not fully understood what constitutes the multifaceted concept of interactivity, and have therefore failed to capitalize on its properties. We have identified four shortcomings with prior research in interactivity that have partly contributed to this lack of comprehension, which this work attempts to address. Our first concern has been the lack of comparability and generalizability of definitions. More specifically, we have noted that prior work has tended to present conceptualizations of interactivity that are anchored to specific technologies, and that are therefore difficult to generalize across all manner of interactive situations. Second, facets contributing to interactivity have been

introduced with little conceptual rationale. In this context, we discuss the inclusion of information control as an essential facet of interactivity in prior work, and argue that although control is an important construct in its own right, it is conceptually distinct from interactivity and its facets. A third shortcoming has been the lack of an explicit consideration of the nonverbal aspects of communication in interactivity research. Recent trends have called for an inclusion of the nonverbal aspects in communication research. We have adopted this approach to enhance understanding of interactive communication. Finally, there exist rich bodies of interactivity-related research in diverse areas such as advertising, communication, information systems, marketing, and educational psychology, which, until now, have been largely isolated. We have explored and used the concepts and empirical findings from these areas to offer insights that can benefit advertising scholars and managers. The proposed conceptualization of interactivity is verified in an experimental context where Web-based interactivity is manipulated with respect to its facets. Implications of the findings for researchers and managers developing advertising strategies using interactive media are discussed.

Grace J. Johnson (Ph.D., Southern Illinois University) is a lecturer at the School of Business Administration, University of Wisconsin–Milwaukee. Gordon C. Bruner II (Ph.D., University of North Texas) is a professor of marketing and director of the Office of Scale Research, College of Business and Administration, Southern Illinois University. Anand Kumar (Ph.D., Indiana University) is an associate professor of marketing, College of Business and Administration, University of South Florida.

The meaning of “interactivity” seems to depend on who you are and the context being referred to. Technologists, for example, typically define interactivity in terms of applications (e.g., the World Wide Web, VoIP, videoconferencing, on-line gaming, IM [Instant Messenger]) or features (e.g., hypertext, multimedia). However, rather than focusing on the

BACKGROUND

This research was funded by a grant to the first author, from the Pontikes Center for Management of Information, Southern Illinois University at Carbondale. Journal of Advertising, vol. 35, no. 4 (Winter 2006), pp. 35–52. © 2006 American Academy of Advertising. All rights reserved. ISSN 0091-3367 / 2006 $9.50 + 0.00. DOI 10.2753/JOA0091-3367350403

36

The Journal of Advertising

FIGURE 1 General Conceptualization of Interactivity

Perception of interactivity

Mediated (technology-based) interactivity

Nonmediated (behavioral) interactivity

General conceptualization of interactivity

technology per se, advertisers are more interested in how the technology can be used to add value to the communication process. To do so, they need a clear understanding of what constitutes the interactivity of such systems, and how these constituents can be modified or manipulated to produce different effects. Related to this is the need for definitions of interactivity to be generalizable across different situations or technologies. How does the interactivity of the Internet compare with that of virtual reality, for example? Can a definition be applicable to a yet-to-be envisioned technology of the future? To be useful and valid, conceptualizations of interactivity need to transcend contexts and applications. A close perusal of several recent definitions in advertising reveals that they tend to be rooted unnecessarily in technology rather than in the concept of interactivity itself (e.g., Coyle and Thorson 2001; Liu and Shrum 2002; McMillan and Hwang 2002 [see Table 1]). Interactivity in communication such as faceto-face discourse existed long before interactive devices were invented. Viewing interactivity too narrowly through the lense of a certain technology fails to provide a suitable basis on which advertisers and consumers can draw parallels between the interactivity in general life, which they are familiar with, and the interactivity of technological devices, which they are trying to understand. A second drawback of current work is that there is little theoretical rationale for what constitutes interactivity. Research has generated dimensions such as two-way communication, synchronicity, and control over the flow of information, but little by way of theory to explain why the aspect should be a part of interactivity. A glaring example of this is the case of control over the flow of information. Most users of this facet cite Steuer’s work (Steuer 1992 [see Table 1]), but a closer examination of

his work reveals that its context is virtual reality (VR) and the mediated interactivity that contributes to the experience of VR systems. His definition of interactivity, then, as the degree to which users of a medium can influence the form or content of the mediated environment (i.e., exert control), is applicable to such systems, but not necessarily to all types of interactivity.1 This study seeks to benchmark interactivity theoretically in terms of general human social experience (referred to as “behavioral interactivity” [Burgoon et al. 2002]). Such a general conceptualization of interactivity can help explain any mediated, technology-based interactivity as well as nonmediated, face-to-face (“FtF”) interactivity (see Figure 1). It can also account for human perceptions of interactivity, whether they are of mediated or nonmediated interactivity. Using this standard, we argue that reciprocity, responsiveness, nonverbal information, and speed of response can be justified as facets that contribute to interactivity, but that information control, which is often discussed as an essential aspect of interactivity, cannot. Prior work has also failed to explicitly account for nonverbal information as a distinct facet of interactivity. The rationale for this approach is explained in greater depth in the section that discusses this facet, but it is stressed here that for an advertiser to understand how different aspects of interactive communication can be manipulated to produce certain effects, nonverbal facet effects need to be considered distinctly. LITERATURE REVIEW When one ponders the meaning of “interactivity,” the first notion that may come to mind is that of bidirectional flow of information between a sender and a receiver in communication. This sounds very similar to the concept of “feedback” in communication, however, which generally refers to the receiver’s response to the sender. Hence, an important question to answer is whether or not “feedback” and “interactivity” are essentially the same. Fiske (1990), a communication researcher, views feedback in terms of two facets: availability of channels and access to channels. “Availability” refers to the extent to which multiple channels are used for communication (e.g., television uses verbal [text, voice] as well as nonverbal channels [pictures]). “Access” to channels refers to the extent to which all participants can use these channels (e.g., television offers very limited opportunity for viewers to respond to senders via the same channel). Does interactivity contain these aspects, too? Existing definitions of interactivity (see Table 1) were reviewed with the aim of identifying a number of variables that help contribute to a perception of interactivity. The criterion used was that the facets should have broad, rather than specific applicability, to a wide variety of “interactive” situations, both mediated and nonmediated. No standard dictionary carries a definition of “interactiv-

Marketing

Advertising

Information systems

Communication

Advertising

Marketing

Marketing

Bezjian-Avery, Calder, and Iacobucci 1998

Burgoon et al. 2000

Burgoon et al. 2002

Coyle and Thorson 2001

Deighton 1996

Deighton 1997

Field

Alba et al. 1997

Name

Consumer marketing using the Internet; using database technologies interphased with Internet technologies (mediated interactivity).

Marketers’ use of the Web to practice interactive marketing (mediated interactivity).

“Interactivity in commercial Web sites” (mediated).

Emerging communication technologies and FtF (behavioral and mediated interactivity).

HCI, CMC, and FtF communication (both behavioral and mediated interactivity).

Advertising and marketing using interactive systems such as the Internet (mediated interactivity).

Interactive electronic home shopping (mediated interactivity).

Context*: Behavioral or mediated interactivity

• Addressability • Responsiveness

(continues)

• Individual-level communication (as opposed to mass communication) • Degree of contingency or responsiveness

The term “interactive” points to two features of communication: the ability to address an individual, and the ability to gather and remember the response of that individual. Those two features make possible a third: the ability to address the individual once more in a way that takes into account his or her unique response. Addressability and responsiveness make a medium interactive. “Addressable” means the communication is directly addressable to individuals (not broadcast to all who can receive it); responsiveness means it is alert to the receiver’s response (it is no longer indifferent to its effect on the receiver).

• Range • Mapping • Speed

Dynamic qualities by which interactivity is experienced as interactive: • Degree of involvement • Interaction ease • Mutuality [Others, such as richness, spontaneity, expectedness, and desirability, may also have an influence.]

By “interactivity” is meant, in the media realm, some form of interdependent message exchange (based on Rafaeli 1988). Structural properties of media that enable interdependent interaction examined in this work: mediation, proximity, modality, and context richness.

Steuer’s (1992) definition is used: “[T]he extent to which users can participate in modifying the form and content of a mediated environment in real time.”

Three properties that create the qualitative experience of interactivity: • Interaction involvement • Mutuality • Individuation [Interaction and mutual involvement are explored.]

• Core dimension—ability to control information [Hierarchical traversal versus linear presentation of information.]

Interactive marketing is “the immediately iterative process by which customer needs and desires are uncovered, met, modified and satisfied by the providing firm.” None. Structural properties that can help distinguish FtF from HCI and CMC: participation, mediation, contingency, media and information richness, geographic propinquity, synchronicity, identification, parallelism, anthromorphism. Operationalized as “interaction involvement” and “mutuality.”

• Response time • Response contingency

Facets (either stated explicitly or implied in the discussion)

“In defining Interactive Home Shopping, we conceptualize interactivity as a continuous construct capturing the quality of two-way communication between two parties.” Two facets are response time and response contingency.

Definition

TABLE 1 Different Definitions of Interactivity in the Literature

Winter 2006 37

Strategy

Communication

Marketing

Advertising

Advertising

Communication

Communication

Heeter 2000

Hoffman and Novak 1996

Liu and Shrum 2002

McMillan and Hwang 2002

Neuman 1991

Newhagen and Rafaeli 1996

Field

Evans and Wurster 1997

Name

Communication on the Internet (mediated interactivity).

New media (mediated interactivity).

WWW (mediated).

“The emphasis of the current definition is on providing a concrete picture of consumers’ on-line communication” (mediated).

Computer-mediated communication (mediated interactivity).

New media with particular reference to “designed mediated experiences” (mediated interactivity, based on behavioral interactivity).

Strategy and the economics of information (mediated interactivity).

Context*: Behavioral or mediated interactivity • Dialogue

Facets (either stated explicitly or implied in the discussion)

“[T]he extent to which communication reflects back on itself, feeds on and responds to the past.”

“[T]he quality of electronically mediated communication characterized by increased control over the communication process by both the sender and the receiver, either can be a microprocessor.”

None. Different definitions in the literature are reviewed.

“The degree to which two or more communication parties can act on each other, on the communication medium, and on the messages and the degree to which such influences are synchronized.”

Use Rafaeli’s definition: “Interactivity is 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.”

• Feedback

• Control over the communication process

• Direction of communication (encompassing the concepts of responsiveness and exchange) • User control (“the way humans control computers and other new media”) • Time

• Two-way communication • Active control • Synchronicity. [Note: “system responsiveness is essential” to this dimension.]

• Feedback

• Actions followed by reactions Interactivity not defined. An interaction is an episode or series of episodes of physical actions and reactions of an embodied human with the world, including the environment and objects and beings in the world; conceptualization of interactivity is based on this. Focus only on “physical interactivity” (actions and reactions that can be observed), as opposed to perceived interactivity.

Interactivity is one aspect of richness of information; it refers to dialogue as opposed to monologue.

Definition

Table 1 (continued)

38 The Journal of Advertising

Communication technology

Communication

Communication

Information systems

Rogers 1986

Steuer 1992

Williams, Rice, and Rogers 1988

Zack 1993

Mediated interactivity of communication media and (behavioral) FtF interactivity.

Communication systems (mediated interactivity)

Virtual reality (mediated interactivity)

New communication technologies (mediated interactivity)

Mediated interactivity of CMCs; FtF (behavioral) interactivity also.

• Control • Exchange of roles • Mutual discourse • Channel bandwidth • Degree of personalization or social presence • Structural organization of interaction (e.g., continuous feedback)

No definition. Bases discussion on interaction theory in the sociology literature, and Rogers’s (1986) interactive model of the communication process, defined as one in which “participants create and share information with one another in order to reach a mutual understanding.”

• Speed of response • Range—the number of parameters that can be modified • Mapping—the way in which human actions are connected to actions within a mediated environment

• Feedback

• Feedback • Responsiveness [implied]

“The degree to which participants in a communication process have control over, and can exchange roles, in their mutual discourse is called interactivity.”

“[T]he extent to which users can participate in modifying the form and content of a mediated environment in real time” (p. 84). Speed of response is one important characteristic. Number of parameters that can be modified (range) is another factor contributing to interactivity, referring to the amount of change that can be effected on the mediated environment. Finally, mapping affects interactivity, referring to the way in which human actions are connected to actions within a mediated environment.

“The capability of new communication systems (usually containing a computer as one component) to talk back to the user, almost like an individual participating in a conversation.”

“Interactivity is 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.” Interactivity merges speaking with listening (Rafaeli 1997).

* The original context of each definition is important. For an explanation of behavioral and mediated interactivity, please see the “Background” section. Briefly, behavioral interactivity refers to face-to-face (nonmediated) interactivity: mediated interactivity refers to technology-mediated interactivity.

Note: HCI = human–computer interaction; CMC = computer-mediated communication; FtF = face-to-face; WWW = World Wide Web.

Communication

Rafaeli 1988

Winter 2006 39

40

The Journal of Advertising

ity.”2 Therefore, to begin with, we examined the root word, “interaction,” from (behavioral) interaction theory in sociology, which refers to reciprocal or mutual action (Goffman 1967). Mutual action refers to two sides participating in a communication exchange (the “access” aspect of feedback). From a perceptual sense, however, the issue is more complex. There are two different components to exchange. At one level, exchange refers to two sides participating in the communication by merely sending and receiving messages. In the literature, this has been termed “reciprocity,” “participation,” “mutual action,” “action-reaction,” and “two-way communication” (see Table 1, last column). A simple example in the context of Internet-based interactivity would be a search engine’s results when a query is typed followed by hitting “Enter” on the keyboard, or when a Web site responds to a mouse-click. There is action followed by reaction (Rafaeli 1988). At another level, the exchange can refer to much more than a reaction, where the response is contingent on, and directly relevant to, the message. Authors have referred to this component of two-way communication as “response contingency,” “mutuality,” “responsiveness,” “feedback,” and “mutual discourse” (Table 1, last column). A message is perceived as being highly “responsive” if it is appropriate, relevant, and sustains the continuity of the interaction. This may be termed the “responsiveness” component of interactive communication. A search engine working with natural language queries as opposed to the traditional Boolean queries, for example, often produces search results of higher relevance, giving rise to the perception of responsiveness. Thus, reciprocity may be seen as one facet of interactivity, and responsiveness, a second. We will now review these two facets of interactivity as well as two others, comparing the concept of interactivity to feedback. Facet 1: Reciprocity The concept of reciprocity has been widely acknowledged in the interactivity literature, although a variety of terms has been used to refer to essentially the same concept. Terms such as “dialogue” (Evans and Wurster 1997), “participate” (Steuer 1992), “iterative” (Bezjian-Avery, Calder, and Iacobucci 1998), “two-way communication” (Alba et al. 1997), “actions and reactions” (Heeter 2000), and “talking back” (Rogers 1986), all refer to this facet, meaning that consumers have the opportunity to participate jointly in conversing with firms as opposed to hearing a monologue from them. Williams, Rice, and Rogers (1988) describe participants exchanging roles, meaning that they alternately play the role of sender and receiver. This helps distinguish interactive communication (where senders take turns transmitting messages) from transactive communication (where senders transmit messages simultaneously).

Facet 2: Responsiveness Several definitions refer to “responsive” communication as a component of interactivity. Alba et al. (1997) identify “response contingency” and explain it as the degree to which the response made by one party is a function of the response made by the other party. A response that is directly related, and relevant, to the query is one that is contingent on it. Burgoon et al. (2002) also use the word contingency, but refer to the extent to which an interface produces answers tailored to a user’s queries, implying that answers need to be appropriate or relevant. They also describe interactivity as “interdependent message exchange,” agreeing that there is more to interactivity than reciprocity. This concept has been incorporated into many of the definitions in Table 1 (Hoffman and Novak 1996; Newhagen and Rafaeli 1996; Rafaeli 1988). Rafaeli comments that the most helpful definition of interactivity would be one grounded in responsiveness, pointing to the importance of this aspect. His definition and Deighton’s (1996) convey the meaning that there is a thread of continuity in interactive dialogue—the continuity is sustained by dialogue that is responsive. A communication that is poor in responsiveness (i.e., low in relevance and appropriateness), fails to fulfill the need of the other participant, thereby making it difficult to sustain the exchange. Human participants expect a high degree of responsiveness in interactive communication. Highly “responsive” communications, as defined here, are appropriate, relevant, and sustain the continuity of the interaction. The concept of feedback does not include this facet. Facet 3: Speed of Response Synchronicity, or speed of response, has been considered by many researchers to be an aspect of interactivity (Alba et al. 1997; Steuer 1992). It refers to the extent to which a message exchange occurs in real time or is delayed (Burgoon et al. 2002). When two parties communicate, a response that occurs with minimum delay contributes to the continuity of the interaction, whereas a delayed response will hinder communication flows. A perception of higher interactivity will, therefore, accompany a quicker response. This has been supported by research in educational technology that has found that as the temporal delay between action and reaction decreases, interactivity is perceived to be greater (Kirsh 1997). Thus, there appears to be some consensus as well as theoretical rationale for speed of response being a relevant aspect of interactivity. Again, feedback does not include this facet. Facet 4: Nonverbal Information In recent years there has been an explosion of research interest in the role of nonverbal behavior in communication (Buck and

Winter 2006

VanLear 2002). This corresponds to the “availability” facet of Fiske’s (1990) feedback construct, which refers to the extent to which multiple channels are used for communication. When more channels are used, feedback is richer. Researchers have questioned the traditional approach to studying communication, which ignores the nonverbal aspects. Instead, they have called for an explicit focus on the nonverbal aspects of communication so that a more holistic understanding of communication emerges. This trend has not, however, been reflected in the literature on interactivity, which has largely overlooked this explicit role of nonverbal components of communication. What, then, are “nonverbal aspects?” In any kind of communication, we rarely send out messages in a single channel. When a person speaks, the verbal element constitutes the primary message. The gestures, facial expressions, tone of voice, voice quality, pace of speech, inflexions, and loudness (the nonverbal elements), which accompany the verbal elements, enrich conversation and imbue it with meaning. This multiple channel situation exists even in print communication. Meaning is conveyed not only by the words in a print ad (the verbal element), but also by the nonverbal elements—the accompanying pictures, paralinguistic codes such as the size and type of the letters, their position on the page, their association with images, and the use of boldface, punctuation, and other typographical devices (see, for example, Townsend 1988). Multimedia research in educational psychology (literally, “multiple channels”) has investigated the use of video, sound, graphics, animation, and pictures (nonverbal elements), along with text-based information (the verbal element). Similarly, when we discuss nonverbal communication on the Internet, we refer to the use of graphics, animation, pictures, video, music, and sound, as well as paralinguistic codes, to present information. The communication/interactivity literature contains scattered references to nonverbal aspects. These references are tied to the structural features of technology (such as bandwidth) and do not explicitly consider nonverbal information as a facet. For example, Zack (1993) stresses the importance of simultaneous multiple-channel, verbal, and nonverbal cues in interactive communication, but uses the structural feature of “channel bandwidth” to convey this facet in mediated communication. In mediated communication, larger bandwidth enables transmission of information using multiple channels, which, in turn, is perceived as communication rich in the nonverbal aspects. Burgoon et al. (2002) discuss how the richness of nonverbal contextual information may reflect the interactivity of the communication. Therefore, we conclude first, that nonverbal effects are important enough to be considered distinct from those of other facets, and second, that communication that is abundant in nonverbal information will be perceived to be more interactive than communication that is poor on this facet.

41

We have, in this section, identified from a theoretical stance, and from the literature, a number of variables that help induce a perception of interactivity. We launched the discussion with a consideration of feedback. Interactivity, with the four facets described here, appears to be a richer, more complex concept than feedback with its two aspects. UNDERSTANDING INTERACTIVITY AND ITS FACETS Having reviewed existing definitions of interactivity, we offer a comprehensive conceptualization, based on behavioral interactivity, which synthesizes the facets. This conceptualization is presented in the following definition: Interactivity is the extent to which an actor involved in a communication episode perceives the communication to be reciprocal, responsive, speedy, and characterized by the use of nonverbal information.

The four facets identified here that constitute perceived interactivity are 1. reciprocity, or the extent to which communication is perceived to be reciprocal or to allow mutual action (Alba et al. 1997; Bezjian-Avery, Calder, and Iacobucci 1998; Heeter 2000; Rogers 1986; Williams, Rice, and Rogers 1988); 2. responsiveness, or the degree to which the responses in a communication are perceived to be appropriate and relevant, and resolving the information need of the interaction episode or event (Alba et al. 1997; Burgoon et al. 2002; Deighton 1996; Newhagen and Rafaeli 1996; Rafaeli 1988); 3. the speed of response, or the extent to which a response to a communication event is perceived to be immediate, or without delay (Alba et al. 1997; Burgoon et al. 2002; Steuer 1992); and 4. the extent to which the communication is perceived to be characterized by nonverbal information (the use of multiple channels for communicating information) (Burgoon et al. 2002; Zack 1993). This definition may be applied to any type of interactive situation—mediated or nonmediated. To illustrate how this definition can apply to any “interactive” technology, we consider company Web sites on the Internet. To determine whether a company Web site is highly interactive, the four facets can be used. If the Web site has many features that provide opportunities for users to act on the Web pages (links, clickable buttons, etc.), it will be perceived to be high on reciprocity. If every response provided by the Web site to a user’s actions is highly relevant and appropriate, it will be perceived to be high on responsiveness. If the Web site contains many sources

42

The Journal of Advertising

of nonverbal information (pictures, graphics, sounds, animation, etc.), it will be perceived to be high on this facet. If the Web site responds immediately to user commands or actions, it will be perceived to be speedy. As an example of behavioral interactivity, a highly animated FtF conversation between two friends is likely to be high on reciprocity (each comment by one of them is reciprocated with one from the other), high on responsiveness (only highly relevant responses encourage the conversation to continue), high on the speed (instantaneous responses), and high on nonverbal aspects (gestures, facial expressions, etc.). There may be facets other than these that constitute interactivity. The aim here, though, is to focus on those that fundamentally contribute to the meaning of the term “interactive.” The four facets likely work in concordance to enhance each other in creating a perception of interactivity, but exploring the exact nature of such interactions is left to future research. Some of the terms used in the definition are clarified: Interactivity takes place within a communication “episode,” which is the bounded (has a beginning and end) set of communication “events” constituting an occasion of interaction. A communication event is either an actor sending a message or one making a response. Interactivity can vary on a continuum from very low to very high (Rafaeli 1988). An aspect that is frequently mentioned as salient to interactivity is control (Ariely 2000; Bezjian-Avery, Calder, and Iacobucci 1998; Neuman 1991; Williams, Rice, and Rogers 1988). These authors all refer to mediated interactivity. Can control be considered an intrinsic aspect of behavioral interactivity? “Control” is defined as the ability to manipulate the duration or timing, content, and sequence of presented information (Ariely 2000). When a person has an interactive face-to-face dialogue, for example, with a salesperson, he or she can manipulate the salesperson’s presentation of information with respect to its duration, timing, content, or sequence only by participating in reciprocal communication, such as by stating, “Could you repeat what you just said?” Where reciprocity is low, as when a sales manager gives a speech without giving listeners the opportunity to participate, the ability of the receiver to manipulate the duration, content, and sequence of presented information is much reduced. Control is facilitated by reciprocity, and it is reciprocity that enables one to have control over the flow of information. Similarly, Steuer (1992) explains that in mediated communication, users have to participate in modifying the form and content of a mediated environment, implying that a user has to have access to reciprocal communication in order to be able to control a medium. Control cannot be easily generalized to all manner of interactive situations. Even in some instances of interactive mediated communication, control may not apply or be relevant. Consider reading print material: Even though there is more

control for the consumer than when watching television (per the definition above), print cannot be said to be more “interactive” than television (see Hoffman and Novak [1996]). We argue that information control cannot be equated with interactivity. We propose first, that the two concepts are distinct, and second, that the precursor to control is actually reciprocity, a facet of interactivity. Information control, though conceptually distinct, is likely to be highly correlated with reciprocity, explaining why it is often viewed as an inherent part of interactivity (Rafaeli 1988). Control is an important construct in its own right, with numerous useful outcomes. The empirical findings with regard to control that Ariely and others present are supported by extensive work in educational psychology, which has demonstrated that learner control, identical to control or control over the information flow described above, is a determinant of learning effectiveness and a range of related dependent variables (see Lawless and Brown [1997] for a review). A final clarification that has to be made concerns the formative nature of the interactivity construct. A formative construct is one where the direction of causality is from the measures to the latent construct, rather than the other way around (Bollen and Lennox 1991; Jarvis, Mackenzie, and Podsakoff 2003). In this case, the four first-order factors (the facets) cause variance in the second-order construct, interactivity. In other words, variance in interactivity is caused by changes in the levels of the four facets, and not the other way around. For example, a message with a high number of nonverbal elements leads to perceptions of that message being more interactive and not the other way around; it is not perceptions of interactivity that lead to a message being considered as high in number of nonverbal elements. Interactivity is thus represented as a second-order latent construct formed by the four facets indicated here. This is another clarification heretofore not presented in earlier research. EMPIRICAL VERIFICATION OF THE FOUR FACETS Figure 2 shows the research model used in the empirical study. Hypotheses 1 to 5 are based on evidence presented earlier.3 H1: There is a positive association between the extent of reciprocity in an episode of communication and the perceived interactivity of the episode. H2: There is a positive association between the extent of responsiveness in a communication episode and the perceived interactivity of the episode. H3: There is a positive association between the extent of nonverbal information in a communication episode (provided the different channels used convey the same information, or complement it) and the perceived interactivity of the episode.

Winter 2006

FIGURE 2 Model Tested

H4: There is a positive association between the speed of response in a communication episode and the perceived interactivity of the episode. H5: Perceived interactivity is a second-order formative construct, with reciprocity, responsiveness, nonverbal information, and speed of response being its four constituent facets. As part of the process of empirically validating the construct, its nomological validity should be established. To do this, the construct “perceived interactivity” is posited to relate to other constructs with which it is expected to have a relationship, on theoretical grounds. Two constructs with which perceived interactivity would be expected to have a positive relationship, in the context of Web-based communication, are attitude to the Web site (AWS) and user involvement. Web sites that are perceived to be more interactive are likely to be evaluated more favorably, and result in greater levels of involvement. Prior research in advertising has both hypothesized and empirically confirmed these two sets of relationships (Fiore and Jin 2003; McMillan and Hwang 2002; Wu 1999). Bosco (1986), Fletcher (1989), and Fletcher (1990) examined 75 learning studies in educational psychology and concluded that people learn faster and have better attitudes to the information presented in an interactive instructional environment. In addition, in the communication literature, Rafaeli and Sudweeks (1997) posited that interactive communication messages are more involving. Hence, based on this logic: H6: There is a positive association between the perceived interactivity of a communication episode and the user’s attitude to the Web site. H7: There is a positive association between the perceived interactivity of a communication episode and the user’s involvement with the communication. Methodology Interactivity was studied in the context of users interacting with a Web site to obtain information to accomplish a task. A computer-based 2 (reciprocity: low/high) × 2 (responsiveness: low/high) × 2 (nonverbal: low/high) × 2 (speed of response: low/high) between-subjects experimental study was designed to test the hypotheses. Preceding this, a pilot study involving 45 participants was conducted for the purpose of determining appropriate levels of each facet to be used to create high-low conditions. Stimuli and Participants A fictitious Web site for an on-line wine retailer named “The Wine Club” was constructed so as to permit a high degree of experimental control over what the participants viewed and

43

RECI

RECIPROCIT Y

+

ATTITUDE TO WEBEBSI SITE

+ RESPONSIVENESS

+

PERCEIVED INTERACTIVITY NONVERBAL INFORMATION

+ +

+

INVOLVEMENT

SPEED OF RESPONSE

did. The cover story was that the company wanted to better understand the kinds of wines young professionals would be interested in buying, and to elicit their input to help design an effective store Web site. Wine was selected as a product about which the participants would generally have little prior knowledge (this was confirmed in the preliminary study results). In prior studies involving multimedia, significant differences in learning effectiveness have been observed between high prior knowledge learners and low prior knowledge learners (Kunz, Drewniak, and Schott 1989; Mayer and Gallini 1990). Where nonverbal information was used, we ensured that the simple graphics and animation that accompanied the text showed closely related, supportive information (see Appendix 2). This is because multimedia research has shown that the mere presence of illustrations does not improve the learning or recall of textual information. The illustrations must be related to information presented in the text (Najjar 1996). In addition, the illustrations were made simple so as not to distract users. Too many sources of nonverbal information can be detrimental to the task at hand because of cognitive overload. If too much effort is involved in processing this information, mental resources available for other tasks at hand (such as comprehension) are reduced. Participants were asked to role-play recently graduated ad executive trainees hosting their first Christmas party for office colleagues and managers. Their task was to go through information on three brands of wine, after which they were to select one to serve at their party. Participants were randomly assigned to 1 of 16 experimental cells. The total information seen by participants in each cell was identical. However, the manner in which the information was presented to the participants varied as a function of the cell to which they were

44

The Journal of Advertising TABLE 2 Factor Intercorrelations and Reliabilities F1

F2

F3

F4

F1 (Reciprocity) F2 (Responsiveness) F3 (Nonverbal)

.77 .45* .40*

.91 .46*

.86

F4 (Speed of response)

.40*

.29*

.23*

.96

Average variance extracted (AVE)

.45

.66

.60

.84

Note: Diagonal values are composite reliabilities. Off-diagonal values are correlations. * Correlations significant at p < .01.

assigned. (See Appendix 2 for details on how each facet was operationalized.) Brand names were fictitious and pretested to ensure neutral pre-exposure brand evaluations. All responses and actions by the participant were captured real-time via an Access database interphased with the Web site. One hundred eighty juniors and seniors from an undergraduate business course at a large U.S. university completed the study in exchange for course credit. All instructions were provided on the computer to minimize extraneous and unmeasured effects that may have systematically biased the results. Measures To develop a measure of perceived interactivity, an item pool was initially generated based on the conceptualization described above, literature review, and interviews with 33 volunteers who were asked to generate descriptors to differentiate high and low interactivity. Sixty-nine distinct items were generated (248 before redundancies were eliminated) and classified into the four facets. Following this, a panel of 11 IT professionals and doctoral students was surveyed. The concept of interactivity was described as a formative construct constituted by the four facets and each judge was asked to classify the 69 items into: (1) highly representative, (2) somewhat representative, and (3) not at all representative of the facets. The 17 items that fell into the “highly representative” category were retained in the item pool and formatted into seven-point rating scales. (There were four items each for reciprocity, nonverbal information, and speed of response, and five items for responsiveness.) The items were perused once more to ensure that together they covered the entire domain of perceived interactivity. The wording of the items was then adapted for use in the particular computer-mediated interactive situation of the experiment. As a simple dependent measure, participants were asked

to rate the extent to which they agreed or disagreed with the description of the Web site as “interactive” (on a seven-point scale). Attitude to the Web site was measured using the threeitem measure used by Bruner and Kumar (2000, 2002). Involvement was measured using a 12-item validated adaptation of the Personal Involvement Inventory (Zaichkowsky 1985). These scales are all shown in Appendix 1. Results Manipulation Checks ANOVAs (analyses of variance) were run to determine whether participants’ perceptions of the low and high conditions of each facet differed significantly. In each case, there was a significant difference indicating that the experimental manipulations were successful: reciprocity: F(1, 178) = 32.13 (p < .01); responsiveness: F(1, 178) = 14.41 (p < .01); nonverbal: F(1, 178) = 488.62 (p < .01); speed: F(1, 178) = 68.87 (p < .01). Measurement Model EQS 6.1 (Bentler 2003) was used to test the measurement and structural models. Confirmatory factor analysis (CFA) was used to assess the psychometric properties of the multi-item scales employed to test the hypotheses. The confirmatory model of five constructs measured by 17 items revealed a good fit of the data. Several indices were examined to assess the overall fit of the model, including comparative fit index (CFI = .96), incremental fit index (IFI = .97), normed fit index (NFI = .90), non-normed fit index (NNFI = .96), goodness-of-fit index (GFI = .88), and adjusted goodness-of-fit index (AGFI = .84). Root mean square error of approximation (RMSEA) was .05. The correlations between the different constructs are shown in Table 2. The item reliabilities, as well as the convergent and discriminant validities, were examined per criteria suggested by Fornell and Larcker (1981). Every item loaded significantly on the construct it was supposed to measure (p < .01) (see Table 3). The composite reliability of each construct was calculated. Values were greater than .70 for all facets (see Table 2, diagonal). The average variance extracted by each construct was found to be greater than the squared correlations between that construct and every other construct in the study. The average variance extracted (AVE) was greater than the recommended .50 with all of the facets except for reciprocity (.45). Since it was just slightly below this criterion and met all of the other criteria, it was considered to be acceptable for use in the study. With this in mind, the scales showed sufficient evidence of unidimensionality, internal consistency, and convergent and discriminant validity to be included in the structural model.

Winter 2006

45

TABLE 3 Measurement Model Facet

Reciprocity

Responsiveness

Nonverbal information

Speed of response

Factor loadings

Items* V1. Perceived number of exchanges V2. Perceived number of actions V3. Extent of participation V4. Perceived number of responses V5. Relevance V6. Appropriateness V7. Met your expectations V8. Suitableness V9. Usefulness V10. Extent that nonverbal features were used V11. Extent of nontext information V12. Graphics-type information V13. Extent that pictures and graphics are used V14. Speed V15. Quickness V16. Extent of delay V17. Immediacy

Furthermore, the low- to moderate-factor intercorrelations suggested that multicollinearity would not be a problem affecting stability of parameter estimates in the structural part of the model. Hypotheses Tests To test Hypotheses 1 to 5, an SEM (structural equation modeling) analysis can be conducted. However, extra care must be taken since interactivity is a formative construct. We closely followed Jarvis, Mackenzie, and Podsakoff’s recommendations for formative latent construct analysis. We first classified our model as a Type II (reflective first-order, formative second-order) model, using their typology. The next issue was resolving the indeterminancy associated with the error term of the second-order formative construct (interactivity). Jarvis, Mackenzie, and Podsakoff recommend three possible ways to deal with this. Of these, we selected the third option: the formative construct must emit paths to one reflective indicator (we used the single-item perceived interactivity) and one latent construct with reflective indicators (we selected attitude to the Web site, AWS [α = .95]). The final issue in dealing with the formative construct was to ensure that the covariances among all exogenous latent constructs and manifest variables are estimated. The above structural model was run, including the latent construct, user involvement (α = .96), to test H7. The multivariate Lagrange Multiplier Test suggested that model fit

.58 .61 .73 .76 .70 .82 .81 .88 .83 .84 .71 .61 .90 .91 .93 .93 .90

t statistic 7.53 8.06 10.05 10.54 10.47 12.94 12.79 14.66 13.27 13.27 10.57 8.68 14.80 15.64 16.36 16.38 15.42

Item reliabilities .33 .37 .53 .58 .49 .66 .65 .78 .69 .71 .51 .38 .82 .82 .87 .87 .81

might improve significantly if a path were introduced from AWS to involvement. The suggested path made intuitive sense, as it seems very plausible that users’ attitudes toward the Web site would either draw them in or turn them away from the Web site, thus affecting their involvement with the site. Hence, we added the suggested path from AWS to involvement and reestimated the model. The results showed good overall fit: CFI = .97, IFI = .97, NFI = .86, NNFI = .97, GFI = .88, AGFI = .82, RMSEA = .04. The average of the off-diagonal absolute standardized residuals was low (.05), also confirming the good model fit. H2, H3, and H4 were supported, as responsiveness (β = .37, p < .01), nonverbal information (β = .41, p < .05), and speed of response (β = .28, p < .05) had significant effects on perceived interactivity. It was found that nonverbal information in communication exchanges was the most important determinant of user perceptions of interactivity. In testing H1, reciprocity was found to have a positive effect on perceived interactivity (β = .16, p = .08), but fell just short of statistical significance. H5 was supported as demonstrated by the overall model fit and the fact that the four facets together constituted 67.7% of the variance in the second-order construct, perceived interactivity. Support was also found for H6 and H7. Perceived interactivity was found to have very strong, positive effects on AWS (β = .65, p < .01) as well as user involvement (β = .42, p < .01), providing evidence for the nomological validity of the perceived interactivity construct.

46

The Journal of Advertising

DISCUSSION AND IMPLICATIONS One of the most important contributions of this study is the development of a definition of interactivity that is robust enough to be applied to a wide variety of situations, mediated or otherwise. This definition, which is independent of hardware or technology, was formulated based on a synthesis of work from diverse fields such as advertising, communication, education, and information systems. At the same time, congruities and differences between interactivity and other similar constructs in these fields such as feedback and control were delineated. In addition, this work investigated the nonverbal aspects of interactive communication, which have not been explicitly considered in prior work. Another important theoretical contribution of this study is that it empirically confirms the relationship between interactivity and its facets. It was found that user perceptions of the interactivity of communication are strongly influenced by the extent of nonverbal information in the exchange, its responsiveness, and by how fast a response is received. The nomological validity of the construct was demonstrated by showing that, as expected, attitude to the Web site and user involvement are positively influenced by the perceived interactivity of the communication. We also found strong empirical support for the theorized second-order interactivity factor, modeled as a formative construct constituted by the four facets. Therefore, interactivity can be defined as more than four distinct antecedents; it consists of the four facets, as well as the structure of interrelationships among these facets. Nonverbal information had the strongest effect on perceptions of interactivity. These effects have been extensively documented in multimedia studies in educational psychology, which provide theoretical reasons for some of the favorable effects associated with this facet. Multimedia provide multiple modalities for representing the real world (Hoffman and Ritchie 1997). According to dual coding theory, the human cognitive system has become specialized for dealing simultaneously with language and with nonverbal objects (Paivio 1986). The theory assumes that there are two cognitive subsystems, one specialized for the representation and processing of nonverbal objects/events, and the other for dealing with language (verbal information such as text or voice). Multimedia information encourages referential processing of information in a dual coding fashion (processing by both subsystems). This occurs, for example, when a person sees a picture of a car and also processes the word “car.” Research has also shown that referential (dual) processing has an additive effect on recall (Paivio 1991) because the learner creates more cognitive paths that can be followed to retrieve this information (Najjar 1996). Future research can explore these effects, tying interactivity with learning, comprehension, or recall. Responsiveness was found to have the second largest impact

on users’ perception of interactivity. Rafaeli’s (1988) comment that the definition of interactivity that would be most helpful would be one grounded in responsiveness finds strong empirical support here. As stated earlier, a communication that is poor in responsiveness makes it difficult to sustain or continue the interaction. For a communication to be “interactive,” it must, it is important to note, be responsive. Reciprocity’s theoretical importance was not supported by the findings. The results could imply that reciprocity is a necessary, but not sufficient condition for creating the perception of interactivity.4 Merely having the opportunity to participate in reciprocal communication may not guarantee that the communication is perceived as interactive (elaborated on below). Another reason for the results, which fell just short of statistical significance, could be the relatively low reliability of the subscale measuring this facet (.77). Further research can seek to confirm whether reciprocity does indeed play a weaker role than the other facets. The multifaceted model of interactivity proposed and validated in this study can be used to classify and understand all types of “interactive” technologies, based on the levels of the four facets, from exercise machines, videoconferencing, and video games, to virtual reality, blogs, VoIP (Voice over Internet Protocol), and interactive technologies yet to be envisioned. The model, therefore, need not be confined to the Internet. The findings of this study can suggest ways in which these applications can be modified to help increase the perception of interactivity, if needed. The results of this study can help managers develop more effective communication strategies for interactive media. The finding that nonverbal information has a relatively greater impact on user perceptions of interactivity than the speed of response can provide guidelines when managers consider trade-offs between using enhanced graphics and sounds, and increasing the speed of the Web site. This is especially relevant since providing more nonverbal information can often result in a lowered speed of response, due to bandwidth constraints. The reciprocity issue remains an interesting one for managers to consider and does call for further research. Is it better to have consumers click a number of times (high reciprocity) before the information they desire is provided them, or is a “one-click” (low reciprocity) approach more satisfactory? From the user’s point of view, the sooner the desired information is obtained, the better. From the company’s perspective, however, higher reciprocity may be more desirable. A real-world example of this dilemma occurred when Ticketmaster filed a lawsuit against Microsoft in 1997 for placing links on its Seattle Sidewalk Web site that linked directly deep inside Ticketmaster’s Web site rather than its front page. Ticketmaster wanted consumers to begin at its “front door,” and proceed via a series of clicks, thereby exposing them to the ads on each

Winter 2006

page. From its point of view, higher reciprocity was desirable, as it also meant that consumers’ visit duration would increase. From a research extension point of view, it would be interesting to explore to what extent consumers are willing to tolerate higher reciprocity conditions without becoming frustrated. Duration of interaction may be an important variable (Rogers 1986). Rogers suggests that interactivity may be desirable, but that its advantages may come “at the cost of more communication message exchanges and the greater time and effort required for the communication process” (p. 5). The salience of the responsiveness facet in the findings suggests that it is of great importance for the firm to show how it has been open to the user’s prior queries, actions, communications, or interactions, and that it has reacted to them. This would mean putting the rich data from consumer interactions to good use to optimize the consumer–firm relationship, as is done by Customer Relationship Management (CRM) programs. Firms sometimes find it difficult to retain users in spite of making their Web sites more interactive. It is possible that in these cases, increase in interactivity may not have involved an increase in responsiveness. The responsiveness findings also have implications for the effective design of Web sites. A problem sometimes encountered with Web portals illustrates this. Since space is limited on the front page of portals such as AOL, users are guided by brief headings or descriptions on the front page, which are hyperlinked deeper within the Web site. Headings can be misleading or inaccurate (low responsiveness). Clicking and not finding the implied information may lead to frustration, dissatisfaction, or disappointment, which can lead to a negative evaluation of the Web site. In one instance, AOL had a linked heading on its front page that read “Eat, Cheat. . . . Melt the Fat Away: Somers Diet,” accompanied by a picture of actress Suzanne Somers. A consumer may click on what appears to be an interesting article. After clicking on successive links, the user is actually led to the Barnes & Noble Web site where the actress’s book can be bought for $20. The consumer may react with frustration—what appeared to be an interesting story turns out to be no more than a cheap sales gimmick. Our research suggests that such communication actions of low responsiveness can erode consumers’ attitudes to the Web site, and possibly the brand as well (e.g., AOL). Web advertisers are able to devise innovative ways to get consumers to click on their ads, but they need to be aware of the possible negative consequences of such tactics. Our findings also suggest that there are different means of increasing interactivity. Firms must first consider why they wish to increase interactivity, and then decide which means will serve their interests best. As noted earlier, our results suggest that, in general, nonverbal information is the most important facet to enhance to increase perceived interactivity. Having said that, there may be occasions when the other facets should be

47

emphasized. For example, speed of response becomes critical if one’s goal is to close sales and get shoppers through the process as quickly as possible. Nonverbal information may be more important as users are browsing information and attempting to make sense of it. Nonverbal information, if it is designed in such a way as to complement textual information, can aid in the comprehension process. On the other hand, nonverbal information such as irrelevant graphics, animations, or sounds can serve to distract consumers and detract from the Web site’s main aims. Managers should understand why more of a facet may not always be better. CONCLUSION The interactivity construct was critiqued with a view to clarifying and generalizing it, and presenting a more complete picture of its constituent facets. In clarifying it, we distinguished interactivity from other concepts in the literature, such as feedback and control. In generalizing it, we expanded its conceptualization to include all manner of interactive situations, both technology-based and nonmediated. The nonverbal facet was presented to complement facets hitherto presented in the literature. The revised concept, which includes reciprocity, responsiveness, nonverbal information, and speed of response, was empirically confirmed. Where other studies have merely provided evidence of the relationships between interactivity and its dimensions, our study clearly demonstrated that the proposed facets together do, indeed, constitute a formative second-order latent construct. It was also demonstrated how this conceptualization can be used to explain a wide variety of interactive situations. From a managerial perspective, we stressed that as on-line ad spending continues to expand at a much higher rate than that for other media (Oser 2005), advertisers need to capitalize on the potential that Internet advertising offers for powerful oneon-one, two-way communication. As Frederick Webster stated during a Harvard discussion of the marketing potential offered by interactive technologies, “Marketers must stop thinking of marketing as persuasion (one-sided communication). . . . They must learn how to communicate with customers, listening to them as well as sending messages to them” (Webster, quoted in Deighton 1996). In the final analysis, if a “marketing orientation” (Levitt 1960) is all about a culture of discovering and satisfying customer needs, then marketers, advertisers, and the entire organization should be prepared to listen and respond appropriately when the customer speaks. NOTES 1. Steuer states: “For the purposes of this paper, interactivity . . . is a stimulus driven variable, and is determined by the technological structure of the medium” (1992, pp. 84–85).

48

The Journal of Advertising

2. We consulted the following dictionaries: Oxford English Dictionary, Merriam-Webster, Funk & Wagnalls New World Encyclopedia, American Heritage Dictionary, NetLingo, and the Internet Language Dictionary. 3. Although these associations appear intuitive, none of them has been verified empirically as yet. 4. We thank an anonymous reviewer for this insight.

REFERENCES Alba, Joseph, John Lynch, Barton Weitz, Chris Janiszewski, Richard Lutz, Alan Sawyer, and Stacy Wood (1997), “Interactive Home Shopping: Consumer, Retailer, and Manufacturer Incentives to Participate in Electronic Marketplaces,” Journal of Marketing, 61 (July), 38–53. Ariely, Dan (2000), “Controlling the Information Flow: Effects on Consumers’ Decision Making and Preferences,” Journal of Consumer Research, 27 (September), 233–248. Bentler, Peter M. (2003), EQS: Structural Equations Program Manual, Los Angeles: BMDP Statistical Software. Bezjian-Avery, Alexa, Bobby Calder, and Dawn Iacobucci (1998), “New Media Interactive Advertising Vs. Traditional Advertising,” Journal of Advertising Research, 38 (July/August), 23–32. Bollen, Kenneth, and Richard Lennox (1991), “Conventional Wisdom on Measurement: A Structural Equation Perspective,” Psychological Bulletin, 110 (2), 305–314. Bosco, James (1986), “An Analysis of Evaluations of Interactive Video,” Educational Technology, 26 (5), 7–17. Bruner, Gordon C., II, and Anand Kumar (2000), “Web Commercials and Advertising Hierarchy-of-Effects,” Journal of Advertising Research, 40 ( January), 35–42. ——, and —— (2002), “Similarity Analysis of Three AttitudeToward-the-Website Scales,” Quarterly Journal of Electronic Commerce, 3 (2), 163–172. Buck, Ross, and C. Arthur VanLear (2002), “Verbal and Nonverbal Communication: Distinguishing Symbolic, Spontaneous, and Pseudo-spontaneous Nonverbal Behavior,” Journal of Communication, 52 (3), 522–541. Burgoon, Judee K., Joseph A. Bonito, Jr., Bjorn Bengtsson, Jr., Artemio Ramirez, 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. ———, ———, Artemio Ramirez, Norah E. Dunbar, Karadeen Kam, and Jenna Fischer (2002), “Testing the Interactivity Principle: Effects of Mediation, Propinquity, and Verbal and Nonverbal Modalities in Interpersonal Interaction,” Journal of Communication, 52 (3), 657–677. 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. Crain, Rance (2004), “The Growing Impact of Consumers’ Web Publishing: Why Advertising Can No Longer Be a OneWay Monologue,” AdAge.com (November 1) (accessed November 1, 2004).

Deighton, John (1996), “The Future of Interactive Marketing,” Harvard Business Review, 74 (November/December), 151–162. ——— (1997), “Commentary on ‘Exploring the Implications of the Internet for Consumer Marketing,’” Journal of the Academy of Marketing Science, 25 (4), 347–351. Evans, Philip B., and Thomas S. Wurster (1997), “Strategy and the New Economics of Information,” Harvard Business Review, 75 (September/October), 71–82. Fiore, Ann Marie, and Hyun-Jeong Jin (2003), “Influence of Image Interactivity on Approach Responses Towards an Online Retailer,” Internet Research, 13 (1), 38–48. Fiske, John (1990), Introduction to Communication Studies, London: Routledge. Fletcher, Dexter J. (1989), “The Effectiveness and Cost of Interactive Videodisc Instruction,” Machine-Mediated Learning, 3 (4), 361–385. ——— (1990), “The Effectiveness and Cost of Interactive Videodisc Instruction in Defense Training and Education,” IDA Paper P-2372, Alexandria, VA: Institute for Defense Analyses. Fornell, Claes, and David F. Larcker (1981), “Evaluating Structural Models with Unobservable Variables with Measurement Error,” Journal of Marketing Research, 18 (February), 39–50. Goffman, Erving (1967), Interaction Ritual, Chicago: Aldine. Heeter, Carrie (2000), “Interactivity in the Context of Designed Experiences,” Journal of Interactive Advertising, 1 (Fall). Hoffman, Bob, and Donn Ritchie (1997), “Using Multimedia to Overcome Problems with Problem Based Learning,” Instructional Science, 25 (2), 97–115. Hoffman, Donna L., and Thomas P. Novak (1996), “Marketing in Hypermedia Computer-Mediated Environments: Conceptual Foundations,” Journal of Marketing, 60 (July), 50–68. Jarvis, Cheryl Burke, Scott B. Mackenzie, and Philip M. Podsakoff (2003), “A Critical Review of Construct Indicators and Measurement Model Misspecification in Marketing and Consumer Research,” Journal of Consumer Research, 30 (September), 199–218. Kirsh, David (1997), “Interactivity and Multimedia Interfaces,” Instructional Science, 25 (2), 79–96. Kunz, Gunnar C., Ute Drewniak, and Franz Schott (1989), “OnLine and Off-Line Assessment of Self-Regulation in Learning from Instructional Text and Picture,” paper presented at the annual meeting of the American Educational Research Association, San Francisco, March 27–31. Lawless, Kimberly A., and Scott W. Brown (1997), “Multimedia Learning Environments: Issues of Learner Control and Navigation,” Instructional Science, 25 (2), 117–131. Levitt, Theodore (1960), “Marketing Myopia,” Harvard Business Review, 38 (4), 45–56. Liu, Yuping, and L. J. Shrum (2002), “What Is Interactivity and Is It Always Such a Good Thing? Implications of Definition, Person, and Situation for the Influence of Interactivity on Advertising Effectiveness,” Journal of Advertising, 31 (4), 53–64. Mayer, Richard E., and Joan K. Gallini (1990), “When Is an

Winter 2006

Illustration Worth Ten Thousand Words?” Journal of Educational Psychology, 82 (4), 715–726. McMillan, Sally J., 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. Najjar, Lawrence J. (1996), “Multimedia Information and Learning,” Journal of Educational Multimedia and Hypermedia, 5 (2), 129–150. Neuman, Russell (1991), The Future of Mass Audience, Cambridge: Cambridge University Press. Newhagen, John E., and Sheizaf Rafaeli (1996), “Why Communication Researchers Should Study the Internet: A Dialogue,” Journal of Communication, 46 (Winter), 4–13. Oser, Kris (2005), “Online Ad Spending Hits New High,” AdAge.com (February 23). Paivio, Allan (1986), Mental Representations, New York: Oxford University Press. ——— (1991), “Dual Coding Theory: Retrospect and Current Status,” Canadian Journal of Psychology, 45, 255–287. Peters, Tom, and Diane Hessan (2003), “If Your Customers Could Talk . . . Oh, Wait a Minute, They Can!” Web seminar, American Marketing Association, January 23. Pilkington, Paul (director, Pricewaterhouse Coopers LLP) (2004), “Online Adspend: Beating All Records,” Internet Advertising Bureau/Price Waterhouse Coopers tracking research. Rafaeli, Sheifaz (1988), “Interactivity: From New Media to Communication,” in Advancing Communication Science: Merging

49

Mass and Interpersonal Processes, R. P. Hawkins, J. M. Wiemann, and S. Pingree, eds., Newbury Park, CA: Sage. ———, and Fay Sudweeks (1997), “Networked Interactivity,” Journal of Computer-Mediated Communication, 2 (4), http:// jcmc.indiana.edu/vol2/issue4/rafaeli.sudweeks.html. Robbins, Erica S., and Richard F. Haase (1985), “Power of Nonverbal Cues in Counseling Interactions: Availability, Vividness, or Salience?” Journal of Counseling Psychology, 32 (4), 502–513. Rogers, Everett M. (1986), Communication Technology: The New Media in Society, New York: Free Press. Steuer, Jonathan (1992), “Defining Virtual Reality: Dimensions Determining Telepresence,” Journal of Communication, 42 (Autumn), 73–93. Townsend, John (1988), “Paralinguistics: It’s Not What You Say, But the Way You Say It,” Management Decision, 28 (3), 36–40. Williams, F., Ronald E. Rice, and Everett M. Rogers (1988), Research Methods and the New Media, New York: Free Press. Wu, Guohua (1999), “Perceived Interactivity and Attitude Towards Websites,” in Proceedings of the American Academy of Advertising, M. S. Roberts, ed., Gainesville: University Press of Florida, 254–262. Zack, Michael H. (1993), “Interactivity and Communication Mode Choice in Ongoing Management Groups,” Information Systems Research, 4 (September), 207–239. Zaichkowsky, Judith Lynne (1985), “Measuring the Involvement Construct,” Journal of Consumer Research, 12 (December), 341–352.

50

The Journal of Advertising

APPENDIX 1 Seventeen-Item Perceived Interactivity Scale Based on the Four Facets Facet 1: RECIPROCITY: All on seven-point scales ranging from “very low” to “very high” Measures: V1. Every time you clicked on a wine attribute, such as “Age of wine” or “Where the grapes were grown,” the Web site responded. You would look at the information and then perhaps click again on another attribute. These are exchanges that took place between you and the Web site. What is your impression of the number of exchanges you had with the Web pages for this task? V2. Did the Web page require you to perform a low or high number of actions? V3. To what extent did you participate in the interaction with the Web site? V4. Did you feel that the number of times the Web site responded to your commands was low or high?

Facet 2: RESPONSIVENESS: All on seven-point scales Measures: V5. Please rate the relevance of the information you were shown when you clicked on the wine attributes as instructed. V6. Please rate the appropriateness of the information you were shown when you clicked on the wine attributes as instructed. V7. When you clicked on the wine attributes as instructed, you expected to get some information. Please rate the extent to which the information met your expectations. V8. How suitable was the information you received when you clicked on the wine attributes as instructed for the task at hand? V9. When you clicked on the wine attributes as instructed, you wanted to get some information that would be useful to you. Please rate the usefulness of the information you received.

Facet 3: NONVERBAL: All on seven-point scales Measures: V10. Many Web sites contain pictures, icons, graphics, animation, and colors to enhance your understanding of the material presented. Please rate the extent to which these features were used to describe the brands. V11. In your opinion, to what extent did the brand descriptions contain nontext information? “Nontext” refers to the use of anything other than words and numbers (e.g., pictures) to convey information. V12. Did you think the brand descriptions had more text-type information (words and numbers) or more graphics-type (pictures, colors, animation, etc.) information? V13. To what extent were pictures and graphics used to enhance your understanding of the brand descriptions? Facet 4: SPEED OF RESPONSE: All on seven-point scales

Measures: V14. Please rate the speed with which the Web pages responded to your commands. V15. Every time you clicked on parts of the Web page, how quickly did the Web site respond? V16. When you performed an action on the Web page, what was your impression of how much delay there was in obtaining a response? V17. What was your impression of the immediacy with which the Web pages responded to your comments?

Winter 2006

Dependent Measure Please indicate the extent to which you agree or disagree with the following description: The Web site was interactive. (seven-point scale ranging from 1 = strongly disagree to 7 = strongly agree)

Three-Item Attitude to the Web Site Scale Seven-point rating scales, ranging from “strongly disagree” to “strongly agree.”

Measures: 1. I like the Web site. 2. I think it is a good Web site. 3. I think it is a nice Web site.

Twelve-Item Involvement Scale Seven-point bipolar adjective scale.

Measures: 1. Unimportant/Important 2. Of no concern/Of concern to me 3. Irrelevant/Relevant 4. Means nothing to me/Means a lot to me 5. Worthless/Valuable 6. Not beneficial/Beneficial 7. Doesn’t matter/Matters to me 8. Boring/Interesting 9. Unexciting/Exciting 10. Unappealing/Appealing 11. Nonessential/Essential 12. Insignificant/Significant to me

51

52

The Journal of Advertising

APPENDIX 2 Operationalization of the Facets Construct

Conceptual definition

Operationalization/measurement

Reciprocity

The extent to which communication is perceived to be reciprocal or to allow mutual action.

Low: Participant clicks on any wine attribute listed on front page to find out more about it. Page opens up, revealing all wine information for the brand. PROCEED button appears so that participant can proceed to next brand. High: Participant has to click on eight different wine attributes. Each time, only a portion of the brand information is revealed. PROCEED button appears only when all attributes have been viewed.

Responsiveness

The degree to which the responses in a communication are perceived to be appropriate and relevant, and resolving the information need of the interaction episode or event.

Low: When a wine attribute such as “Age of wine” is clicked, information not directly relevant to the selected attribute (but related to wines) is revealed, such as “Light golden colored with citrus and spicy aromas.” High: When a wine attribute such as “Type of wine” is clicked, information that is appropriate and relevant is shown, such as “This is a White Zinfandel, a type of varietal wine.”

Speed of response

The extent to which a response to a communication event is perceived to be immediate or without delay.

Low: Each time a wine attribute is clicked, page takes 60 seconds to appear, while message reading “Please wait a moment” flashes. After 60 seconds, the wine information is revealed. High: Each time a wine attribute is clicked, page appears immediately, revealing wine information.

Nonverbal information

The extent to which communication is characterized by nonverbal information (the use of multiple channels for communicating information).

Low: Wine descriptions are in plain text, such as “A perfect lunchtime wine to accompany cold meat and salad.” High: Wine descriptions are accompanied by simple graphics, animations, and sounds next to descriptive terms. Next to the words “cold meat,” for example, there is a picture of a plate of meat, and next to the word “salad,” there is a bowl of salad. Next to the word “fizzy” in a description, there is a small animation of a bubbling glass of wine.