Congress on the Internet: Messages on the Homepages of the U.S. House of Representatives, 1996 and 2001
Sharon E. Jarvis,
Assistant Professor in the Department of Communication Studies and the Associate Director of The Annette Strauss Institute for Civic Participation at the University of Texas at Austin. She teaches and conducts research in the area of political communication. She is a co-author of Political Keywords: Using Language that Uses Us (Oxford University Press) and her articles have been published in such journals as Journal of Communication, American Behavioral Scientist, Political Psychology, and Political Communication.
Address: Department of Communication Studies, 1 University Station A1105, University of Texas at Austin, Austin, TX 78712 USA
Address: 3403 Wheat Street, Columbia, SC 29205 USA
This article explores the World Wide Web homepages of members of the U.S. House of Representatives from 1996 and 2001, focusing specifically on what the content posted to these sites suggests about their ideal audiences. Despite considerable changes in the content of business, entertainment, and political campaign sites over this five-year time period, we found few notable changes in the Congressional.gov pages over time. Moreover, the pages harbor particular assumptions about audiences that depart from other forms of Congressional communications and lag far behind the interactive innovations found in political campaign sites. These findings lead to suggestions for the creation of sites that are more in line with the potential the medium presents for the online public.
Part of what this is going to do is to get legislation and legislative materials beyond the cynicism of the elite and you are going to see a dramatic expansion of an intellectual populism that Jefferson would have dreamed of. If every citizen had access to information that the Washington lobbyists have, we would have changed the balance of power in America towards the citizens and out of the beltway. This program is really a major step in that direction. - Representative Newt Gingrich (Ubois, 1995, p. 45)
In 1995, Speaker of the U.S. House of Representatives Newt Gingrich (R-GA) held an ambivalent view toward politics on the Internet. On the one hand, the Speaker touted the Jeffersonian ideal that by increasing access to information, government could empower a larger segment of the citizenry. On the other hand, it took a full year after Gingrich made this statement before he launched his own World Wide Web page. His uncertainty about posting information about himself, however, was not unique to him or to the times. Three years later, Speaker Dennis Hastert (R-IL) and his staff had yet to set up a way to reply to email messages, even though his office was receiving almost 1,000 messages per day (Ault & Jones, 1999). Also in 1999, 22 members of the U.S. House had yet to establish email addresses, and some of those who did feared that emailing constituents “would place too great a strain on my resources and my staff's ability to keep up with their already heavy work load” (Rep. Barney Frank, D-MA), “might force me to hire additional staff to wade through non-constituent communications and ‘spam’ junk mail” (Rep. C. W. Bill Young, R-FL), and “may have the capacity to over-tax our own resources” (Rep. Philip Crane, R-IL) (all cited in Ault & Jones, 1999).
The idea that Congress might establish a presence on the Web in the mid-1990s was met with initial praise, and garnered impressive media coverage at a time when many portrayals of the institution were negative (Cook, 1989; Lichter & Amundson, 1994) and when public support for the U.S. House was low (Hibbing & Theiss-Morse, 1995). The idea also seemed to make sense in cultural time. During the 1990s, Web-based retail sales mushroomed, growing from $700 million in 1996 to $32.6 billion in 2001 (Chandrasekaran & Pressler, 1997; Regan, 2002), and major media organizations poured resources into their Web sites, integrating such pages into news, infotainment, and entertainment programming. Citizens also gained access to the Web at an unprecedented rate—the Web reached 50 million U.S. households in less than five years, as compared to 38 years for radio, 13 years for television, and 10 years for cable television (Schroeder, 2001). Although universal access was (and still is) a concern, by 2002 more than 174 million people, or 55 percent of homes in the U.S., had Internet access, encouraging Pew Internet & American Life Project Director Lee Rainie to observe, “The Internet has gone from novelty to utility for many Americans. They are beginning to take it for granted, but they can't imagine life without it” (Horrigan & Rainie, 2002).1
In the mid-to-late 1990s, there were many public calls for the U.S. Congress to use the Web. Reasons included to make governance transparent, to post messages to resurrect its image, and to bypass traditional news media filters to connect directly with the electorate. Yet practical concerns also surrounded the dissemination of information and the management of intellectual and staff resources. In the current study, we examine trends in the first five years of Congressional Web sites linked to official U.S. House site http://www.house.gov/house/MemberWWW.html. In doing so, we (1) investigate Congressional sites in 1996 and 2001, focusing on three concepts central to early promises about Internet politicking, and (2) discuss these sites in light of their ideal audiences as well as related genres of Congressional and online communication to gain insight into the motivations behind, implications of, and potential future of these Web sites.
Democratic Opportunities on the World Wide Web
Much of the research on Web politics published between 1994 and 1997 anticipated a bright future for online politicking—a pattern not unlike the early expectations for the democratic potential of radio (Hartley, 2000) and television (Croteau & Hoynes, 2000). Scores of hopeful scholars claimed that the Web offered renewed opportunities for democratic theory (Grossman, 1995), political learning (Browning, 1996), social movements and social organizing (Boncheck, 1995), citizen empowerment (Dertouzos, 1997), issue-based politics (Aikens & Koch, 1996), thoughtful, deliberate politics (Kurtz, 1995), communitarianism in virtual space (Jones, 1995), and electronic voting (Browning, 1996). A few forecasted a key role for the Internet in electoral politics as early as 1996; others suggested that the year 2000 would be the year when the Internet would become the primary way that candidates would campaign (see Johnson, Braima, & Sothirajah, 1999).
Many of these sanguine expectations were tempered with time, however, as scholars publishing in the late 1990s made a number of observations. Changes in electoral politics due to the Internet would most likely be incremental (Neuman, 2000); politicians were more likely to engage in media interactivity, not human interactivity (Stromer-Galley, 2000); Internet campaigns seemed to encourage individualized campaigns and shallow candidate narratives (Klotz, 1997); citizens continued to experience uneven levels of Internet access (Corrado & Firestone, 1996); and campaign managers generally refrained from directing traffic to campaign sites (Dulio, Goff, & Thurber, 1999). Academics and professionals began to believe that campaigns were more interested in the Web for its fundraising—rather than its participatory—activities (Rosenkranz, 2000), that campaigns were more innovative with their sites than were government agencies (Musso, Weare & Hale, 2000), and that the Internet had taken its place alongside other media outlets as a source of information rather than as a replacement for traditional ways of learning about politics (Althaus & Tewkesbury, 2000).
Democratic Opportunities of Congressional Web Sites
The present study is inspired by Bimber's (1999, 2000) call to focus on the features and types of information appearing online rather than conceptualizing technology as the variable of interest. Although a variety of political topics could be explored in an analysis of Congressional sites, three subjects appear particularly meritorious of study, as they were central to widely reported hopes for the Web's role in re-energizing American politics: information, identification, and participation. The following paragraphs detail why these three notions—which are also key to off-line politics—were selected for this project.
Information has been regarded as the “most extensive and most substantive factor of a Congressional Web site” (Browning, Caldow, Fose, McShea, & Ronen, 2001). Representative Gingrich and others have touted it as a democratic good that should be made available to the American people (Casey, 1996). The low cost, voluminous capacity, and high reach of the Web positions it as an ideal medium for posting information once difficult for the average citizen to obtain (e.g., campaign finance figures, lobbying data, voting information, position papers, press releases, biographies, historical facts about Congress, etc.). Two concerns, of course, faced the creators of governmental sites with regard to information: a desire to offer citizens direct access to political information, and a temptation to limit the amount of data shared with the public.
Early theorizing about the Web suggested that a variety of “communities” might emerge online, the byproduct of like-minded individuals finding each other in cyberspace. Rhetorician Burke (1969) might suggest that these communities could be built either through messages emphasizing shared commitments or common frustrations. In his mind, both “association” and “division” are key to identifying with an audience, as individuals can feel connected by both forces. The creators of the Congressional web sites were charged with developing an interface that would present members of Congress favorably, as well as crafting messages that would promote a sense of connection with the audience. The creators thus had to consider the following: Will the sites present the members as elites in Washington D.C. or as citizens of their districts? Will party cues be sharpened or eschewed? Will policy proposals, preferred agendas and political friendships be featured or hidden? Tracing these choices can inform us of efforts to identify with online audiences.
In the early 1990s, many cyber optimists viewed the Web as a tool that could decrease the transaction costs of organizing—a step that might increase participation, augment the legitimacy of decisions, and revitalize the democratic system. While few could discount the desirability of such civic blessings, the prospect of engaging citizens may not have squared with the key goals of Representatives (Mayhew, 1972) or with the size of their staffs. Creators of these sites had to navigate between these two temptations: posting opportunities for online engagement while protecting the objectives and operations of Congressional offices.
In this study, we analyze the content and form of the information, the markers of identification, and the opportunities for participation on Congressional Web pages. Seminal work from communication and political science provide a theoretical backdrop for this endeavor. First, Black's (1970) essay on the second persona outlines how all messages harbor cues about their authors as well as their implied or ideal auditors. According to Black, a valuable task for those interested in political messages is to examine texts for cues that describe the audience that the author would prefer to engage in communication. This perspective is valuable to the current study, as it permits us to analyze the information, identification, and participation content appearing on the Web sites in light of the audiences they imply—a strategy evident in other analyses of Web pages (e.g., Esrock & Leichty, 2000). Purposely or unwittingly, these Web sites contain cues as to their intended audiences; unpacking these cues allows scholars to gain additional insight into these texts.
Fenno's work on Congressional communication also informs this analysis. From 1970 to 1977, Fenno ventured into U.S. Congressional Districts with elected members of the House of Representatives and “looked over their shoulders,”“tried to observe and inquire into anything and everything these members did,” and “worried about whatever they worried about” (1977, p. xiv). Through this approach, Fenno learned that elected officials see at least four constituencies in their districts, including the geographical (the outlines of a district), the reelection (supporters who may re-elect a member), the primary (individuals who strongly support a member), and the personal (strong supporters and close advisors). While his research is based on strategic face-to-face interactions between Representatives and constituents, the categories derived from such observations are illustrative of similarly strategic but mediated Congressional communication. Elected officials know that it is important to edit messages differently in front of various audiences—saying certain things in certain places, avoiding the same statements in others. Fenno's findings regarding how homestyles are edited for specific constituencies encourages us to attend to the material that is highlighted as well as that which is downplayed for the online audience.
Thus in this study, we ask the following questions: What types of information, cues of identification and opportunities for participation appear on Congressional homepages? What type(s) of audiences appear to be addressed via these pages? Did the pages change between the years 1996 and 2001? Finally, what are some of the potential implications of these pages for the future content of Congressional sites?
A guiding premise behind this analysis is that the messages posted to the sites exist in a dynamic communication environment whereby audience members come to them with expectations formed by prior experience with politics, Congress, and the Web. As with any communication text sent via a medium, these messages are likely to be judged (consciously or subconsciously) in accordance with the expectations for that medium. To this point, elected officials know that the delivery of their public messages can be as important as their content, and that marginal messages (e.g., television advertisements with low production values, unpolished pieces of direct mail, or poorly delivered speeches or soundbites) can affect their public image and opportunities for success in politics. Thus, as the Web has moved from a technological curiosity in 1996 to a tool that is heavily integrated into American life by 2001, the mere existence of Congressional Web sites means that texts have been posted to the public domain that stand to be judged in light of other texts in the medium and to contribute to members' reputations, regardless of whether these sites are a priority to the elected officials.
This study utilizes the method of content analysis to examine the Web pages of members of the U.S. House of Representatives. Content analysis is a technique that allows researchers to classify items objectively and systematically according to explicit rules and criteria. This method of research ranks high on external validity as the findings are generalizable to other populations and/or social settings, and is low in internal validity as it does not make strong causal links between the variables being tested (Krippendorff, 1980). Although these sites could be examined from rhetorical or critical perspectives, we have chosen to analyze them via content analysis so that the findings can be compared systematically across time. To maximize the understanding of the web pages, both quantitative and qualitative measures are used and attention is paid to verbal and nonverbal messages.
Congressional Web pages were still in their infancy during the first wave of the study, thus analyzing them in 1996 and in 2001 permits an inspection of continuities and changes during their early years (coinciding with the .com boom in the U.S.). These sites were selected because of the excitement that accompanied Congress' movement online, because Congressional communication receives far less attention than that of the President, and because the size of their chamber provides greater potential variance for the analysis, .gov (“dot gov”) sites—as opposed to .org or .com campaign sites—were chosen because most of the initial praise for government use of the Web concerned increasing the legitimacy of governing beyond electioneering, and because campaign sites were difficult to locate (Davis, 1999), inconsistently advertised (Davis, 1999), and often not fully developed (Casey, 1996) at the start of this study.2
All 119 sites posted to the page http://www.house.gov/house/MemberWWW.html in 1996 were saved and examined. Five years later these same 119 sites were again archived and analyzed (minus the site of a Representative who died during the five-year interim). Thirty-eight members from the 1996 sample were no longer in the U.S. House of Representatives in 2001, so the sites of their replacements were included in the analysis. Naturally, these pages are imperfect substitutes, and all comparisons between these sites and the sites of their predecessors should be treated with caution. Nevertheless, the replacement pages reflect the sites that constituents would visit were they to look for their elected officials online.
In both 1996 and 2001, the pages were visited four to five months after a new Congress had convened, and just prior to the summer recess. Over the course of this study, the House of Representatives had a computer support office that employed staff members to help in the development of homepages: the House Information Resources (HIR). The HIR created a stock page that many Representatives borrowed, and in interpreting their web pages, it is important to acknowledge that many of the sites were developed from these templates. This pattern, however, does not diminish the importance of the amount and types of information on the stock templates; indeed, the templates provide a shell of the priorities developed by the HIR for interactions with an audience.
It is important to note that there were restrictions on the types of content that could be posted to the sites, and during the course of this study the Committee on Rules and Administration was in charge of establishing guidelines for the pages. Set rules were: (1) as early as 1996, Senators were prohibited from using their official, government-provided homepage within 60 days of an election, and (2) members of Congress were not allowed to use these government-provided homepages to solicit funds, to provide personal material, to electioneer, or to provide links to sites with these characteristics. After conducting the analyses, we contacted 10 Congressional webmasters of the sites we had examined to discuss any additional rules or norms regarding these sites. The webmasters confirmed the aforementioned rules, adding that web sites “may not include grassroots lobbying or solicit support for a Member's position; generate, circulate, solicit, or encourage signing petitions; include any advertisement for any private individual, firm, or corporation; or imply in any manner that the government endorses or favors any specific commercial product, commodity, or service.”3 It is critical to interpret the content of the sites in light of these regulations. At the same time, however, these rules are not totally unlike those enforced by other agencies (e.g., the FEC) on other forms of mediated communication such as television advertisements. The HIR restrictions are factored into this analysis and all interpretative work is mindful of the concerns they raise.
Unit of Analysis
The unit of analysis for this study is the Congressional homepage, accessible via http://www.house.gov/house/MemberWWW.html (cf. McMillan, 2000). Conceptually, the homepage was selected because it serves as a content gateway that “strongly influences whether or not a visitor will stay and investigate or leave for another cyberspace destination,” and displays commitments to “publics, issues, and communication styles” (Esrock & Leichty, 2000, p. 330). Many of the sites in the 1996 sample posted all of their content and links directly on the homepage (Casey, 1996), a choice that Ryan, Field, and Olman (2003) call the “long list of links” style, and that Davis (1999) and others have called “brochureware.” (Indeed, Davis found that 48% of Congressional campaign sites in 1996 did not feature any links to other pages; instead, campaigns directly uploaded their campaign brochures to the Web such that the homepage resembled one long page of text.) An example of the “long list of links” style is given in abbreviated format (the links continue below the portion of the page shown in the image) in Figure 1. In 1996, each homepage was visited and archived, and the coders opened all the hyperlinks, icons, and images to record what appeared on the homepage as well as one link beyond these features (see Ha & James, 1998).
In 2001, when the same homepages were visited and archived, we found that many of the sites were formatted somewhat differently. To accommodate the shift in homepage design, and so as not to omit potentially significant content, we elected to examine the entire site in 2001 to look for items that had been directly posted to the homepages in 1996 in the “long list of links” or “brochureware” formats (see Casey, 1996; Davis, 1999; Ryan, Field, & Olman, 2003). While we agree with Weare and Lin that “researchers can and should employ individual pages as the recording unit whenever possible” (2000, p. 282), we also agree with McMillan that “researchers must use rigor and creativity” (2000, p. 93) to protect reliability in comparing early sites to those appearing in later years.
The content analytic scheme is as follows. All 119 sites in 1996 and 118 pages in 2001 were read and archived, and all links off the pages were opened and their contents recorded. To track the information content of the pages, each site was observed for the characteristics appearing in Table 1. These items were selected because they were featured in early promises about Web politics (Casey, 1996). To investigate the identification content on these pages, the appearance of party labels or symbols of partisanship (e.g., graphics or pictures with recognizable partisans, textual references to same or other party leaders) were documented. Additionally, more subtle qualitative attempts to communicate association or division with the audience via issues, photos, or references to other political figures were noted. To examine the participation content on the pages, the appearance of contact information (phone, postal addresses, and email addresses) was logged. Moreover, technological variables often included in business sites to encourage return visits (such as “last updated” features and interactive polls and surveys) were observed.
Table 1. Content of web sites (in percentages of sites containing feature)
Note: 1996, n=119; 2001, n=118
U.S. House Site Link
Contact (Phone/U.S. Post)
These coding categories were devised to avoid judgment criteria. Nevertheless, a key methodological issue in content analysis is the uniformity of the interpretation and categorization of the items of interest. Thus, the second author examined all of the sites and two undergraduate students were trained as independent coders. Intercoder reliabilities for these objective items were all above .928.4 Both authors examined and discussed the qualitative nature of all sites in 1996 and 2001.
The quantitative results of the content analysis appear in Table 1. As the two columns illustrate, despite a few shifts over time, the content of the homepages did not change substantially overall between 1996 and 2001. The following sections describe the quantitative and qualitative patterns on the homepages, noting how they reveal unique assumptions about their audiences, particularly when viewed in light of other forms of Congressional communication and expectations of online communication.
The types of information posted to the homepages appear in the left column of Table 1. As illustrated there, the most common features on the pages were: (1) biographies (91% of pages had them in 1996, 88% in 2001), (2) photos (81% in 1996, 97% in 2001), (3) press releases (71% in 1996, 95% in 2001), (4) constituent service information (61% in 1996, 100% in 2001), (5) state concerns (67% in 1996, 69% in 2001), and (6) legislation (59% in 1996 and 56% in 2001). After these six categories, no other type of information appeared on more than half of the homepages.
At first blush, these purely quantitative data might suggest that these pages are a meeting place that is open to the public relations needs of a Congressmember (via biographies and photos) and the information gathering needs of the visitor (via press releases, constituent information, state concerns, and legislation). Upon further inspection, however, it emerges that these types of information presume a meeting with a certain kind of visitor: one who is informed about the member, aware of the inner workings of Congress, and interested in the members' Washington behavior.
To begin, certain beliefs about the knowledge of visitors are embedded in the 1996 texts. The first one might seem trivial initially, but in 1996 it was difficult for individuals even to locate their member's homepage unless they knew the member's name and/or district number. As most political scientists know, this no small feat for the average American (see Delli Carpini & Keeter, 1996). Shortly after the first analysis, the HIR placed “zip code readers” on the http://www.house.gov/house/MemberWWW.html site so that virtual visitors could type in their zip code and find their member—a step that made it easier for citizens to find their own member, but that did little to help citizens find Representatives from other cities without knowledge of their names, district numbers, or the zip codes they represented. Table 1 offers another datum related to geographical districts: fewer than 20% of the sites featured explicit information about district concerns in 1996 or 2001. Fenno (1977) observed that in face-to-face settings, Representatives have a heightened awareness of their geographical district, always remembering and emphasizing where they come from. While this may be a priority in face-to-face communication, cues marking links to a Representative's district (other than the number) were not prominently placed on these homepages.
A second pattern in these pages is that they “promoted the boss,” featuring cues that underscore the distance between the elected official and the visitor (see Browning et al., 2001). This pattern was particularly evident in the biographies, photos, and descriptions of Washington activities. First, as regards the biographies, 91% of the pages in 1996 and 88 % in 2001 featured links to biographical essays on the homepage—patterns that do not necessarily spell detachment between the Congressmember and the visitor. Qualitatively, however, the biographies tended to focus on the accomplishments and professional successes of members in very reserved and formalized ways (see Klotz, 1997). This choice may follow a practice of traditional Congressional communication (establishing the legitimacy of the elected official), but was far less personal than Fenno's (1977) description of visits with legislators in face-to-face settings, and far less intimate than many personal homepages that appeared during the late 1990s (Dominick, 1999). Consider just one example. Someone familiar with (1) the political setting in 2001 (in which a Republican president advanced an agenda that led one GOP Senator to break with the party), (2) Congressmember J. D. Hayworth (R-AZ) (a lively former television sports news anchor), and (3) trends in personal Web pages (often used to create relationships with visitors) might be surprised by Hayworth's site in 2001, which featured three paragraphs in the biography starting with phrases such as: “Congressman J.D. Hayworth has written…”“Congressman J.D. Hayworth was recently named…,” and “J.D. Hayworth is pleased to announce that he has been selected Chairman of the Conservative Opportunity Society (COS)…” In the terminology of Watzlawick, Beavin, and Jackson (1967), all messages feature content and relational dimensions. When compared to either spoken conversations with constituents or personal homepages, the priorities in this and other biographies may send detached and formalized relational cues to visitors.
So may the photos on the sites. In 1996, 91% of the sites featured a photo of the member and by 2001, 97% did the same. In 1996, almost all of the photos were suit-and-tie headshots in front of U.S. flags or the Capitol, a trend still evident in 90% of the sites five years later. Visually, this choice emphasizes the trappings of the office, signifying the position of the Representative rather than his or her relationship with constituents. Examples of such photos are shown in Figure 2. Casey (1996) believed that this practice made certain members look self-important. While it may seem natural to post a headshot or a photo in front of the Capitol on a .gov site, this would not be the first instinct of a campaign manager or a communications professional offering advice for interactions with one's constituency. Both sets of actors would most likely counsel the member to adopt nonverbal behaviors and dress in a way that emphasizes a connection with, rather than distance from, their audience. They might advise the types of practices appearing in the 7% of pages in 2001 that hosted photos of members interacting with constituencies, images that underscored interaction with a community. When compared to Congressional homepage styles or other online sites, the emphasis on the Congressmember in the biography and photos contributes to member-centered sites, pages that pay less attention to the visitor than other related forms of communication that aim to profit politically and economically by flattering the needs of the audience.
An emphasis on Washington D.C. is also evident in the content and form of the pages. Many homepages were decorated with visual images of the Hill, the most notable being the layout of Rep. Jane Harman's (D-CA) page which was literally a visual model of her office that invited visitors to imagine her engaging in D.C. activities (see Figure 2). A majority of the “constituent service” links in the sites in 1996 and 2001 did not address issues facing Congressional districts, but instead featured instructions for requesting flags that had been flown over the Capitol and information to help visitors organize their vacations to Washington D.C. The orientation towards Washington was so apparent on these pages that the 2001 sites of 18 Representatives featured no images of their home states. These patterns stand in sharp contrast to Fenno's observation that in empirical districts, “the Washington community is often described as a group of people all of whom come from somewhere else” (1977, p. 884). At a time when many individuals ran for Congress by running against it, these sites featured visual images connoting greater comfort with the symbology of Capitol Hill than with the individual districts.
The information content on the homepages also revealed assumptions that the audience understood the inner workings of Congress. While links to press releases, constituent services, state concerns, and legislation may offer quantitative proof that the sites featured materials to connect with broad audiences, the qualitative features of these links may have prevented them from being useful to visitors with only a causal understanding of how Congress operates. In reading through the links, we noticed that they took the form of “lists” rather than “stories” (Browning, 1992), of data posted without interpretation, background description, legislator commentary, or editorial input. Unlike news audiences, who are accustomed to learning about Congress via narratives that emphasize elites, conflict, power, and even the voice of the journalist (Bennett, 1996; Patterson, 1994), most of the information in Table 1 is not anchored by an authorial voice. For instance, the press releases are posted as they would appear for the mass media, constituent information appears without a backdrop of the responsibilities of the U.S. House of Representatives, and legislation materials appear without a description of the committee system. It thus stands to reason that the information presented is more meaningful for visitors who already have a thorough understanding of current events, the duties of the federal government, and the organization of the House of Representatives.
The design and content of the homepages in 1996 may have emerged to satisfy the early visitors to the sites. In Corman's (1994) study of the use and users of governmental Web sites, he found that pages were often visited by .gov and .edu domains, (i.e., individuals in government and at universities). Bonchek, Hurwitz, and Mallery (1996) also found that many of the users were already highly informed. Given these early understandings of the Web audience, it may be understandable that Web designers offered lists with limited commentary rather than richer stories to the visitor, the former marking an offertory that was efficient to a politically involved, sophisticated information seeker, and probably less meaningful to a less involved, less knowledgeable visitor. There could be other explanations for the types of information as well, including a general uncertainty with regard to the relative homogeneity or heterogeneity of Web site visitors, a desire to emulate prior pages, uncertainty due to the newness of the medium, and/or a general reluctance to be innovative in communications. Regardless of the exact motivation, these pages displayed member-centric preferences that depart from face-to-face Congressional communication in which strategic members would be less self-centered, more given to narrative, and less oriented to Washington than they appeared on the sites.
A second pattern in the data describes the ways in which cues of association or division appeared in content to encourage the visitor to identify with the homepage. The data show that few overt markers of connection with citizens were visible, a finding that builds upon the lack of local cues (and predominance of member and Washington-centric ones) on the pages. Perhaps the most prominent finding of this study is that members seldom featured their partisanship on their homepages. In 1996, only 21% of the Representatives featured the term “Democrat” or “Republican” (or a visual image indicating partisanship such as graphics of a donkey or elephant, or photos of a member with a leading political figure in their party) on their home page. This number did not change significantly over the five-year period; only 17% of members featured this information in 2001. To ensure that party symbols were not against the HIR guidelines, we interviewed 10 Congressional webmasters about this pattern. All of our interviewees told us that there are no restrictions forbidding posting partisan cues.5
A reluctance to feature party symbols held for prominent members in the leadership as well as for ambitious junior members. For instance, in 1996, there were no cues signaling Speaker Newt Gingrich's partisanship on his page—despite the fact that he (1) was the Speaker of the House (the leader of the majority party in this branch of the legislature), (2) displayed partisanship abundantly in other communication outlets, and (3) led the Republican revolution in the 1994 election, taking back the House for the Republican party after 40 years of Democratic control. Rep. Dick Gephardt (D-MO), the ranking Democrat at the time, featured one link to the Democratic leadership page in 1996, but no clear references to his partisanship in 2001. The pages of Rep. Patrick Kennedy (D-RI) and Rep. Nancy Pelosi (D-CA)—two “up and comers” in Democratic politics on the Hill in 1996—similarly hid partisanship in 1996 and again in 2001. In both years, Kennedy featured a headshot with an image of a sailboat (presumably from the coast of Rhode Island), and Pelosi included a head shot above a San Francisco cable car (an emblem of her district); like the leadership, these two did not include verbal (textual) or visual (photos, icons, etc.) emblems of membership in the Democratic party.
When links to partisanship were featured, they took the following form. First, consider Republican cues. Rep. George Radanovich (R-CA) was one of the members to label himself a “Republican” on his homepage in 1996; in 2001 his homepage also featured a picture of him and Republican President George W. Bush, as well as two news articles highlighting his partisanship. Rep. Deborah Pryce (R-OH) called attention to her partisanship in 2001 by posting a link to the House Republican Conference, as well as an article praising Republican tax cuts during spring and early summer of 2001. In addition, Rep. Ernest Istook's (R-OK) site featured the headline “Istook, Bush Celebrate Tax Cut for Oklahomans,” and a Tax Relief Calculator, which invited visitors to “see how much you will save from the tax cut bill.” Instances of Democratic cues from 2001 were a bit less pronounced than those of the Republicans. For instance, Rep. Barney Frank (D-MA) posted a link to the House Democratic Leadership, Rep. John Conyers (D-MI) placed a picture of himself and Democratic Senator Hillary Rodham Clinton on his site, and Rep. Peter DeFazio (D-OR) incorporated 10 articles describing himself as a Democrat on his page. While trends toward candidate-centered campaigning are well documented (Wattenberg, 1998), the choice to avoid party cues on Congressional websites is notable, particularly considering that partisanship organizes the House, serves as a powerful predictor of voting behavior, and carries meaningful cues for voters (Aldrich, 1995).
As Table 1 shows, issues were not a prominent feature of these homepages, and when they did appear, they often took the form of valence (e.g., economic prosperity, public corruption, resolute leadership), rather than sharpened, politics (Stokes & DiIulio, 1993). An example of a verbal instance of valence politics in 2001 can be found in Rep. Brian Baird's (D-WA) announcement of the formation of a bipartisan Congressional caucus to “bring the growing nationwide problems and dangers associated with the abuse and production of methamphetamine to the attention of Members of Congress.” It is difficult to imagine a visitor who would not applaud this move. Other pages featured visual images of valence politics in 2001, as for example the page of Zoe Lofgren (D-CA), which did not show a head shot of the member, instead offering a picture of Lofgren surrounded by smiling students from her district. In refraining from posting cues to partisanship or preferred policy positions, the sites seem related to Fenno's member facing a heterogenous district (1977, p. 907); that is, the strategies for identification behind these pages may be described as those of inclusion rather than division, of safe and positive politics rather than pointed, ideological discussions.
Our qualitative analysis revealed another apparent aversion on these pages: the mention of an “other.” While television coverage of Congress often emphasizes partisan bickering and conflicts with the President (Cook, 1989), these pages featured very few indications of political opposition of any kind, be it with another party, another member of the institution, another branch of government, the very idea of Congress, or another country. In the few cases when any opposition was shown, it tended to be mild. For instance, Rep. Diana DeGette (D-CO) posted a news release headline on her site that read, “Rep. DeGette Statement on Bush Administration Decision on Roadless Map.” When accessed, the release began with a bland statement, “I'm relieved that the Bush Administration has pulled back from their inclination to gut sound environmental regulations and let stand the US Forest Service Roadless Rule today.” Buried in the third paragraph was a statement forging a stronger partisan appeal: “I remain concerned that the Bush Administration will attempt to slowly weaken the roadless proposal in the coming months by allowing local forest managers to compromise protections. This decision would be an unacceptable return to old policies that have not been protective of our nation's forests.” This news release stood out on three counts: (1) an individual would have had to search for this partisan opposition on the site, (2) when located, this type of statement paled in comparison to the political conflict evident in news reporting and some Congressional debates, and (3) the webmasters interviewed did not regard such statements as a violation of HIR rules.
Not only were other actors or forces not named on these pages, but “bad news” of any kind did not appear on the pages. During the years of this study, a citizen would have little idea after visiting the sites that the U.S. House of Representatives had recently been in a conflict with the President so severe that it led to a shutdown of the federal government in 1995, nor would they have known that a U.S. Senator was in the process of leaving the Republican party for Independent status during the spring of 2001. While it may seem natural to forgo commenting on potentially harmful topics, it is important to mention that doing so would not automatically violate HIR rules. Indeed, two sites that we visited after this analysis—those of Rep. Barbara Lee (D-CA) and Rep. Gary Condit (D-CA)-both posted interpretations of difficult situations. For Lee, a rationale for voting against a post-September 11 bill was highlighted; for Condit, a pledge to work hard for his constituents (during a moment of personal scandal related to his inappropriate relationship with a missing intern) was underscored. In both instances, .gov Web pages became places where members could become the bearers of their own (controversial) news, to share their side(s) of complicated stories.
These two sites aside, a key assumption about the audience of these pages was thus that visitors were more interested in detached member information and valence issues than in partisanship, ideology, group membership, adversarial messages, or descriptions of individual choices. Taken together, these priorities constitute a heightened sense of the importance of individual members of Congress, and a limited understanding of the centrality of parties (Hetherington, 2001) and relationships (Uslaner, 1993) in the House of Representatives.
A third pattern in the data involves the expectations for Web visitors. What, based on the participation-oriented web site content, were visitors supposed to do? A few answers to this question suggest themselves.
A first expectation for the audience appeared to be that they were to visit, but not interact with, the sites. In 1996, 52% of pages featured email addresses, a figure that increased to 96% in 2001. While this is an impressive quantitative shift, the qualitative cues surrounding the “contact member” links in 2001 seemed to promote distance as much as they invited interaction. At that time, many sites posted the “U.S. House of Representatives Privacy and Security Notice” right next to the email address of the member. This notice described how members were not collecting data about visitors without their approval. While the notice is a legitimate and important cue to send a visitor on the site, when positioned next to the email address it has the effect of heightening privacy concerns, framing Web interaction as potentially risky rather than mutually beneficial.
The opportunity (Ault & Jones, 1999) and threat (Stromer-Galley, 2000) of email interaction with constituents has received coverage in both popular and academic circles, and relates directly to how many resources Congressmembers can or wish to allocate to this medium. With regard to the former, Ault and Jones (1999) suggest that “e-mail could save some money—Congress spent $31.4 million on postage alone in 1998, not including costs for paper and preparation of mail,” and use the case of Representative Jay Inslee (D-WA) to support this suggestion. Inslee's staff believed that their electronic communication saved both time and money, and that the 500 to 1,000 emails the Representative received daily in 1999 dwarfed his incoming postal mail.
Yet concerning the latter, Casey (1996, p. 143) argued that “members of Congress should not treat the Internet as a broad-cast only medium, pumping it full of their speeches and statements, if they are not prepared to allow for feedback.” As Weise (2000, p. 39) observed,
[t]he problem (and the blessing) of the Net is that it is most potent in its original one to one form. E-mail was and is likely to be the killer app. It's e-mail, instant messaging and chat rooms that lure most people on-line, not the chance to shop, whatever retailers want to believe. But except in small local elections, there's been almost none of that kind of one-on-one interaction happening between candidates and the electorate. Instead, we're given what in Web parlance is called “shovelware” (think regurgitated) speeches, position papers, and photos shoveled onto Websites to fill them up.
A key goal for many business sites between 1996 and 2001 was to use the Web to reduce costs and increase profit by encouraging visitors to answer their own questions and conduct their own transactions online (Esrock & Leichty, 1999). To meet these goals, Web designers created sites that would be pleasant to navigate, encourage return visits, reduce support costs, engage customers in dialogue about products and services, and educate the company about customer concerns (King, 2003; Lesser & Fontaine, 2002). Our data show that Congressional sites met just one of these criteria, becoming a bit easier to navigate over time (due to slight organizational changes, see Casey, 1996; Ryan, Field & Olfman, 2003). While .gov sites should only be compared to .com sites with great caution (conceptually, their missions are different; rhetorically, some business practices are forbidden by HIR), viewing the Congressional sites in light of pages created with a competitive vision of the user in mind can be illustrative.
First, a common practice on .com sites is to feature cues that the page is current—either through constantly changing content, dating postings, or presenting icons indicating that the material has been updated. In our data, only 14% of the sites featured a “last updated” feature in 2001, and the average update was 37 days from the recording date. A second innovation in Web sites is to feature interactive items, letting the visitor engage with the page and potentially allowing the site to learn from such engagement. Just 5% of the sites in 1996 and 2001 featured such interactive features. Another innovation for business and political campaign sites is to post links to media articles that praise the organization or candidate. This public relations practice allows a visitor to read up on what other voices have to say about the product or candidate promoted on the site. Very few Congressional sites posted links to news content in either 1996 or 2001 (although the Condit site did so); our analyses of the pages since 2001 suggest that more pages are adding such links over time. Taken together, the Congressional web sites did not encourage a return visit as other business and campaign sites do; moreover, members appeared to try to reduce support costs by leaving the sites dormant for long periods of time. Both of these strategies may have the short-term benefit of reducing traffic, but their long-term public relations consequences could be damaging.
If citizens come to regard these sites as part of constituent service work, and if the sites continue to pale in comparison to other types of sites, with pages that are member-centric, do not inspire identification, and are rarely updated, they may serve as more of a public relations liability than as a benefit for Congress. In face-to-face situations, elected officials work to keep citizen expectations below what can be delivered in order to avoid disappointment. This process often involves emphasizing symbols or relationships to specific audiences when preferred policy outcomes cannot be provided. Congressional sites did not follow this logic over the first five years. Whereas Fenno (1977) observed elected officials editing their message for constituencies, offering broad but inclusive statements to diverse groups and more specific ones to loyal groups, Congressional sites do not include the cues of identification that would be given to a diverse audience, nor do they share the specifics that would be given to a loyal one. To better manage visitor expectations, it would make sense either to visibly embrace the promise of the medium and offer sites that are informative, interactive, and regularly updated, or pare down the number of pages and offer thin but aesthetically pleasing sites that present cues of identification and do not pretend to house vast amounts of information. Although the former would be regarded as more desirable by cyber-optimists, either approach would be more strategic than the current practices.
The limited amount of interactivity on Congressional homepages might surprise observers of on-line campaigning—a topic that has generated considerable media and citizen interest. One of the bigger stories of the 2000 presidential primaries was the ability of Republican candidate John McCain to raise over two million dollars via his website in the week following his victory in the New Hampshire primary (Mintz, 2000). In campaign 2002, web strategists applauded the role of the Internet in campaign organizing, connecting volunteers with the campaign, allowing supporters to email materials to their friends, and linking canvassers to campaign headquarters (Drinkard & Lawrence, 2003). In the early days of campaign 2004, former Vermont Governor Howard Dean garnered citizen attention and news coverage for his web presence which featured meetups and web logs (blogs) and raised $19 million in web contributions (Baker, Green, & Hof, 2004). 2004 also witnessed the blossoming of web-based activist groups, including MoveOn.org, a grassroots effort born out of opposition to the impeachment of President Bill Clinton in 1998, that has raised money for several candidates' causes (largely Democratic) and boasts an email subscriber list of 1.7 million individuals (Cochran, 2003).
The limited amount of interactivity on Congressional homepages, however, might not surprise those who compare the activities of candidates and elected officials. In analyses of off-line campaign and governing speech styles (Hart, 2000), of the adoption of webpages (D'Alessio, 2000), and of the effort put into early campaign pages (Davis, 1999), researchers have consistently found that candidates and viable challengers (e.g., those with sufficient resources) are far more likely than elected officials or incumbents to embrace communication styles and technologies that involve audiences. As Hart (2000) puts it, the act of campaigning can frustrate an imperial style; in his mind, campaigns bring immediacy to governance and re-engage leaders in the lives of the electorate. An example of this tension can be located in a public statement made by Representative Ehlers (R-MI), one of the early supporters of on-line efforts:
Our nation, as a republic, depends upon the discerning and analytical judgment of information by individual Representatives. While information technology promises to enhance the democratic marketplace of ideas, it is appropriate that decisions finally rest with the authority of those elected Representatives. (Casey, 1996, p. 109)
Two aspects of the statement are notable here. First, the statement focuses on elected Representatives, rather than on citizenry. Second, the statement addresses the intersection of information dissemination and resource management in a democratic republic, ultimately siding with the latter over the former. These words are reminiscent of a basic observation about power in a political system: those who have it rarely want to give it away. This truism may explain why.gov pages have lagged behind the innovations found in business, entertainment, and campaign sites in a booming information economy (Esrock & Leichty, 1999). In summary, despite the empowering rhetoric surrounding the movement of Congress onto the Web, and despite innovations by individual candidates to engage the public in their attempts to get elected, the few interactive opportunities found on Congressional homepages in 1996 and 2001 encourage us to believe that the ideal virtual visitor would observe, and ultimately defer to, members of Congress.
This study has examined the homepages of the U.S. House of Representatives in 1996 and again in 2001. Public officials know that the delivery of their messages can be as important as their content and that it is important to edit messages for different audiences. Viewing the homepages as texts and querying about their intended audiences and potential changes over time has allowed us to interrogate these sites and gather insight as to how they resemble and depart from other forms of Congressional communication. Attending to assumptions about the information, identification, and participation content of value to visitors provides fresh insight into these sites as well as prospects for their future.
In analyzing the sites' content, we found that the pages did not change between 1996 and 2001 as much as pages in the business, entertainment, and campaign sectors did. Moreover, they offered information content that assumed the audience was aware of the inner workings of Congress, impressed by the qualifications of the members, and interested in the members' Washington behaviors. Thay also hesitated to build a connection with visitors by posting party cues, preferred policy information, references to other political figures, or explanations for political decisions, and they refrained from presenting material to encourage online interaction or even return visits from the audience. On one hand, these findings may be related to the relative newness of the sites and a general ambivalence of Congressmembers (like Gingrich in the introduction) to embrace the medium fully. On the other hand, this ambivalence may send a message in itself.
The early debate about political Web sites focused largely on the categories and amounts of data that would be posted to the sites—the material that would appear on the site. This is important, but so, too, is the material that does not appear on the site. Sites that emphasize the Congressmember him- or herself, that limit shared cues, and that are rarely updated may unwittingly distance users from pages, and potentially from elected officials as well. While the verbal messages on the pages resemble the careful and inoffensive messages that Fenno (1977) observed being uttered to “heterogenous” constituencies, there are few concordant nonverbal messages inspiring the visitor to identify with the communication transaction. At a time when constituent visits, campaign messages, and online business texts “are all about the audience,” the pairing of careful information and meager attempts at identification may limit the sites' impact. By not keeping up with offline or online communication practices, these pages run the risk of (1) disappointing visitors by not offering current information or overtures of identification, (2) missing opportunities to explain political positions (as was done by Lee) or defend oneself (as was done by Condit), and (3) overlooking chances to provide positions to the media.
Critics of current Congressional Web efforts claim that government has always lagged behind developments in communications technology. In a speech addressing Congress' willingness to share information with citizens, a USA Today foreign editor stated that in 1889, Congressmembers worried that the speed of sending messages by telegraph would ignite international conflict by “robbing (diplomats) of sufficient time to think” (“Congress ‘Clueless’ on Using Internet,” p. 1). Graeme Browning advanced a similar view, noting that “Congress is not interacting with voters (using technology such as email) because they really don't want to. They liked it the way it was 20 years ago … when they were kind of walled off and things were filtered through (their staffs)” (“Congress ‘Clueless’ on Using Internet,” p. 1). Such reluctance may mean that members are not excited about their web sites; it does not mean, however, that citizens will not read messages into their reluctance. User data suggest that individuals are visiting these sites. In 2001, Congress received more than 117 million email messages, averaging almost 330,000 per day—a 186% increase since 1999 (BBC, 2002). Additionally, Browning et al.'s (2001) data show that U.S. House Web sites logged roughly 500 million hits in 2001, and the top House sites received almost 3,000 unique visitors each month. Thus, even if members are not energetic about their sites, the content that is and is not featured on their homepages may have an impact.
A few possibilities exist for the future of Congressional web sites. If campaign sites continue to become more innovative, politicians and their staffs may become more comfortable and creative with .gov pages. If constituents become more interested in the sites and reward individuals who host pages that are innovative, a spirit of competition may lead to improvements across sites (Scammell, 2000). If the media continue to visit the sites, and lead the way in technological innovation (Cornfield, 2000), incentives may emerge to allocate more resources to .gov sites. Stephens (1998) suggests that digital technologies may change the way individuals look at the world, leading citizens out of the “age of insufficient information.” While this may be true, political communication scholars would be wise to recall that there is more to a Web site than just information. The topics addressed and ignored, the associations inspired or neglected, and the audiences invited and uninvited all contribute to the political messages on homepages and have potential effects for the political images of Congressmembers as well as their opportunities for success. Attending to the audience has allowed us to gain fresh insight into these homepages and learn how they depart from other types of Congressional and .com messages; a similar attention to the audience may allow for the creation of future homepages that are more in line with the potential of the medium to serve the online public.
As of 1996, there were around 12 million Internet users in the United States and 35 million users worldwide (Weber, 1996). During that time, approximately 35 percent of U.S. households had a computer (Marketing Tools, 1996). Five years later, this figure had jumped dramatically. According to the Central Intelligence Agency's World Factbook (2001), the number of Internet users in the U.S. during the second phase of this study in 2001 had risen to 166 million, and topped 513.4 million worldwide. Additionally, nearly two-thirds of Americans (or 65%) reported having a home computer in 2001 (Horrigan & Rainie, 2002)
Although Davis (2000) reports that after the 1996 campaign, “megalists” of campaign sites (“Project Vote Smart,””GTE.Net,” and “Campaign 96 Online”) featured up to 521 Congressional campaign sites for that year (p. 191), he also notes how difficult it was for citizens to find those pages (p. 88) because many of the Congressional campaign sites were not listed with major search engines (p. 110).
Email communication with webmaster Bill Cole, Office of Representative Elijah Cummins (May 2, 2002).
Intercoder reliability was measured as percent agreement. As the classifications were designed to be objective and categorical, the intercoder reliabilities are expectedly high: .932 for the first coder and .928 for the second.
Email interviews were held with ten webmasters for Congressional offices. The justification that webmaster Kevin Bishop (Office of Representative Lindsey Graham) provided for this trend is that, “members represent everyone in a district, not just Democrats or Republicans” (email communication, May 2, 2002).