TY - JOUR
AU - McMahan, Carolyn
AB - Abstract Analysis of interactivity in Web sites is an important extension of a long tradition of analyzing content of media messages. But both interactivity and online analysis of content and features offer unique challenges to researchers. This study develops and tests a tool for measuring interactivity in the context of health-related Web sites. The tool was flexible enough to distinguish among multiple types of interactivity and powerful enough to show differences in interactivity based on domain type. Thus, it should have a relatively long life as a multifaceted tool for the tough job of measuring interactivity online. A Multifaceted Tool for a Complex Phenomenon: Coding Web-Based Interactivity as Technologies for Interaction Evolve Analysis of mediated messages has a long history in communication research and has been steadily evolving as a more sophisticated tool for analysis (Riffe & Freitag, 1997). Recent research has found content analysis to be a key methodology used in many different types of communication studies (Chang & Tai, 2005; Cho & Khang, 2006). Another growth area in communication research is examination of messages delivered via the internet (Cho & Khang, 2006; Tomasello, 2001). Relatively early in the study of internet communications, researchers began to note unique challenges in attempting to analyze content – particularly on the World Wide Web (McMillan, 2000b; Weare & Lin, 2000). But despite problems such as rapidly changing messages and increasingly complex structures, scholarly evaluation of Web sites has grown and evolved. Analysis has also come to include examination of features (De Marsico & Levialdi, 2004; Jankowski, Foot, Kluver, & Schneider, 2005; Trammell, Williams, Postelnicu, & Landreville, 2006). Feature analysis has been used extensively in the context of interactivity – a characteristic that is often seen as one of the critical differences between the internet and earlier communication media (Bucy, 2004; Cho & Cheon, 2005; Hwang, McMillan, & Lee, 2003; Massey & Levy, 1999; McMillan, 1999; McMillan, Kim, McMahan, & Fall, 2006; Schultz, 1999; Wise, Hamman, & Thorson, 2006). However, the absence of consistency in how researchers have coded Web site interactivity mars the ability to synthesize results and truly advance our understanding of this construct. Some early studies counted hyperlinks, a fairly basic component of all Web sites, as a way of measuring interactivity (Ha & James, 1998). More recent studies have focused on technological developments such as interactive spokes-characters (Phillips & Lee, 2005). The primary purpose of this study is to test a multifaceted and flexible tool for measuring interactive features of Web sites. A tool such as this one is an important “next step” in the evolution of content analysis as that method moves toward more rigorous implementation in online environments (McMillan, 2000b; Riffe & Freitag, 1997; Weare & Lin, 2000). Study of interactivity is also important to the evolving field of online communication (Chang & Tai, 2005; Cho & Khang, 2006; Kim & McMillan, Forthcoming). While earlier studies have attempted to code interactive content and features, no good tool has yet been developed that can keep pace with both changing technologies and changing perceptions of what makes Web sites interactive. Thus, it is critical that scholars have a rigorous, yet flexible tool for the tough job of analyzing Web-based interactivity. Literature An earlier review of the academic literature on interactivity (McMillan & Hwang, 2002) found studies that define interactivity based on features (the characteristics of the communication environment that make it interactive), processes (actually using an interactive feature), and perceptions (whether or not users perceive the communication environment to be interactive). While some authors have argued that interactivity should be defined primarily as a perceptual variable (Bucy & Chen-Chao, 2007; Liu, 2003), others have focused largely on process (Rafaeli & Sudweeks, 1997; Sundar & Kim, 2005). The literature is also rich with researchers who have defined interactivity based on the features of Web sites. This study focuses on that feature/content-based approach to interactivity. Content analysis of Web sites is difficult because of problems ranging from the selection of the sample to ensuring all coders analyze the same content in the same way (McMillan, 2000b; Weare & Lin, 2000). It is hard to capture the sampling unit in the online environment when content is customized by the user and is responding to user interactivity on that site. This precipitates the need for what is essentially an interactive tool to assess interactivity. Coding of Interactivity In the past 10 years, researchers have developed a variety of tools for measuring interactivity in the context of Web sites for topics ranging from politics to corporate communications (Ambre, Guard, Perveiler, Renner, & Rippen, 1997; Bedell, Agrawal, & Petersen, 2004; Bucy, 2004; De Marsico & Levialdi, 2004; Gonzalez & Palacios, 2004; Ha & James, 1998; Jankowski et al., 2005; Massey & Levy, 1999; McMillan, 1998; Okazaki, 2005; Paul, 2001; Yun, 2007). But what seems to be lacking in earlier studies examining interactive content of Web sites is an enduring way of coding interactive features. For example, none of the studies cited above included references to instant messaging – an interactive feature that has boomed in popularity in recent years. Additionally, little systematic work has been done in categorizing interactivity features or in determining how “deep” into a Web site researchers must go to determine overall interactivity of the site. Categorizing Interactivity Features New ways of implementing interactivity are sure to continue evolving. But rather than trying to anticipate new features, a coding scheme is needed that will provide a general framework into which those new technologies can be accounted for as they evolve. Scholars have begun to point to the importance of building content analysis measures on theoretical rather than technological foundations (Gerbic & Stacey, 2006). This study attempts to develop such systematic measures by considering multiple types of interactivity. Several researchers have suggested a three-dimensional construct of interactivity reviewed here as: human-to-computer, human-to-human, and human-to-content (Barker & Tucker, 1990; Haeckel, 1998; Jensen, 1998; McMillan, 2005). The first research question will use the content analysis tool developed for this study to broadly explore the nature of interactivity among these three basic types: RQ1: How are the three types of interactivity (human-to-computer, human-to-human, and human-to-content) implemented and what is their relationship to the overall interactivity of the Web site? The study also seeks additional insights into the nature of interactivity by further refining each of these types of interactivity based on characteristics found in the literature. Human-to-Computer The study of human-to-computer interactivity is not new. Early uses of the term interactive in the context of computers refers to a computer interface that was different from, and more user-friendly than, batch processing (Alter, 1977; Miles, 1992; Zeltzer, 1992). Human-to-computer interactivity may focus on navigation. Some studies focused on ways to categorize hyperlinks from sites, identify types of menu bars, and code for various search functions (Ha & James, 1998; McMillan, 2000a; McMillan & Downes, 2000; McMillan, Hwang, & Lee, 2003). Enabling actions such as online surveys and personalized login scripts is another form of interactivity that has been identified in some studies (Blattberg & Deighton, 1991; Bryant & Street Jr., 1988; Coyle & Thorson, 2001; Jensen, 1998; McMillan, 2000a, 2002, 2005). Finally, some scholars suggest the ultimate form of human-to-computer interaction is the ability to transact business online (Blattberg & Deighton, 1991; Coyle & Thorson, 2001; Haeckel, 1998; Lynn, Lipp, Akgun, & Crotez Jr., 2002; Pavlou & Stewart, 2000; Wells & Chen, 2000). Regardless of whether human-to-computer interaction is viewed as navigation, action, or transaction, some human-to-computer interactions are more personalized than others (McMillan, Kim, & Hwang, 2004). For example, standard navigational tools include menu bars and hyperlinks. But, if the developer adds customizable search engines, navigation becomes more personalized. The various implementations of human-to-computer interactivity lead to the following research question. RQ2: Within the context of human-to-computer interactivity: What is the overall distribution of navigation, action, and transaction features and are they more likely to be standard or personalized? How well do each of these subtypes of human-to-computer interactivity predict the overall level of human-to-computer interactivity at the site? Human-to-Human Human-to-human interactivity focuses on ways individuals interact with each other through computers. This tradition is based in human communication research. Some researchers position interactivity as being closely tied to new tools for facilitating old techniques of human communication. Fundamentally, media such as computer networks and telecommunication systems add a layer of technology between communicating partners (Chilcoat & DeWine, 1985). Human-to-human interactivity may focus on ways to facilitate communication between the organization and the visitors to the site through e-mail links, “Contact Us” pages, and so on (Levins, 1999; Liu & Shrum, 2002; Reardon & Rogers, 1988; Ruben, 1975). However, some scholars have suggested that it is important to also distinguish human-to-human interaction that enables individuals to communicate with other individuals through tools such as “tell a friend” links (Liu & Shrum, 2002; Moon & Nass, 1996; Rafaeli, 1988; Walther, 1996). Some scholars have focused on similarities and differences in human-to-human interaction when it occurs in synchronous or “real time” versus time-lagged asynchronous communication (Burke, Aytes, Chidambaram, & Johnson, 1999; Hanssen, Jankowski, & Etienne, 1996; Marvin, 1983; McGrath, 1991; McMillan & Hwang, 2002). The various implementations of human-to-human interactivity lead to the following research question. RQ3: Within the context of human-to-human interactivity: What is the overall distribution of features that allow for organization-to-individual and individual-to-individual communication and are they more likely to be asynchronous or synchronous? How well do each of these subtypes of human-to-human interactivity predict the overall level of human-to-human interactivity at the site? Human-to-Content People interact with the computer and with each other, but they also interact with the content of computer-mediated communication. This type of interaction takes place in two basic forms. First, is the ability to actually contribute to the content of the Web site. This may involve posting text messages in chat rooms or bulletin boards. Users may also be offered the opportunity to upload pictures, audio, video, or other elements. This type of contribution to content has been identified in several different studies that examine interactive features (Armstrong & Hagel, 1996; McMillan, 1998; McMillan et al., 2004; Schultz, 1999). Another conceptualization of human-to-content interactivity is that content dynamically responds to individual actions. Straubhaar and LaRose suggested that interactivity refers to situations where real-time feedback is collected from the receivers of a communications channel and is used by the source to continually modify the message as it is being delivered to the receiver (Straubhaar & LaRose, 1996). Similar concepts have been explored in applied fields such as education (Barker & Tucker, 1990; Hester, 1999) and marketing (Barker & Tucker, 1990; Blattberg & Deighton, 1991; Hester, 1999; Xie, 2000). The various implementations of human-to-content interactivity lead to the following research question. RQ4: Within the context of human-to-content interactivity: Are users more likely to be allowed the opportunity to add content or customize content? How well do each of these sub-types of human-to-content interactivity predict the overall level of human-to-content interactivity at the site? Depth of Analysis Researchers have not always agreed on how “deep” in a Web site one must go to measure interactivity. Some have focused on the opening page (Hwang, McMillan, & Lee, 2002; Paul, 2001) while others have analyzed two or more levels of the Web site (Okazaki & Rivas, 2002). Some research has suggested that it is important to go beyond the first page because users spend about twice the amount of time on interior pages as compared to home page viewing (Nielsen & Loranger, 2006). Recent research (McMillan et al., 2006) has suggested the best tradeoff between finding interactivity and the time it takes to find it is achieved by coding two levels of the Web site – the home page (level 1) and all pages accessed from it by a single “click-through” (level 2). The coding tool developed for this study will be used to further explore the question of how well overall interactivity can be predicted by features found at levels 1 and 2 of Web sites: RQ5: How well is the overall interactivity of a site predicted by the interactivity found at levels 1 and 2 of the site? Domain Type Research has begun to explore ways that different types of Web sites present information differently. For example, in recent research among older Americans, participants clearly differentiated between health-related .com sites that were seen as simply “peddling their wares” and government and educational sites that were perceived as being less biased and more information-rich (Macias & McMillan, 2008). A content analysis of diabetes Web sites offered confirmation of the critique of .com sites by including only one such site in its list of top-five sites related to the disease. That study also found that noncommercial sites were most likely to include content on a broad range of disease-related topics that have been determined to be important to patients (Bedell et al., 2004). The final research question explores differences in interactivity based on the domain type of sites: RQ6: What differences are found in interactivity within each of the following domain types: .com, .gov, .edu, .net, and .org. Method Sample Selection To examine the research questions posed by this study and evaluate the interactivity assessment tool, a purposive sample of 100 health-related Web sites was selected (20 each for each of the five domain types identified in RQ6). Health Web sites provide an interesting domain of study because a growing body of literature examines structural characteristics of internet health information (Berland, Elliott, Morales, & Algazy, 2001; Curro et al., 2004; Keltner, 1998; McMillan, 1999; Morahan-Martin, 2004; Risk & Petersens, 2002; Ritterband et al., 2006; Walther, Pingree, Hawkins, & Buller, 2005). Within the context of health sites, interactivity has been identified as an important feature for analysis (Ritterband et al., 2006). The sites were selected from the Yahoo! directory of health sites (found at http://dir.yahoo.com/Health/). This directory is an organized listing of health sites that has been successfully used in the past as a sampling frame for Web site analysis (McMillan, 1998). While it is not comprehensive (e.g., it does not include “hidden” Web sites), it does allow for the kind of purposive selection needed for this study. Five separate searches were conducted. A table of random numbers was used to identify the category within the list at which to start for each of the five searches. The first four sites of each of the five domain types (e.g., .com, .org, etc.) were selected within each category. This approach assured a purposive sample of domain types and also broadly distributed the health sites among a range of health topic areas. See Appendix 1 for a full list of all sites analyzed for the study. Interactivity Assessment Tool Development: Coding Scheme Unlike traditional static content analysis forms that ask coders to identify the presence or absence of features (Ambre et al., 1997; Bedell et al., 2004; Bucy, 2004; De Marsico & Levialdi, 2004; Gonzalez & Palacios, 2004; Jankowski et al., 2005; Okazaki, 2005; Paul, 2001), the interactivity assessment tool developed and tested in this study identifies all interactive site features in a way that is flexible, enduring, and interactive. Coders were not limited to noting predetermined features. Instead, if they found a feature that enabled interactivity, they wrote a brief description of that feature on the form indicating what type of interactivity was enabled and further categorizing the type based on key criteria identified in the literature. They also tallied all interactive features in any given category and for each of the two levels of the site being analyzed. The full coding sheet is presented in Appendix 2 and coding instructions are provided in Appendix 3. For human-to-computer interactivity, the coding form provides sections for coding navigation, action, and transaction. Within each of these primary types of human-to-computer interactivity, the coding scheme separately notes tools that are standard and those that are more personalized. In the context of human-to-human interaction, all features that allow organization-to-individual communication as well as those that facilitate individual-individual communication were recorded. For both types of human-to-human interaction, features that allow real-time (synchronous) communication and those most likely to occur with a time lag (asynchronous) were assessed. For human-to-content communication, the coding scheme captures features that enable site visitors to add content and those that enable customization. To validate this coding scheme an additional step was taken. Thirteen researchers who have studied interactivity (all have published one or more articles on the topic of interactivity in the context of the internet) were identified (these researchers are referred to hereafter as experts). All of the experts were presented with a very brief (one short paragraph) definition of each of the types of interactivity identified earlier in this study. All of the features that were identified for the current study (see Appendix 2) were presented in random order. Experts were asked to indicate which categories of interactivity were most closely aligned with each feature. For every individual feature, the most commonly selected category was the same category that had been identified by the study authors. Table 1 summarizes the overall agreement between experts and authors. Table 1 Agreement between experts and authors on feature categories Feature . % Experts Agreeing on Category . Navigation tools (menus, text links, search etc.) 71.8 Actions (games, surveys, registration, etc.) 71.2 Transactions (order physical and/or digital items) 71.2 Organization/individual communication (chat, e-mail, feedback, etc.) 69.2 Individual/individual communication (chat, send virtual postcard, etc.) 76.9 Add content (text, graphics, audio, etc.) 80.8 Customize content (select language, etc.) 100.0 Feature . % Experts Agreeing on Category . Navigation tools (menus, text links, search etc.) 71.8 Actions (games, surveys, registration, etc.) 71.2 Transactions (order physical and/or digital items) 71.2 Organization/individual communication (chat, e-mail, feedback, etc.) 69.2 Individual/individual communication (chat, send virtual postcard, etc.) 76.9 Add content (text, graphics, audio, etc.) 80.8 Customize content (select language, etc.) 100.0 Open in new tab Table 1 Agreement between experts and authors on feature categories Feature . % Experts Agreeing on Category . Navigation tools (menus, text links, search etc.) 71.8 Actions (games, surveys, registration, etc.) 71.2 Transactions (order physical and/or digital items) 71.2 Organization/individual communication (chat, e-mail, feedback, etc.) 69.2 Individual/individual communication (chat, send virtual postcard, etc.) 76.9 Add content (text, graphics, audio, etc.) 80.8 Customize content (select language, etc.) 100.0 Feature . % Experts Agreeing on Category . Navigation tools (menus, text links, search etc.) 71.8 Actions (games, surveys, registration, etc.) 71.2 Transactions (order physical and/or digital items) 71.2 Organization/individual communication (chat, e-mail, feedback, etc.) 69.2 Individual/individual communication (chat, send virtual postcard, etc.) 76.9 Add content (text, graphics, audio, etc.) 80.8 Customize content (select language, etc.) 100.0 Open in new tab The one instance in which the overall agreement between experts and study authors dropped to below 70% was organization/individual asynchronous communication. Many of the experts coded these as fitting in the human-computer interaction tradition. This is understandable in the context of corporate Web sites where navigating to an e-mail feedback form may not, in itself, lead site visitors to believe that they will interact with another human. The high level of agreement between experts and study authors, adds support to the coding tool and its three dimensions of human-to-human, human-to-computer, and human-to-content interactivity. However, it also points to the importance of coder training. Any researcher who wishes to use this tool will need to carefully describe for coders why a given feature is being coded in a certain way. For example, it may be that some authors want to focus on actual interchanges and thus would code e-mail links as interactive only if the researcher receives a personalize reply from the site creator. Interactivity Assessment Tool Development: Coding Instructions All interactive features found on both the first page (level 1) of the Web site and interior pages accessed with a single “click through” (level 2) were coded. While this type of content analysis form is more complex than a typical form that simply assesses presence or absence of a feature or type of content, it is also much more flexible for evaluating interactivity, which is a rapidly changing technological feature. It is also rigorous in that no form of interactivity is omitted simply because it was not listed as a predetermined option. This form is designed to be applicable as an interactivity assessment tool across a variety of contexts while also adapting to the ongoing evolution of the internet as a site for interactivity. One key instruction given to coders was that level 2 coding should include only NEW features. Thus if standard navigational tools (e.g., menu bars) appear on both levels 1 and 2, they would be coded only as level 1 features. Coders were also instructed to code only the content of the “home” site – not of any external sites to which it linked. Within all types of interactivity, “like items” on a page were coded as one “type” of interactivity. This was done to try to reduce the influence of content on the coding of interactivity. For example, a content-rich site might have many more hyperlinks on the front page than a site with less content. But this study is interested in the fact that a site uses hyperlinks as a standard navigational tool rather than in how many hyperlinks are used. Coder Training and Intercoder Reliability Three coders were trained to code for interactivity using the assessment tool developed for this study. All became familiar with the literature on interactivity and then coded a group of five “test sites” that were selected to represent a broad range of health-related Web sites. With very minor changes to the coding forms and instructions, coders were able to reach agreement on the coding of the test sites. Examination sites were distributed to coders in a stratified manner so that each coder reviewed sites from all domain types. The coders evaluated their assigned sites within a 2-week period. Content analysis generally requires that 10-20% of the content be checked for intercoder reliability (Riffe, Lacy, & Fico, 1998). A total of 15 of the 100 sites were checked for intercoder reliability. As Bucy noted, it is difficult to use standard measures of intercoder reliability when coding for interactive features – particularly in a study that allows for a virtually unlimited number of different coding options (Bucy, 2004). Intercoder reliability was checked in two ways. First, reliability was considered in terms of how often coders agreed simply about whether a particular kind of interactivity exists. This method is similar to traditional measures of intercoder reliability that measure the ability of coders to identify and code a phenomenon (Berelson & Lazarsfeld, 1948; Holsti, 1969). Thus reliability was measured using Holsti’s (1969) formula for comparing agreement on whether each type of interactivity (e.g., standard navigation) were present at a given site. Reliability was tested for each pair of coders. Using this measure, average reliability was 85.90% which is within what is generally considered an acceptable range (Riffe et al., 1998). A second, and more rigorous, measure for testing reliability would be to compare the total number of features found by coder A and those found by coder B for each coding pair. Overall, the average for this measure of reliability was 70.12%. What seems to have happened is that coders were able to reliably identify the presence or absence of specific features comparable to applying the content analysis method to traditional media. However, given the complexity of the coding form, they sometimes missed an instance of the feature resulting in slightly different counts for numbers of features. In all cases where exact agreement was not achieved, a fourth trained coder reviewed the coding with the initial pair of coders and all three came to agreement on feature coding. As expected, most differences were because of coder oversights. Findings Interactivity Prevalence and Predictability between Levels In the 100 sites coded, an average of 21.25 total interactive features were found per site with a range of 0 to 56 features. Table 2 addresses research question 1 with a summary of the prevalence of the basic types of interactivity and how they were distributed. Table 2 Types of Interactivity by Level (N = 100) . Human-to-Computer . Human-to-Human . Human-to-Content . Level 1 Features mean = 7.89; range 0–25 mean = 1.31; range 0–5 mean = .43; range 0–4 mean = 9.66 r = .962*** r = .292*** r = .403*** Level 2 Features mean = 8.92; range 0–31 mean = 2.22; range 0–8 mean = .40; range 0–3 mean = 11.59 r = .963*** r = .594*** r = .243** Total Features mean = 16.81; range 0–53 mean = 3.53; range 0–11 mean = .83; range 0–6 mean = 21.25 r = .968*** r = .502*** r = .338*** . Human-to-Computer . Human-to-Human . Human-to-Content . Level 1 Features mean = 7.89; range 0–25 mean = 1.31; range 0–5 mean = .43; range 0–4 mean = 9.66 r = .962*** r = .292*** r = .403*** Level 2 Features mean = 8.92; range 0–31 mean = 2.22; range 0–8 mean = .40; range 0–3 mean = 11.59 r = .963*** r = .594*** r = .243** Total Features mean = 16.81; range 0–53 mean = 3.53; range 0–11 mean = .83; range 0–6 mean = 21.25 r = .968*** r = .502*** r = .338*** ** p < .01, *** p < .001 Open in new tab Table 2 Types of Interactivity by Level (N = 100) . Human-to-Computer . Human-to-Human . Human-to-Content . Level 1 Features mean = 7.89; range 0–25 mean = 1.31; range 0–5 mean = .43; range 0–4 mean = 9.66 r = .962*** r = .292*** r = .403*** Level 2 Features mean = 8.92; range 0–31 mean = 2.22; range 0–8 mean = .40; range 0–3 mean = 11.59 r = .963*** r = .594*** r = .243** Total Features mean = 16.81; range 0–53 mean = 3.53; range 0–11 mean = .83; range 0–6 mean = 21.25 r = .968*** r = .502*** r = .338*** . Human-to-Computer . Human-to-Human . Human-to-Content . Level 1 Features mean = 7.89; range 0–25 mean = 1.31; range 0–5 mean = .43; range 0–4 mean = 9.66 r = .962*** r = .292*** r = .403*** Level 2 Features mean = 8.92; range 0–31 mean = 2.22; range 0–8 mean = .40; range 0–3 mean = 11.59 r = .963*** r = .594*** r = .243** Total Features mean = 16.81; range 0–53 mean = 3.53; range 0–11 mean = .83; range 0–6 mean = 21.25 r = .968*** r = .502*** r = .338*** ** p < .01, *** p < .001 Open in new tab The means in each cell of Table 2 (as well as Tables 3 and 4) are for average number of features of a type (column label) within the category defined by the row label (e.g., there were an average of 7.89 human-to-computer features at level 1). While standard deviations would provide similar information to the ranges shown in each cell, we report minimum and maximum number of features to provide depth of understanding of variance in implementation of features. The r values in Tables 2–4 report Person correlation coefficients for bivariate correlations between average features in a column (e.g., human-to-computer features) and total features for the row (e.g. level 1). Thus, the r of .962 in the upper left portion of Table 2 indicates human-to-computer features at level 1 are strongly correlated with overall interactivity at level 1. Table 3 Human-to-Computer Interactivity (N = 100) . Navigation . Action . Transaction . Standard mean = 11.68; range 0–39 mean = .74; range 0–5 mean = 1.16; range 0–7 mean = 13.58 r = .953*** r = .402*** r = .533*** Personalized mean = 1.73; range 0–6 mean = 1.44; range 0–9 mean = .06; range 0–2 mean = 3.23 r = .691*** r = .729*** r = .219** Total Human-to-Computer mean = 13.41; range 0–43 mean = 2.18; range 0–13 mean = 1.22; range 0–7 mean = 16.81 r = .962*** r = .624*** r = .490*** . Navigation . Action . Transaction . Standard mean = 11.68; range 0–39 mean = .74; range 0–5 mean = 1.16; range 0–7 mean = 13.58 r = .953*** r = .402*** r = .533*** Personalized mean = 1.73; range 0–6 mean = 1.44; range 0–9 mean = .06; range 0–2 mean = 3.23 r = .691*** r = .729*** r = .219** Total Human-to-Computer mean = 13.41; range 0–43 mean = 2.18; range 0–13 mean = 1.22; range 0–7 mean = 16.81 r = .962*** r = .624*** r = .490*** ** p < .01, *** p < .001 Open in new tab Table 3 Human-to-Computer Interactivity (N = 100) . Navigation . Action . Transaction . Standard mean = 11.68; range 0–39 mean = .74; range 0–5 mean = 1.16; range 0–7 mean = 13.58 r = .953*** r = .402*** r = .533*** Personalized mean = 1.73; range 0–6 mean = 1.44; range 0–9 mean = .06; range 0–2 mean = 3.23 r = .691*** r = .729*** r = .219** Total Human-to-Computer mean = 13.41; range 0–43 mean = 2.18; range 0–13 mean = 1.22; range 0–7 mean = 16.81 r = .962*** r = .624*** r = .490*** . Navigation . Action . Transaction . Standard mean = 11.68; range 0–39 mean = .74; range 0–5 mean = 1.16; range 0–7 mean = 13.58 r = .953*** r = .402*** r = .533*** Personalized mean = 1.73; range 0–6 mean = 1.44; range 0–9 mean = .06; range 0–2 mean = 3.23 r = .691*** r = .729*** r = .219** Total Human-to-Computer mean = 13.41; range 0–43 mean = 2.18; range 0–13 mean = 1.22; range 0–7 mean = 16.81 r = .962*** r = .624*** r = .490*** ** p < .01, *** p < .001 Open in new tab Table 4 Human-to-Human Interactivity (N = 100) . Asynchronous Features . Synchronous Features . Organization/Individual mean = 2.74; range = 0–8 mean = .04; range = 0–2 mean = 2.92; range = 0–8 r = .990*** r = .080 Individual/Individual mean = .46; range = 0–6 mean = .09; range = 0–2 mean = .61; range = 0–6 r = .944*** r = .458*** Total Human-to-Human Features mean = 3.40; range = 0–9 r = .981*** mean = .13 range = 0–2 r = .393*** . Asynchronous Features . Synchronous Features . Organization/Individual mean = 2.74; range = 0–8 mean = .04; range = 0–2 mean = 2.92; range = 0–8 r = .990*** r = .080 Individual/Individual mean = .46; range = 0–6 mean = .09; range = 0–2 mean = .61; range = 0–6 r = .944*** r = .458*** Total Human-to-Human Features mean = 3.40; range = 0–9 r = .981*** mean = .13 range = 0–2 r = .393*** *** p < .001 Open in new tab Table 4 Human-to-Human Interactivity (N = 100) . Asynchronous Features . Synchronous Features . Organization/Individual mean = 2.74; range = 0–8 mean = .04; range = 0–2 mean = 2.92; range = 0–8 r = .990*** r = .080 Individual/Individual mean = .46; range = 0–6 mean = .09; range = 0–2 mean = .61; range = 0–6 r = .944*** r = .458*** Total Human-to-Human Features mean = 3.40; range = 0–9 r = .981*** mean = .13 range = 0–2 r = .393*** . Asynchronous Features . Synchronous Features . Organization/Individual mean = 2.74; range = 0–8 mean = .04; range = 0–2 mean = 2.92; range = 0–8 r = .990*** r = .080 Individual/Individual mean = .46; range = 0–6 mean = .09; range = 0–2 mean = .61; range = 0–6 r = .944*** r = .458*** Total Human-to-Human Features mean = 3.40; range = 0–9 r = .981*** mean = .13 range = 0–2 r = .393*** *** p < .001 Open in new tab Human-to-computer interactivity is the dominant form of interaction at these sites. Site navigation tools include using menus and submenus, text and graphic links, drop-down boxes, search functions or site maps. Completing a survey, member registration or taking a virtual tour offered actions users could take. Ordering a physical item, downloading a document or Powerpoint ™ presentation or requesting a customized report are transaction examples. An average of about eight types of human-to-computer interaction are found on opening pages and about nine additional features are added at level 2. Because this is a dominant form of interaction, it is not surprising that there are significant positive correlations between human-to-computer interaction and total interactive features at both levels 1 and 2 and for total number of features. Human-to-human interaction is the second most common form of interaction with one feature on the average opening pages and slightly more than two additional features found on subsequent pages. Almost all sites include some kind of “contact us” link at level 1. Additional contact information appearing at level 2 is often for specialized e-mail addresses for different groups within the organization (e.g., ask a nurse, contact the public relations department, etc.). Human-to-content interaction was the least common form of interactivity found in this sample of health-related Web sites. Examples included posting to message boards and customizing content, usually in the form of viewing a Spanish mirror site or enlarging the font or image. However, sites that did have human-to-content interactivity were also more likely to be among the most interactive sites. Human-to-Computer Interactivity Research question 2 related to human-to-computer interactivity, which was the most common and most complex form of interactivity examined. Features were divided into those that were standard with the same functionality for all visitors and those that were personalized so that they were responsive to specific inputs of individual visitors. Additionally, features were coded as facilitating navigation, action, or transaction. As shown in Table 3, personalized features were found less frequently overall with an average of 3.23 features per site compared to 13.58 standard features (t = 17.09, df = 99; p < .001). Significant differences were also found for each type of human-to-computer interactivity. Standard features were more common than personalized features for both navigation (t = 17.23, df = 99, p < .001) and transaction (t = 7.46, df = 99, p < .001). But for action features, personalized features were more common than standard features (t = 6.15, df = 99, p < .001). Table 3 also shows that the most common form of human-to-computer interaction is standard navigation with an average of slightly more than 11 features per site. This includes menus, buttons, hyperlinks, etc. Standard navigational tools are a necessary component of interactivity. It is important to note that the coders only included tools that were unique to the site and not part of the browser function (e.g., scroll bars). Also, the coding scheme focused on different “types” of standard navigation tools. Thus a menu bar that appeared at the top of the page was counted as only one navigation tool regardless of how many items were on that menu. Thirty-two percent of the sites had no personalized navigation tools at all and the average number of these features was about two. These features include search tools, options for limiting content the site visitor wishes to view, and so on. Because of the dominance of navigation features it is not surprising that navigation features correlate strongly with standard, personalized, and overall human-to-computer features. Action features enable site visitors to participate in games and activities that are standard for all users and to take actions such as log-in and surveys that are more personalized to the individual. As noted earlier, this was the only form of human-to-computer interaction for which personalized functions were dominant – usually driven by activities that require a log in. Transactions allow users to transact some type of exchange with the site creators. This might involve activities that are the same for all visitors such as ordering a book or an online newsletter. Some transactions are more personalized such as the ability to order a customized report based on individual health symptoms. As illustrated in Table 3, standard transactions are more common than personalized ones. In this sample, the most common types of transactions were ordering books, brochures, reports, etc. Only 4% of the sites had personalized transaction – thus it is not surprising that weaker correlations were found between transactions and personalized features. However, overall transactions do appear to be a fairly good predictor of human-to-computer interaction. Human-to-Human Interactivity Research question 3 related to human-to-human interaction, which was considered in terms of the interactive dyad and the timing of the communication. Two interactive dyads were coded. The first was organization-to-individual interaction. Second was individual-to-individual interaction in which the site visitors were given tools to communicate with any individual that they choose on some matter related to the site. Timing was considered in terms of asynchronous functions in which there is no expectation for immediate response (e.g., e-mail) and synchronous functions that have a high expectation for immediate response (e.g., instant messaging). As illustrated in Table 4, overall organizational/individual interactivity was significantly higher than individual/individual with means of 2.92 and .61 features respectively (t = 16.96, df = 99, p < .001). This same pattern of stronger organization/individual communication was also seen for asynchronous (t = 5.09, df = 99, p < .001) but not for synchronous (t = 1.67, df = 9, p = .05) human-to-human interactivity features. The most common form of human-to-human interaction at these sites was asynchronous communication between the organization and the individual. This was most often a link or form that enables site visitors to send e-mail to organizations and individuals associated with the site. Ninety-one percent of sites had at least one of this type of feature. It is important to note here that each “type” of e-mail link was coded separately but e-mail addresses were not totaled. So, for example, a link that enabled site visitors to e-mail staff members would be coded as one feature regardless of how many staff member e-mail addresses were present. Only three percent of the sites offered synchronous communication between the organization and the individual. The second most common form of human-to-human interaction at these sites was tools that facilitated asynchronous communication between site visitors and other individuals. However 70% of sites had no such tools resulting in a mean of less than .5 overall. The most common tool for facilitating individual-to-individual interaction was a link that enabled the site visitor to e-mail site content to a friend. Few tools for facilitating synchronous communication were found – only three “live chat” functions for customer support (organization/individual) were found. Six sites provided links to instant messaging functions for individual/individual synchronous communication. Human-to-Content Interactivity Research question 4 related to human-to-content interactivity. Only 21% of sites had any features that allowed for the addition of content. Eleven of those sites had a single feature (most often a bulletin board or chat room). Thus, it is not surprising that the average number of features for adding content was only .39 and that overall correlation between adding features and total interactivity was modest (r = .216; p < .01). About a third (30%) of sites had some customization capability. The number of customization features ranged from 1–5 with the most common features being ability to change a language and to reformat a page to optimize for printing. Sites that allow customization are also moderately correlated with high levels of overall interactivity (r = .251; p < .01). When both types of human-to-content interactivity are considered in combination, there is a stronger correlation between human-to-content interactivity and overall interactivity (r = .338; p < .001). This suggests that even though relatively few sites are using human-to-content interactive functions, those that do add those features are among the most interactive sites. And it may also be possible that as interactivity evolves these “less-common” features will become even more important in differentiating levels of interaction with visitors. Interactivity by Level The breakdown of features by level 1 and 2 is central to research question 5. On average, fewer features were found at level 1 (9.66) than level 2 (11.59). That difference was significant (t = 19.25, df = 99, p <.001). There was also a positive correlation between levels 1 and 2 (r = .597; p < .001). This suggests that the interactivity found at level 1 is a fairly good predictor of level 2 interactivity. But, this study also points to the importance of going beyond level 1 in examining interactivity. The correlation between level 2 and total interactivity was stronger (r = .934; p < .001) than correlation between level 1 and total interactivity (r = .845; p <.001). This suggests that important insights about interactivity at Web sites might be lost if researchers examine only level 1. Interactivity by Domain Type This study purposively selected health-related Web sites from among five different types of internet domains. As detailed in research question 6, the researchers wanted to determine if there were any significant differences in interactivity among these domains. Table 5 shows that there are significant differences among sites in terms of human-to-computer interactivity features. Posthoc analysis (using Tukey HSD) shows that there is no significant difference between .gov and .com sites (t = .160; p = .05), which are the richest in human-to-computer interactive features. Similarly .net and .org sites are not significantly different (t = .014; p = .05) from each other as both are relatively low in human-to-computer features. Sites from the .edu domain have midlevel human-to-computer features. Educational sites were not significantly different from .net (t = .53; p = .05) or .org (t = .591; p = .05) sites. All relationships other than those noted above were significantly different at the p < .05 level. The greatest source of difference is between .gov domains and .org domains (t = 2.28; p < .05). Table 5 Mean Interactivity Scores by Domain Type and Interactivity Type . Human-to-Computer . Human-to-Human . Human-to-Content . Total . .gov (N = 20) 20.90 3.95 1.10 26.05 .com (N = 20) 20.45 4.50 1.05 26.10 .edu (N = 20) 15.35 3.25 .30 18.95 .net (N = 20) 13.70 3.10 1.00 17.95 .org (N = 20) 13.65 2.85 .70 17.20 Total (N = 100) 16.81 3.53 .83 21.25 F 2.83* 1.95 1.58 3.44** . Human-to-Computer . Human-to-Human . Human-to-Content . Total . .gov (N = 20) 20.90 3.95 1.10 26.05 .com (N = 20) 20.45 4.50 1.05 26.10 .edu (N = 20) 15.35 3.25 .30 18.95 .net (N = 20) 13.70 3.10 1.00 17.95 .org (N = 20) 13.65 2.85 .70 17.20 Total (N = 100) 16.81 3.53 .83 21.25 F 2.83* 1.95 1.58 3.44** * p < .05; ** p < .01. Open in new tab Table 5 Mean Interactivity Scores by Domain Type and Interactivity Type . Human-to-Computer . Human-to-Human . Human-to-Content . Total . .gov (N = 20) 20.90 3.95 1.10 26.05 .com (N = 20) 20.45 4.50 1.05 26.10 .edu (N = 20) 15.35 3.25 .30 18.95 .net (N = 20) 13.70 3.10 1.00 17.95 .org (N = 20) 13.65 2.85 .70 17.20 Total (N = 100) 16.81 3.53 .83 21.25 F 2.83* 1.95 1.58 3.44** . Human-to-Computer . Human-to-Human . Human-to-Content . Total . .gov (N = 20) 20.90 3.95 1.10 26.05 .com (N = 20) 20.45 4.50 1.05 26.10 .edu (N = 20) 15.35 3.25 .30 18.95 .net (N = 20) 13.70 3.10 1.00 17.95 .org (N = 20) 13.65 2.85 .70 17.20 Total (N = 100) 16.81 3.53 .83 21.25 F 2.83* 1.95 1.58 3.44** * p < .05; ** p < .01. Open in new tab The general distribution of features is similar for human-to-human features, but no overall significant differences were found. Slightly more human-to-human features were found in .com sites than .gov sites, but a comparison of means for those two types of sites did not reveal a significant difference (t = .648; p = .05). The only significant differences found in posthoc analysis of human-to-human features was between .com and .org sites (t = 2.04; p < .01) and .gov and .org sites (t = 2.28; p < .05). This finding highlights the relatively low level of human-to-human interaction that is facilitated at these largely nonprofit sites. No overall significant difference was found among domain types for human-to-content interactivity. This is probably, at least in part, because of the low overall implementation of this type of interactivity. Again .gov and .com sites are most likely to use this type of interactivity as they did with other types. But interestingly, .net sites often seemed to approach the level of human-to-computer interactivity found at the “higher-end” sites. Posthoc analysis of this type of interactivity found the only significant differences to be between .gov and .edu sites (t = 2.53; p < 05) and .com and .edu sites (t = 2.62; p < .01). Table 5 shows that there is an overall difference in total interactivity based on site type. Posthoc analysis shows that the general patterns found in each of the types of interactivity are consistent with overall findings. The .gov sites were significantly more interactive than were .org (t = .264; p < .01) .net (t = 2.16; p < .05), and .edu (t = 2.39; p < .05) sites. Similarly, the .com sites were significantly more interactive than were .org (t = 2.71; p < .01), .net (t = 2.21; p < .05) and .edu (t = 2.37; p < .05) sites. No other significant differences were found in post hoc analysis of total interactivity differences based on domain types. Discussion The goal of this study was to develop and apply an assessment tool for interactivity. Health related Web sites provided an appropriate context. These findings offer insights for researchers and practitioners. Considerations for Interactivity Assessment One of the most promising outcomes is that coders were able to reliably use a complex, but highly flexible, tool to assess interactive features at Web sites. The coding tool was relevant to subtopics and domains that ranged from very simple sites created by nonprofit organizations to complex government sites and sophisticated commercial sites. However, the study also suggested possible needs for further refinement of the interactivity assessment tool. The coding scheme may need to differentiate between instances when additional features add interactivity and when they are simply increasing redundancy at sites. For example, a menu could appear at the top of the page, buttons on the side, and text links at the bottom of the page. The current coding form would report three standard navigation tools. In some cases, these tools would lead to different types of content that are organized logically to correspond to the different navigational elements. But in other cases, all of these formats link to the same content. In these cases, additional features don’t add functionality and may even frustrate the user who is looking for new content. Thus while a Web site could score high on interactivity, typically viewed as a positive characteristic, because of the level of redundancy a site visitor could experience frustration as s/he explored the site yet found limited new content. Future studies may seek ways to fine-tune the interactivity assessment tool to address this concern. The findings also suggest researchers need to carefully consider the goals of a study before applying the coding scheme to their work. For example, if a study is focused on human-to-human and/or human-to-content interactions, some sort of “weighting” scheme may need to be used to keep those features from being “masked” by more dominant navigation features. As noted earlier, it will also be critical to train coders for each research project so that they can appropriately adapt the coding scheme not only to the specific type of content being coded but also to new features that are continually emerging. An important part of the coder training will be to carefully delineate how the researchers decide which category should be used for coding a specific feature (e.g., should e-mail links only be coded as human-to-human interaction if they actually produce an individualized response). These decisions should be guided by the goals of the specific study. The context of the study will also be important to researchers. Health Web sites provided a venue that allowed for selection of multiple domains and a variety of topics, but findings may be somewhat idiosyncratic. For example, health topics can often be “life or death” issues. Thus, Web site creators may hesitate to provide synchronous communication unless they can fully staff it. If site visitors were lead to believe that “online chat” could substitute for a 911 emergency call, the results could be catastrophic if those expectations could not be met. Future research should also build and test on some very basic assumptions made in this study. Correlations between interactivity at a given level and overall interactivity is only a very rudimentary way to determine relationships among features at any part of the Web site and overall interactivity. Future studies should also engage consumers in analysis of perceived interactivity to determine relationships between specific features and overall interactivity. Considerations for Practitioners Preliminary findings revealed several potential areas of improvement for Web site creators. These areas for improvement cut across all three forms of interactivity. With respect to human-to-computer interactivity, this study found significantly fewer personalized compared to standardized features. Unlike other mass media, the internet offers the technological opportunity to create an individualized experience that could further enhance the sense of a personal relationship with the content creator. One somewhat unexpected finding was that action functions (unlike navigation and transaction functions) are more likely to be personalized than standardized. This may suggest that the easiest area to enhance personalization on Web sites such as those examined in this study would be increase action features. For example, visitors could engage with the sites more fully through quizzes, surveys, games, and other activities that require minimal programming input but allow the individual more opportunities to interact with the site. Both forms of human-to-human interactivity showed areas that could be improved. Although nearly all sites allowed for asynchronous interaction between the site visitor and the organization (e.g., e-mail us or contact form), “real time” or synchronous communication was virtually absent. Research has found that 42% of internet users engage in instant messaging to some degree (Shie & Lenhart, 2004); Web site creators may need to “catch up” in terms of expected tools for human-to-human interaction. Only a few sites provided opportunities for individual to individual interaction, or online word-of-mouth. Macklin reports that the tactics most used by more experienced Web site creators include encouraging e-mail forwarding (91%), “tell a friend” options (80%), and offering e-cards (47%) (Macklin, 2006). Such features are technically simple to implement and offer quick impact at a relatively low cost. Opportunities also exist for enhancing human-to-content interactivity. This form of interactivity was least used, but may be most innovative. By allowing site visitors to customize and add content, the Web site becomes more personal and thus may be more likely to become central to the individual’s online life (Strauss & Frost, 1999). Suggestions for Future Research This study developed and tested a tool that has potential to be highly valuable to both researchers and practitioners. Interactivity is an important characteristic of Web sites and an assessment tool is needed that is both comprehensive and sensitive to the evolution of the Web. The tool used for this study seems flexible enough to be applied regardless of site content or technological developments. From here, future research could develop along two paths. The first would be to further examine health-related Web sites with an expanded sample to verify this study’s preliminary findings and offer more specific recommendations to enhance interactivity. Additionally, there is the need to engage in a large-scale content analysis that would include a variety of content areas (rather than health Web sites only) to further gauge generalizability of the assessment tool. A related direction would be comparison of Web sites by target audience especially comparing those sites targeting younger adults, who are “born digital” and may have high expectations for interactivity, and older adults for whom the concept of computer-based interactivity remains somewhat preplexing. The second path to pursue is examination of site visitor responses to various forms of interactivity. What specific features do they expect to find? What features enhance evaluation of the site? When do visitors want redundancy? When might it be a distraction or irritation? This line of inquiry could also explore questions such as how much time and effort individuals wish to invest in internet activities. Is there such a thing as “too much” interactivity? If so, should site creators avoid the types of interactivity that require a large investment of time/effort – most notably contribution to the content of Web sites? Interactivity is a complex phenomenon. Not only is it multifaceted involving humans’ interaction with the computer, each other, and content, but it is also constantly changing. The hyperlink that seemed revolutionary as a tool for interactivity in the early 1990s is an expected feature of the contemporary Web site. The multifaceted tool presented here should be flexible enough to be applicable to any current Web site and adaptable enough to assess new interactive features as they evolve. Thus, it should have a relatively long life as a tool for the tough job of measuring online interactivity. References Alter , S . ( 1977 ). Why Is man-computer interaction important for decision support systems? Interfaces , 7 ( 2 ), 109 – 115 . Google Scholar Crossref Search ADS WorldCat Ambre , J. , Guard , R., Perveiler , F. M., Renner , J., & Rippen , H. ( 1997 ). Criteria for assessing the quality of health information on the Internet . Available online: http://www.mitretek.org/hiti/showcase/documents/criteria.html. Health Information Technology Institute of Mitretek Systems, Inc Armstrong , A. , & Hagel , J. I. ( 1996 ). The real value of on-line communities . Harvard Business Review , 74 ( 3 ), 134 – 140 . Google Scholar OpenURL Placeholder Text WorldCat Barker , J. , & Tucker , R. N. ( 1990 ). The interactive learning revolution: multimedia in education and Training . London : Kogan Page . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bedell , S. E. , Agrawal , A., & Petersen , L. E. ( 2004 ). A systematic critique of diabetes on the World Wide Web for patients and their physicians . International Journal of Medical Informatics , 73 , 687 – 694 . Google Scholar Crossref Search ADS PubMed WorldCat Berelson , B. , & Lazarsfeld , P. F. ( 1948 ). The analysis of communication content . Chicago and New York : University of Chicago and Columbia University . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Berland , G. K. , Elliott , M. N., Morales , L. S., & Algazy , J. I. ( 2001 ). Health information on the Internet: Accessibility, quality, and readability in English and Spanish . JAMA: The Journal of the American Medical Association , 285 ( 20 ), 2612 – 2621 . Google Scholar Crossref Search ADS WorldCat Blattberg , R. C. , & Deighton , J. ( 1991 ). Interactive marketing: Exploiting the age of addressability . Sloan Management Review , 32 ( 1 ), 5 – 14 . Google Scholar OpenURL Placeholder Text WorldCat Bryant , J. , & Street Jr., R. L. ( 1988 ). From reactivity to activity to action: An evolving concept and weltanschauung in mass and interpersonal communication . In R. P. Hawkins & J. M. Wiemann & S. Pingree (Eds.), Advancing Communication Science: Merging Mass and Interpersonal Process (pp. 162 – 190 ). Newbury Park, CA : Sage . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bucy , E. P . ( 2004 ). Second generation net news: Interactivity and information accessibility in the online environment . The International Journal on Media Management , 6 ( 1/2 ), 102 – 113 . Google Scholar Crossref Search ADS WorldCat Bucy , E. P. , & Chen-Chao , T. ( 2007 ). The mediated moderation model of interactivity . Media Psychology , 9 ( 3 ), 647. Google Scholar OpenURL Placeholder Text WorldCat Burke , K. , Aytes , K., Chidambaram , L., & Johnson , J. J. ( 1999 ). A study of partially distributed work groups: The impact of media, location, and time on perceptions and performance . Small Group Research , 30 ( 4 ), 453 – 490 . Google Scholar Crossref Search ADS WorldCat Chang , T.-K. , & Tai , Z. ( 2005 ). Mass communication research and the invisible college revisited: The changing landscape and emerging fronts in journalism-related studies . Journalism & Mass Communication Quarterly , 82 ( 3 ), 672 – 694 . Google Scholar Crossref Search ADS WorldCat Chilcoat , Y. , & DeWine , S. ( 1985 ). Teleconferencing and interpersonal communication perception . Journal of Applied Communication Research , 13 ( 1 ), 14 – 32 . Google Scholar Crossref Search ADS WorldCat Cho , C.-H. , & Cheon , H. J. ( 2005 ). Cross-cultural comparisons of interactivity on corporate Web sites . Journal of Advertising , 34 ( 2 ), 99 – 115 . Google Scholar Crossref Search ADS WorldCat Cho , C.-H. , & Khang , H. ( 2006 ). The state of Internet-related research in communications, marketing, and advertising: 1994–2003 . Journal of Advertising , 35 ( 3 ), 143 – 163 . Google Scholar Crossref Search ADS WorldCat Coyle , J. R. , & Thorson , E. ( 2001 ). The Effects of Progressive Levels of Interactivity and Vividness in Web Marketing Sites . Journal of Advertising , 30 ( 3 ), 65 – 77 . Google Scholar Crossref Search ADS WorldCat Curro , V. , Buonuomo , P. S., Onesimo , R., DeRose , P., Vituzzi , A., DiTanna , G. L., & D’Atri , A. ( 2004 ). A quality evaluation methodology of health Web-pages for non-professionals . Medical Informatics & the Internet in Medicine , 29 ( 2 ), 95 – 107 . Google Scholar Crossref Search ADS WorldCat De Marsico , M. , & Levialdi , S. ( 2004 ). Evaluating Web sites: Exploiting user’s expectations . International Journal of Human-Computer Studies , 60 , 381 – 416 . Google Scholar Crossref Search ADS WorldCat Gerbic , P. , & Stacey , E. ( 2006 ). A purposive approach to content analysis: Designing analytical frameworks . Internet and Higher Education , 8 , 45 – 59 . Google Scholar Crossref Search ADS WorldCat Gonzalez , F. J. M. , & Palacios , T. M. B. ( 2004 ). Quantitative evaluation of commercial Web sites: An empirical study of Spanish firms . International Journal of Information Management , 24 , 313 – 328 . Google Scholar Crossref Search ADS WorldCat Ha , L. , & James , L. ( 1998 ). Interactivity reexamined: A baseline analysis of early business Web sites . Journal of Broadcasting & Electronic Media , 42 ( 4 ), 457 – 474 . Google Scholar Crossref Search ADS WorldCat Haeckel , S. H . ( 1998 ). About the nature and future of interactive marketing . Journal of Interactive Marketing , 12 ( 1 ), 63 – 71 . Google Scholar Crossref Search ADS WorldCat Hanssen , L. , Jankowski , N. W., & Etienne , R. ( 1996 ). Interactivity from the perspective of communication studies . In N. W. Jankowski & L. Hanssen (Eds.), Contours of Multimedia: Recent Technological, Theoretical, and Empirical Developments (pp. 61 – 73 ). Luton, UK : University of Luton Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hester , J. B . ( 1999 ). Using a Web-based interactive test as a learning tool . Journalism and Mass Communication Educator , 54 ( 1 ), 35 – 41 . Google Scholar Crossref Search ADS WorldCat Holsti , O . ( 1969 ). Content analysis for the social sciences and humanities . Reading, MA : Addison-Wesley . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hwang , J.-S. , McMillan , S. J., & Lee , G. ( 2002 ). Corporate web sites as advertising: An analysis of message strategies and features . In A. M Abernathy (Ed.), Proceedings of the American Academy of Advertising (pp. 56 ). Auburn, AL : Auburn University . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Hwang , J.-S. , McMillan , S. J., & Lee , G.. ( 2003 ). Corporate Web sites as advertising: An analysis of function, audience, and message strategy . Journal of Interactive Advertising , 3 ( 2 ), Available http://jiad.org Jankowski , N. W. , Foot , K., Kluver , R., & Schneider , S. ( 2005 ). The Web and the 2004 EU election: Comparing political actor web sites in 11 EU member states . Information Polity , 10 , 165 – 176 . Google Scholar OpenURL Placeholder Text WorldCat Jensen , J. F . ( 1998 ). Interactivity: Tracing a new concept in media and communication studies . Nordicom Review , 19 ( 1 ), 185 – 204 . Google Scholar OpenURL Placeholder Text WorldCat Keltner , K . ( 1998 ). Networked health information: Assuring quality control on the Internet . Federal Communications Law Journal , 50 ( 2 ), 417 – 439 . Google Scholar OpenURL Placeholder Text WorldCat Kim , J. , & McMillan , S. J.. (Forthcoming). Evaluation of Internet advertising research: A bibliometric analysis of citations from key sources . Journal of Advertising . OpenURL Placeholder Text WorldCat Levins , I . ( 1999 ). Relationship marketing gets more personal – and it works . Healthcare PR & Marketing News , 8 ( 25 ), 1 . Google Scholar OpenURL Placeholder Text WorldCat Liu , Y . ( 2003 ). Developing a scale to measure the interactivity of websites . Journal of Advertising Research , 43 ( 2 ), 207 – 216 . Google Scholar Crossref Search ADS WorldCat Liu , Y. , & Shrum , L. J. ( 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 . Google Scholar Crossref Search ADS WorldCat Lynn , G. S. , Lipp , S. M., Akgun , A. E., & Crotez Jr., A. ( 2002 ). Factors impacting the adoption and effectiveness of the World Wide Web in marketing . Industrial Marketing Management , 31 , 35 – 49 . Google Scholar Crossref Search ADS WorldCat Macias , W. , & McMillan , S. J. ( 2008 ). The return of the house call: The role of internet-based interactivity in bringing health information home to older adults . Health Communication , 23 ( 1 ), 1 – 11 . Google Scholar Crossref Search ADS PubMed WorldCat Macklin , B . ( 2006, September 28 ). Viral …and stealthy . emarketer.com. Marvin , C . ( 1983 ). Space, time, and captive communications history . In M. S. Mander (Ed.), Communications in Transition (pp. 20 – 38 ). New York : Prager . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Massey , B. L. , & Levy , M. R. ( 1999 ). Interactivity, online journalism, and English-language Web newspapers in Asia . Journalism & Mass Communication Quarterly , 76 ( 1 ), 138 – 151 . Google Scholar Crossref Search ADS WorldCat McGrath , J. E . ( 1991 ). Time, interaction, and performance (TIP): A Theory of Groups . Small Group Research , 22 ( 2 ), 147 – 174 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J . ( 1998 ). Who pays for content? Funding in interactive media . Journal of Computer Mediated Communication , 4 ( 1 ), Available http://www.ascusc.org/jcmc/vol4/issue1/mcmillan.html McMillan , S. J . ( 1999 ). Health communication and the Internet: Relationships between interactive characteristics of the medium and site creators, content, and purpose . Health Communication , 11 ( 4 ), 375 – 390 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J . ( 2000a ). Interactivity is in the eye of the beholder: Function, perception, involvement, and attitude toward the Web site . In M. A Shaver (Ed.), Proceedings of the American Academy of Advertising (pp. 71 – 78 ). East Lansing, MI : Michigan State University . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McMillan , S. J . ( 2000b ). The microscope and the moving target: The challenge of applying content analysis to the World Wide Web . Journalism and Mass Communication Quarterly , 77 ( 1 ), 80 – 98 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J . ( 2002 ). A four-part model of cyber-interactivity: Some Cyber-Places are More Interactive Than Others . New Media and Society , 4 ( 2 ), 271 – 291 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J . ( 2005 ). The researchers and the concept: Moving beyond a blind examination of interactivity . Journal of Interactive Advertising , 5 ( 2 ), Available online: http://jiad.org McMillan , S. J. , & Downes , E. J. ( 2000 ). Defining interactivity: A qualitative identification of key dimensions . New Media and Society , 2 ( 2 ), 157 – 179 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J. , & Hwang , J.-S. ( 2002 ). Measures of perceived interactivity: An exploration of communication, user control, and time in shaping perceptions of interactivity . Journal of Advertising , 31 ( 3 ), 41 – 54 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J. , Hwang , J.-S., & Lee , G. ( 2003 ). Effects of structural and perceptual factors on attitudes toward the website . Journal of Advertising Research , 43 ( 4 ), 400 – 409 . Google Scholar Crossref Search ADS WorldCat McMillan , S. J. , Kim , J., & Hwang , J.-S. ( 2004 ). Searching for the riddle in the Web site: Coding interactivity in online content analysis . In P. B. Rose (Ed.), Proceedings of the American Academy of Advertising (pp. 255 – 262 ). Miami, FL : University of South Florida . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McMillan , S. J. , Kim , J., McMahan , C., & Fall , L. T. ( 2006 ). Analyze this: Developing a coding scheme for measuring interactivity features in the context of state tourism Web sites . In J. Richards (Ed.), Proceedings of the American Academy of Advertising (pp. 222 – 231 ). Austin, TX : University of Texas . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Miles , I . ( 1992 ). When mediation is the message: How suppliers envisage new markets . In M. Lea (Ed.), Contexts of Computer-Mediated Communication (pp. 145 – 167 ). New York : Harvester-Wheatsheaf . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Moon , Y. , & Nass , C. ( 1996 ). How “real” are computer personalities? Psychological responses to personality types in human-computer interaction . Communication Research , 23 ( 6 ), 651 – 674 . Google Scholar Crossref Search ADS WorldCat Morahan-Martin , J. M . ( 2004 ). How Internet users find, evaluate, and use online health information: A cross-cultural review . Cyberpsychology & Behavior , 7 ( 5 ), 497 – 510 . Google Scholar Crossref Search ADS WorldCat Nielsen , J. , & Loranger , H. ( 2006 ). Prioritizing Web usability . Berkley, CA : New Riders . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Okazaki , S . ( 2005 ). Searching the Web for global brands: How American brands standardise their Web sites in Europe . European Journal of Marketing , 39 ( 1/2 ), 87 – 109 . Google Scholar Crossref Search ADS WorldCat Okazaki , S. , & Rivas , J. A. ( 2002 ). A content analysis of multinationals’ Web communication strategies: Cross-cultural research framework and pre-testing . Internet Research: Electronic Networking Applications and Policy , 12 ( 5 ), 380 – 390 . Google Scholar Crossref Search ADS WorldCat Paul , M. J . ( 2001 ). Interactive disaster communication on the Internet: A content analysis of sixty-four disaster relief home pages . Journalism & Mass Communication Quarterly , 78 ( 4 ), 739 – 753 . Google Scholar Crossref Search ADS WorldCat Pavlou , P. A. , & Stewart , D. W. ( 2000 ). Measuring the effects and effectiveness of interactive advertising: A research agenda . Journal of Interactive Advertising , 1 ( 1 ), Available at: http://jiad.org/vol1/no1/pavlou Phillips , B. , & Lee , W.-N. ( 2005 ). Interactive animation: Exploring spokes-characters on the Internet . Journal of Current Issues and Research in Advertising , 27 ( 1 ), 1 – 17 . Google Scholar Crossref Search ADS WorldCat Rafaeli , S . ( 1988 ). Interactivity: From new media to communication . In R. P. Hawkins & J. M. Wiemann & S. Pingree (Eds.), Advancing communication science: Merging mass and interpersonal process (pp. 110 – 134 ). Newbury Park, CA : Sage . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rafaeli , S. , & Sudweeks , F. ( 1997 ). Networked interactivity . Journal of Computer Mediated Communication , 2 ( 4 ), Available: http://www.usc.edu/dept/annenberg/vol2/issue4/rafaeli.sudweeks.html Reardon , K. K. , & Rogers , E. M. ( 1988 ). Interpersonal versus mass media communication: A false dichotomy . Human Communication Research , 15 ( 2 ), 284 – 303 . Google Scholar Crossref Search ADS WorldCat Riffe , D. , & Freitag , A. ( 1997 ). A content analysis of content analyses: Twenty-five years of journalism quarterly . Journalism & Mass Communication Quarterly , 74 ( 4 ), 873 – 882 . Google Scholar Crossref Search ADS WorldCat Riffe , D. , Lacy , S., & Fico , F. ( 1998 ). Analyzing media messages: Using quantitative content analysis in research . Mahwah, NJ : Lawrence Erlbaum Associates . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Risk , A. , & Petersens , C. ( 2002 ). Health information on the Internet: Quality issues and international initiatives . JAMA: The Journal of the American Medical Association , 287 ( 20 ), 2713 – 2715 . Google Scholar Crossref Search ADS WorldCat Ritterband , L. M. , Cox , D. J., Gordon , T. L., Borowitz , S. M., Kovatchev , B. P., Walker , L. S., & Sutphen , J. L. ( 2006 ). Examining the added value of audio, graphics, and interactivity in an Internet intervention for pediatric encopresis . Children’s Health Care , 35 ( 1 ), 47 – 59 . Google Scholar Crossref Search ADS WorldCat Ruben , B. D . ( 1975 ). Intrapersonal, interpersonal, and mass communication processes in individual and multi-person systems . In B. D. Ruben & J. Y. Kim (Eds.), General Systems Theory and Human Communication (pp. 164 – 190 ). Rochelle Park, NJ : Hayden Book Co . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Schultz , T . ( 1999 ). Interactive options in online journalism: A content analysis of 100 U.S. newspapers . Journal of Computer Mediated Communication , 5 ( 1 ), Available: http://www.ascusc.org/jcmc/vol5/issue1/schultz.html Shie , E. , & Lenhart , A. ( 2004, 1 September ). How Americans use instant messaging . Pew Internet and American Life Project . Available: http://www.pewinternet.org/pdfs/PIP_Instantmessage_Report.pdf [2006, 7 September] Straubhaar , J. , & LaRose , R. ( 1996 ). Communications media in the information society . Belmont, CA : Wadsworth Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Strauss , J. , & Frost , R. ( 1999 ). Marketing on the Internet: Principles of online marketing . Upper Saddle River, NJ : Prentice Hall . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Sundar , S. S. , & Kim , J. ( 2005 ). Interactivity and persuasion: Influencing attitudes with information and involvement . Journal of Interactive Advertising , 5 ( 2 ), N.PAG. Google Scholar OpenURL Placeholder Text WorldCat Tomasello , T. K . ( 2001 ). The status of Internet-based research in five leading communication journals, 1994-1999 . Journalism & Mass Communication Quarterly , 78 ( 4 ), 659 – 674 . Google Scholar Crossref Search ADS WorldCat Trammell , K. D. , Williams , A. P., Postelnicu , M., & Landreville , K. D. ( 2006 ). Evolution of online campaigning: Increasing interactivity in candidate Web sites and blogs through text and technical features . Mass Communication & Society , 9 ( 1 ), 21 – 44 . Google Scholar Crossref Search ADS WorldCat Walther , J. B . ( 1996 ). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction . Communication Research , 23 ( 1 ), 3 – 43 . Google Scholar Crossref Search ADS WorldCat Walther , J. B. , Pingree , S., Hawkins , R. P., & Buller , D. B. ( 2005 ). Attributes of interactive online health information systems . Journal of Medical Internet Research , 7 ( 3 ), 21 – 33 . Google Scholar Crossref Search ADS WorldCat Weare , C. , & Lin , W.-Y. ( 2000 ). Content analysis of the World Wide Web: Opportunities and challenges . Social Science Computer Review , 18 ( 3 ), 272 – 292 . Google Scholar Crossref Search ADS WorldCat Wells , W. D. , & Chen , Q. ( 2000 ). The dimensions of commercial cyberspace . Journal of Interactive Advertising , 1 ( 1 ), Available online: http://jiad.org/vol1/no1/wells/index.html Wise , K. , Hamman , B., & Thorson , K. ( 2006 ). Moderation, response rate, and message interactivity: Features of online communities and their effects on intent to participate . Journal of Computer-Mediated Communication , 12 ( 1 ), Available at: http://jcmc.indiana.edu/vol12/issue11/wise.html Xie , H . ( 2000 ). Shifts of interactive intentions and information-seeking strategies in interactive information retrieval . Journal of the American Society for Information Science , 51 ( 9 ), 841 – 857 . Google Scholar Crossref Search ADS WorldCat Yun , G. W . ( 2007 ). Interactivity concepts examined: Response time, hypertext, role taking, and multimodality . Media Psychology , 9 ( 3 ), 527 – 548 . Google Scholar Crossref Search ADS WorldCat Zeltzer , D . ( 1992 ). Autonomy, interaction, and presence . Teleoperators and Virtual Environments , 1 ( 1 ), 127 – 132 . Google Scholar Crossref Search ADS WorldCat About the Authors Sally J. McMillan is an Associate Professor in Advertising and Public Relations at the University of Tennessee. Her research focuses on exploring interactivity, definitions, and history of new media, online research methods, health communication, and impacts of communication technology on organizations and society. Address: 476 Communications Building, University of Tennessee, Knoxville, TN 37996. Mariea G. Hoy is a Professor in Advertising and Public Relations at the University of Tennessee. Her research focuses on issues such as consumer privacy, disclosures in advertising messages, and advertising to children. Address: 476 Communications Building, University of Tennessee, Knoxville, TN 37996. Juran Kim is an Assistant Professor in Advertising and Public Relations, School of Business Administration, Jeonju University. Dr. Kim’s research focuses on interactivity and consumer behavior in the interactive marketing communication context. Address: Jeonju University, School of Business Administration, Advertising and Public Relations, 518 Research Bldg., 1200 Hyojadong, Wansangu, Jeonju, Jeollabukdo 560-759, South Korea. Carolyn McMahan is an Assistant Professor of Advertising at the University of North Florida. Her research focuses on gender portrayals in advertising and digital media. Address: 567 St. John’s Bluff Road South Jacksonville, FL 32224. Appendix 1 – List of Sites coded http://Planet-Therapy.com http://www.4therapy.com http://www.aboutpsychotherapy.com http://www.adolescent-substance-abuse.com http://www.ahd.com http://www.allaboutcounseling.com http://www.answersforteens.com http://www.bhworld.com http://www.bubblemonkey.com http://www.cha.com http://www.checkyourself.com http://www.counselingnetwork.com http://www.freevibe.com http://www.chem-tox.com http://www.hospitalsoup.com http://www.robertperkinson.com http://www.soberrecovery.com http://www.WholeHealthMD.com http://www.wvha.com http://www.drgreene.com http://www.aaap.org http://www.antipsychiatry.org http://www.awarefoundation.org http://www.banshock.org http://www.bhj.org http://www.copecaredeal.org http://www.cybershrink.org http://www.delamed.org http://www.inhalant.org http://www.deliciousdecisions.org/ http://www.MindFreedom.org http://www.nostigma.org http://www.planetreesanjose.org http://www.purificationpgm.org http://www.recoveryresources.org http://www.soilandhealth.org http://www.teenshealth.org http://www.victimsofcedars.org http://www.wildernesstherapy.org http://www.womenfdn.org http://rarediseases.info.nih.gov http://www.ahrq.gov/ http://www.nida.nih.gov/ http://www.4girls.gov http://www.4woman.gov http://www.bt.cdc.gov http://www.healthfinder.gov http://www.lbc.nimh.nih.gov http://www.nal.usda.gov http://www.nccam.nih.gov http://www.nia.nih.gov http://www.niaaa.nih.gov http://www.niaid.nih.gov http://www.nida.nih.gov http://www.nimh.nih.gov http://www.ninds.nih.gov http://www.os.dhhs.gov http://www.recalls.gov http://www.teens.drugabuse.gov http://www.thecoolspot.gov http://www.hopelinks.net http://www.alzonline.net http://anxiety.mentalhelp.net http://www.newideas.net http://www.abstinence.net http://www.respectcampaign.net http://www.contraception.net http://www.teensforteens.net http://www.familyhospital.net http://www.angryparents.net http://www.renew.net http://www.bereavement.net http://www.aliveness.net http://www.marriages.net http://www.integrativepsychiatry.net http://www.freepsychotherapy.net http://www.ChildrenofHopeFamilyHospital.dmi.net http://www.soberforever.net http://www.buildingbridgesinc.net http://www.altmed.net http://arcc.stanford.edu http://ec.princeton.edu/ http://med.stanford.edu/ http://orthogenicschool.uchicago.edu http://utsurg.uth.tmc.edu/ http://www.adolescentpregnancy.unc.edu http://www.allencollege.edu http://www.alz.washington.edu http://www.chw.edu http://www.cmps.edu http://www.fmhi.usf.edu http://www.hypnosis.edu http://www.kumc.edu http://www.mclean.harvard.edu http://www.nursing.upenn.edu http://www.psy.edu www.slhs.sdsu.edu http://www.swc.edu http://healthletter.tufts.edu http://www.yourdiseaserisk.harvard.edu Appendix 2 – Coding Form Open in new tab Open in new tab Appendix 3 – Coding Instructions Coding Overview Enter the URL for the site you are coding and your name on the top of the coding form. Code for levels 1 and 2 for each web site and use the definitions of types of interactivity provided later in this document to guide your coding. For all items coded provide a description for the features (e.g. top menu bar is enough, you don’t need to describe what is on the menu bar) as well as providing a numeric summary of how many features you found. Try to code each Web page from top left to bottom right. Code the level 1 page first. This is generally fairly straightforward as you don’t have to worry at all about what to do about multiple “instances” of a feature. You should click on each menu item, hyperlink, button, etc. found on page 1. The results of these “clicks” are all level-2 items and should be coded for interactive features. The key thing to keep in mind here is that you are only counting things that appear NEW at level 2. For example, if the level two pages have all the navigation items of page 1 but also add a new side menu that did not appear on page 1, the only thing you code for navigation at level 2 is the NEW side menu. And you would only code it ONCE, even if it appears on every level 2 page. Generally, as you code level 2, you are looking for NEW elements. For example, if there is a search function that appears on level 1 and all subsequent levels, you would code it as level 1 but not code it again for level 2. But, if there is NOT a search function at level 1, but a search function DOES appear on one or more level 2 pages, then you would code it for level 2. If the same search function appears on several level 2 pages, code it only once; if there are different kinds of searches (e.g. search doctors, search diseases, etc.), code each separate TYPE of search that you find. If you code anything as “other” provide a description for it. If a link takes you to an external site, do not code the features of that external site for features, but note if the link enables a fundamental type of interaction (e.g. download PDF). If the site uses “frames” it is sometimes difficult to tell if you have actually moved “off site.” If the site has a substantially different look and feel within the “frame” version, assume that you have moved off site. It is not necessary to actually conduct searching (using the search tool) and report on the results of those searches. Simply note that the search tool exists. It is also not necessary to try to register and/or login for the site. Simply record what you can without login. It is most important to code the feature and the right level (1 or 2) and within the right general category (e.g. navigation/standard, navigation/personalized, etc.) than to be sure that you have identified exactly the right “subcategory” (e.g. menu/submenu, text link, etc.). The subcategories should be used primarily to help you identify types of features within a category. General Rules Usually, items are coded for only one type of interactivity. The exception is items that are both a navigational tool and something else. For example, a menu bar would be coded as a navigational tool, but a “contact us” item on a menu bar would also be coded as human-to-human interaction. If text, graphics, etc. at a given level indicate that interactivity will occur at a subsequent level, code the element at the first level at which it is found. For example, if there is a “contact us” link on level 1 and clicking on that link opens a level 2 contact form, then organization/individual asynchronous interactivity would be coded for both levels 1 and 2. Rule 2 only applies if the initial element clearly suggests interactivity will occur on a subsequent page. For example, if “cool stuff” on the front page links to a screen for ordering “cool stuff” at level 2, do NOT code interactivity at level 1 (cool stuff is not clearly interactive), but DO code interactivity at level 2. Generally, base coding on what is manifest at a given level, not what you decide the interactivity “really” is after you click through.” However, use good judgment. Like items on a page that seem to be a “set” should be coded only once. For example, multiple items in a menu bar count as one interactive feature. However, if items are substantially different, code once for each “type.” For example, if there are multiple navigational systems (e.g. side bars, top bars, hyperlinks) code each different type. Similarly, if there were three “lookalike” buttons that lead to similar functions (e.g. reading about a specific disease) you would could “1” for buttons. But if there was a substantially different graphic element (e.g. a banner ad) that would be a different “type” of graphic and you would now code “2” for the two types of graphic links. In general, the rule is to code “like items” as a single block, but to code different instances of those items separately. Here are some examples of how this rule is manifest in the different “types” of interactivity: For navigation, count the total number of menu bars on a page (if submenus are visible either directly or with a mouse rollover, these are additional menu bars). But you don’t count the total number of “items” on each menu bar. Note portions of the page where hyperlinks appear, but don’t count every link (e.g. body copy, side bar story, etc.). This principle holds for all kinds of navigational tools. For action, count the total number of different types of surveys, games, registration options, etc. But count each type only once – even if they appear on multiple pages. For transaction, count the total number of opportunities for ordering materials. For example, if there is an option for ordering brochures, you would count that as one transaction. If there are multiple brochures available, do NOT count every brochure. For the various types of human-to-human communication, count different “types” of options, but not every single opportunity. For example, if you have an option to e-mail staff, that would be one type of e-mail; but do not count each staff member’s e-mail address separately. E-mail to an external Web master would be another “type” of e-mail. For opportunities to add content, count each option separately – a guest book is different from adding an item to a calendar. But if there are multiple different calendars all of which add content in the same way, you would only count that once. For opportunities to customize content, you code each “type” of customization opportunity. So if you can select a language option, that is one feature (even if there are multiple languages available). But if you can also customize the background of the site, that is another type of customization. Types of Interactivity Following are general descriptions of the types of interactivity that will be coded. The coding form provides places to code some specific common features in each of these categories, but you should feel free to use the “other” option for any feature that seems to fit one of these general descriptions of interactivity but for which there is not a specific coding category on the form. Human-to-Computer Interaction: Features that enable interaction with the computer but do NOT facilitate communication with another person and do NOT contribute to or customize the content of the site. Navigation is the “baseline” that makes Web sites function and allows users to find their way among various elements of the site. ○  Navigation/Standard– Use this category to code features that offer users options for how to navigate among the content. Menus, hyperlinks, buttons, and banners are clearly navigational tools. Things such as “clickable maps” and “drop down boxes” that offer users options for which part of the site they want to explore are also standard. Tools that help the user navigate through long and/or complex copy are also standard navigation (e.g. a “top of page” button) ○  Navigation/Personalized– These are navigational tools that give the user more control. For example, a search function would be personalized because it allows users to type their own search terms. Actions allow users to give information to the computer but do not result in a purchase or other clearly transactional exchange. ○  Action/Standard– code here actions that offer users a “set” of options that are the same for all users. For example, online surveys and polls would be standard because all users answer the same questions. Also tools such as “print this page” and “bookmark this page” are standard actions. ○  Actions/Personalized– These are actions that respond to specific information provided by the user. For example, login and registration activities would be personalized because the computer requires specific information that must be provided by the user. Transactions mean that the individual will receive something that has been requested through the Web site but is often delivered outside of the site itself. ○  Transaction/Standard– this category is for “unchangeable” items such as a book, brochure, or e-mail newsletter. ○  Transaction/Personalized– these transactions allow the individual to customize what he/she wants. For example, requesting a report based on a set of symptoms is personalized because the individual indicates specific information needed. Additional notes on Human-to-Computer Interactions ○  The line between standard and personalized is not always crystal clear. Generally, if the only selections are “yes/no” type options (e.g. I want this brochure, but not that one) code it as standard. But if the input from the user requires more individual information (e.g. requesting that a specific kind of content no longer be sent to a personal e-mail address) then it is personalized. ○  Even though graphic elements often “feel” more interactive than text, our rules do not give any preference to graphic elements. So a link that plays a video is not inherently any more interactive than a link that leads to additional text. ○  Navigation is clearly the most common form of interaction. What we want to code for here is navigations that are UNIQUE to the site. So do not include scroll bars or any other tools that are part of the browser function. ○  If a navigational tool leads to an action or transaction of any kind, then there is an added element of H2C interactivity. So, for example if the content is simply text describing a medical condition, there is no action or transaction and thus no additional H2C interaction. But, if it is an online “game” that teaches about a condition through an activity such as a crossword puzzle that would be an action/standard. ○  Do not “double-code” things that are all elements of a single H2C function. For example, if a login screen, has a place to type your name as well as a “button” that you select to actually login, you code only once for the tool (in this case it would be “Action/personalized”). ○  Remember that if there is a “solid clue” at level 1 that there is underlying interaction, you code both at level 1 where the clue exists and at level 2 where the feature actually appears. For example, if there is a login “box” on the first page, you code an “action/personalized” at level 1. When you click on the login function, if takes you to a screen where you provide your personal information, you’d code that new screen as another “action/personalized” at level 2. ○  Remember, that if there is a “solid clue” from level 2 that some new form of interactivity will appear at level 3 (e.g. a sub-menu for games) go ahead and code for the interactivity at level 2 (in this case action/standard). Human-to-Human Interaction: Features that enable people to communicate with other people. Organization/Individual Synchronous– Code here any tools that allow communication between the organization and the individual in “real time.” This could include live customer support, etc. Individual/Individual Synchronous– Code here any tools that allow for “real time” conversation among visitors to the site. This could be an IM function, for example. Organization/Individual Asynchronous– This is the most common form of H2H interaction. It includes “contact us” forms, e-mail links, etc. As per the general rules, multiple e-mail addresses in a “contact us” section would be coded only once. But if there are different types of contact information, they would be coded separately. Individual/Individual Asynchronous– Any tools that allow for “lagged time” two-way communication among site visitors. Examples include the option for e-mailing content to a friend, sending a virtual postcard to a friend, etc. Additional Notes on Human-to-Human Interactions ○  In making decisions about whether to use the Individual/Individual or Organization/Individual categories, follow these general rules. Code any feature that facilitates communication between any organizational entity and the individual as interaction between organizations and individuals (e.g. contacting third-party contractors would fit here because they are organizations providing services through the site). Reserve the individual-to-individual coding for interactions that allow the site visitor to communicate with other individuals whom they choose to contact. ○  Make your best judgment call about whether the interaction is synchronous or asynchronous, (e.g. online customer support is probably synchronous, e-mail sent to the Web master is probably asynchronous). ○  Do NOT include here postings to bulletin boards/chat rooms that become part of the content of the site. Those are Human-to-Content interaction. However, use the “general rules” as appropriate. For example, if there were an “ask us” link at level 1 that WOULD be coded here because it looks like it will facilitate communication with the organization. But, if at level 2 what you find is a “bulletin board” where questions are posted and answered, that level 2 interaction would be Human-to-Content and would be coded in a later section of the form because it is an opportunity for the site visitor to add to the content of the site. Human-to-Content Interaction: Features that allow the user to “engage” with content. Add Content – Typically features that add content also have some elements of human-to-human interaction, but they go further by allowing the user to contribute content that others can see. ○  For example, sending a “virtual postcard” would be human-to-human interaction because the message goes only to the intended recipient(s). But signing an online guest book would be human-to-content because anyone who visits the site can see the information. ○  Try as much as possible to code the “type” of content that sites allow to be added. There may be some overlap, but code what is dominant. For example, sites might allow users to post “scrap books” of their experiences. The dominant content type is probably photos although there may be some text in captions for those photographs. Customize Content – Typically features that customize content have some element of human-to-computer interaction in that they are often related to navigations and/or actions performed at the site, but they go further by allowing users to change the actual appearance of the site. ○  For example, an interactive map that allows the user to select a geographic area of interest is a navigational element. But if the user can choose to limit the site so that it ONLY shows information about one geographic region that would be a change to the content of the site. ○  Generally, most customizing will probably be to text. But note if other elements can also be customized. For example, changing screen backgrounds would be another type of customization. Other Interactivity: Use this category for anything that seems to fit the general category of interactivity, but can not be captured by any of the other categories. Provide a detailed description of the interactive feature. © 2008 International Communication Association
TI - A Multifaceted Tool for a Complex Phenomenon: Coding Web-Based Interactivity as Technologies for Interaction Evolve
JF - Journal of Computer-Mediated Communication
DO - 10.1111/j.1083-6101.2008.00420.x
DA - 2008-07-01
UR - https://www.deepdyve.com/lp/oxford-university-press/a-multifaceted-tool-for-a-complex-phenomenon-coding-web-based-Hbpv8e4R0Y
SP - 794
EP - 826
VL - 13
IS - 4
DP - DeepDyve
ER -