Communication Institute for Online Scholarship
Communication Institute for Online
Scholarship Continous online service and innovation
since 1986
Site index
 
ComAbstracts Visual Communication Concept Explorer Tables of Contents Electronic Journal of Communication ComVista

Presidential Web Sites as Sources of Information
EJC logo
The Electronic Journal of Communication / La Revue Electronique de Communication
***** REAVY ******* EJC/REC Vol. 7, No. 3, 1997 ************


PRESIDENTIAL WEB SITES AS SOURCES OF INFORMATION


Matthew M. Reavy
Manship School, Louisiana State University

David D. Perlmutter
Manship School, Louisiana State University


        Abstract.  Scholars recognize that knowledge
     represents power, particularly in a participatory
     democracy.  In modern society, the Internet offers
     a unique opportunity for candidates to interact
     with the citizenry, potentially increasing the
     transfer of political knowledge.  This study
     examines the Web sites of three top candidates in
     the 1996 GOP primary to gauge their channel
     effectiveness - the degree to which candidates
     take advantage of the medium's unique capacity for
     immediacy, interactivity, sourcing and multimedia.
     Attention is devoted to how effective
     communication on the Web might promote a knowledge
     gap between those who have access to the Internet
     and those who do not.


     "My sources are unreliable, but their information
     is fascinating."
                                   --Ashleigh Brilliant

     That knowledge is a form of potential power for
individuals and an influence on the division of power in
society has long been recognized by philosophers and
statesmen.  Aristotle was said to have claimed that those
with knowledge differ from those without "as much as the
living from the dead" (Laertius, 1942 ed.).  This is
distinctly true in a participatory democracy, where the
citizens' knowledge of politics and politicians plays a
crucial part in determining the course of a nation's
government.  Those who possess knowledge about their
politicians tend to adopt a more active role in determining
which of those politicians shall govern and which shall be
governed (Rosen & Merritt, 1994).

     In modern society, the Internet offers a unique opening
for political candidates to interact with the citizenry,
potentially increasing the political knowledge of those who
take advantage of this opportunity.  Ideally, such a
situation creates a more enlightened electorate before which
"tyranny and oppressions of body and mind will vanish like
evil spirits at the dawn of the day" (Jefferson, 1899a ed.).
Perhaps the most salient ground for this interaction lies on
Web sites operated by the various candidates where an
electronic form of old fashioned retail politics may be
practiced, in theory, if not in fact:  candidates speaking
directly to votes, voters responding to leaders, all without
interpolation by other agents of media.  However, this
presumes that universal access is inherent to such
communication.  In fact, even in the United States only 13
percent of the population has access to the online
environment (Nielsen Media Research, 1995).  Moreover, the
U.S.  Internet population differs systematically in its
demographic and psychographic profile from that of the
nation as a whole, being significantly more homogenous in
terms of race, education, occupation, gender, age and income
(Pitkow & Kehoe, 1995).

     Researchers have noted that different groups in society
acquire, retain and employ knowledge at different rates and
with varying effectiveness.  This gap between the
"information rich" and the "information poor" is and has
been a real problem (Price & Zaller, 1993).  But that
political Web sites should contribute to such a divide might
not be so obvious.  Indeed, some journalists have argued
that politicians' use of new media constitutes little more
than glitzy advertisement with little or no information
content (Beckel, 1996).  On the other hand, the Web has been
cited as a great potential venue for political discourse,
potentially eclipsing all other media (Gates, 1996).  It is
of importance for communications researchers a) to ask how
can the Web's characteristics as a medium fit into and
expand traditional conceptions of mass and interpersonal
communication in politics, b) to assess the types, quantity,
quality and accessibility of information available on the
WWW, and c) project the future development and impact of
political communication on the Web.

     This study is a first step toward such a research
program.  We examine the Web sites of three top candidates
-- Bob Dole, Patrick Buchanan, and Steve Forbes -- in the
1996 Republican presidential primary in an effort to gauge
the existence, abundance, form, and persuasive ability of
their information.  Many aspects of the Web sites, including
their accuracy, content, and quality of the message
themselves, can be studied.  However, we argue that a first
step is to evaluate the channel effectiveness -- the degree
to which candidates take advantage of the new medium's
unique capabilities such as immediacy, interactivity,
sourcing and multimedia.  To this end we analyzed
information at these sites, more than 1,600 printed pages of
data, employing an instrument that attempts to measure
channel effectiveness within the framework of traditional
mass communication research.  We also speculate on the
latent meanings inherent in Web political discourse and its
possible effects on the electorate, the democratic process
and media representation of politics.

                     Literature Review

     Knowledge -- as opposed to purely wealth and birth in
the countries of the old world -- has been considered one of
the key elements of social and political mobility in the
United States (Watkinson, 1990).  Jefferson argued that no
nation can expect to be both "ignorant and free" (Jefferson,
1899b ed.), a sentiment echoed by James Madison who noted
that "liberty and learning" support each other like marble
columns--if one were to fall, so would the other (Madison,
1910 ed.).  Thus, the acquisition of knowledge is not merely
a pastime nor an outlet for excess energy, but rather
essential to the continued existence of a free society.  It
follows that, while all areas of knowledge may have some
relevance and importance to groups or individuals, political
knowledge is the keystone to participation in the democratic
process and to mobility within the social system.

     Despite the crucial need for information in a
democratic society, certain groups perennially lack
knowledge of important social and political issues (Hyman &
Sheatsley, 1947).  If information is indeed power, then
differences in knowledge contribute to the power of the
elites over the majority (Moore & Tumin, 1949).
Discrepancies in the quantity and quality of knowledge held
by groups might perpetuate differences in power within a
society (Donohue, Olien & Tichenor, 1987; Olien, Donohue &
Tichenor, 1983).  Such differential absorption of knowledge
has profound political implications in nations, like the
United States, that profess universal suffrage.

     Within communication studies, the strongest model for
noting and predicting social differences in knowledge
acquisition and retention is the knowledge gap hypothesis of
Tichenor, Donohue & Olien (1970).  As originally conceived,
the theory posits that:

         As the infusion of mass media information into a
     social system increases, segments of the
     population with higher socioeconomic status tend
     to acquire the information at a faster rate than
     the lower status segments, so that the gap
     between these segments tends to increase rather
     than decrease (Tichenor, Donohue & Olien, 1970:
     159-160).

In a recent review of 25 years of research on the
knowledge gap hypothesis, Viswanath and Ferguson noted that
one of the theory's foremost strengths is that it "predicts
the impact of information flow on knowledge equalization"
(1996, p. 218).

     The effects of information flow should be measurable
and may occur even when knowledge equalization is the goal
of information diffusion.  One celebrated example of such a
paradoxical knowledge gap involves the program "Sesame
Street," which had been targeted at children of lower
economic status in an effort to raise their educational
level.  The program was later found to unintentionally
accelerate the learning of higher status children, whose
home environments were more conducive to absorbing,
retaining and applying the new knowledge (Ball & Bognatz
1970; Katzman, 1974).  Some researchers believe this problem
illustrates that higher SES groups utilize more cultural and
economic resources to aid in assimilating new knowledge
(Price & Zaller, 1993; Gaziano, 1984; Hyman, Wright & Reed,
1975).

     Since its inception, the knowledge gap hypothesis has
undergone many refinements and extensions (see Gaziano
1983), expanding from its original application in the
analysis of newspaper readership rates to use in examining
any communication situation where individuals collect
information.  Rather than a unilateral hypothesis, knowledge
gap is a "scientific research program" (cf.  Latakos, 1968,
1970) that is a "contribution to the understanding and
explanation of society" (Viswanath & Finnegan, 1996, p.
218).  It can thus be applied to new learning situations,
new knowledge campaigns and novel technologies -- such as
the World Wide Web.

     Previous research on so-called knowledge gaps has
labored under two implicit assumptions.  First, it is taken
as self-evident that access to the mode of communication is
potentially available to all citizens.  For example, studies
of diffusion of televised information assumes that everyone,
even if they don't own a television, potentially has access
to one.  Indeed, television ownership and viewing among
disadvantage groups is relatively high.  Second, knowledge
gap research assumes that the information requires no
special technical skill, beyond literacy to absorb.  In
other words, the information, though perhaps in its present
form may be too complex for certain groups to ingest, is
potentially alterable to be informative to all.  An example
might be information on new farming methods in developing
country where radio spots fail to take local customs into
account when phrasing the ads.  These assumptions cannot be
made about the Internet.  It is not clear whether web-based
information can ever be affordable or technically accessible
to all citizens or can be restructured to be easily absorbed
by all.  At this early stage of Web popularity, when only
fraction of the population has access, a preliminary
question to ask, therefore, is whether the information on
the Web really has any value beyond information available
elsewhere.  In other words, we want to ask if people without
the Web are missing anything in either information content
or richness.  In the case of new media, no claim of a
"knowledge gap" can be made without such a first step.

     Many factors have been suggested as moderating,
affecting or intervening with differential rates of
knowledge acquisition.  These include the social structure
of a community, the geographic limits of a community, the
geographic limits of an issue, the impact of an issue on a
community and the flow of information--how much and how
often information appears within a community (Donohue,
Tichenor & Olien, 1975).  One study (Delli Carpini, et al.
1994) found that citizens of Northern Virginia tended to be
more informed about national issues, but less informed about
state issues than residents of Richmond.  The study
suggested that differences were probably due to location --
the former group proximate to Washington, D.C., and the
latter residing in the state capital.

     The relevance of an issue and the motivations of groups
have also been suggested as interpolating knowledge gap
rates (Ettema & Kline,,1977).  Specifically, lower SES
groups do not have the interest or perceived need to acquire
certain types of knowledge (see also Pan, 1990; Tichenor,
Donohue & Olien, 1980; Wade & Schramm, 1969).  Research also
suggests that a community's "boundedness" -- not merely
geographical, but also social, racial, religious or ethnic
-- may affect the presence and degree of a knowledge gap
(Webber, 1963).  For example, the knowledge gap between
African Americans and other citizens is relatively narrow on
the issues of civil rights and crime, despite the former
group's generally lower SES (Viswanath, Kosicki, et al.
1993).  The researchers suggested that these issues have
greater immediate relevance to the African American
community.

     Motivation of the receiver has been noted as a possible
intervening variable in knowledge acquisition (Ettema &
Kline 1977; Ettema et al., 1983).  Indeed, idiosyncratic
characteristics of audiences and individuals can ameliorate
the knowledge gap (Sears & Freedman, 1967; Dervin, 1980).
Furthermore, certain types of knowledge in certain contexts
may encounter a "ceiling effect" whereby the have-nots can
catch up with haves (Ettema & Kline, 1977).  Other
researchers persist in arguing against wholly individual
level explanations for knowledge gaps and instead continue
to cite differential SES status as the major determinant in
most situations (Gandy & El Waylly, 1985; Viswanath, 1990).

     Studies have explored relationships between the
structure of a knowledge gap and the type of information
being disseminated, the complexity of that information and
the channel employed for its transmission (Gaziano, 1983;
Viswanath & Finnegan, 1996).  For example, topic-oriented
studies have examined the presence or diffusion of
information on public affairs (Robinson & Levy, 1986), the
environment (Donohue et al., 1975), health (Viswanath, 1990;
Viswanath, Finnegan & Kahn, 1993; Zandpour & Fellow, 1992),
and agricultural innovation (Hornik, 1989).  The complexity
of information has been found to be a significant factor in
differential rates of diffusion.  That is, gaps in awareness
of simple knowledge (Viswanath et al., 1994) have been found
to be smaller than gaps in complex data (Galloway, 1977;
Viswanath et al., 1994; Gaziano, 1984) between low and high
SES groups.

     In addition, studies have examined the effect on
knowledge gaps relative to the type of medium employed for
information transmission, including print media (McLoed &
Pearse, 1994, Price & Zaller, 1993), television (Galloway,
1977, Simmons and Garda, 1982), and interpersonal discussion
(Donohue et al., 1975; Griffin, 1990; Hornik, 1989).
Several studies have also compared television-derived
knowledge to that acquired through newspapers (McLoed &
Pearse, 1994, Lang & Lang, 1984).  Channel processing
research has not produced a clear set of effects, but the
absorption of information seems to be equally governed by
SES status of the receivers and the level of exposure to the
information (Robinson 1972).  For example, higher SES groups
seem to use newspaper derives knowledge for making
citizenship decisions like voting more than lower SES groups
(Loges & Ball-Rokeach 1993).

     New media types present a challenge for researchers
exploring knowledge gaps in contemporary society.  Some have
seen cable television, videotext, and computers as agents of
equalization, closing gaps between the haves and the
have-nots (Compaine, 1986; Parker & Dunn, 1972).  However,
such optimistic views are premised on such technology being
both universally accessible and of interest to all groups.
By contrast, it is precisely the higher SES haves who are
more likely to have the means and the interest required to
adopt and employ new technology (Berg, 1984; Ettema, 1984;
Finnegan, Viswanath & Loken, 1988; Rogers, 1976; Scherer,
1989; Tomita, 1989).

     Data available from the U.S.  Census Bureau confirm
unequal access to technology among the citizenry.  The
statistics show that, while 9.6 percent of white Americans
had access to a computer in 1983, only 4.4 percent of
African Americans and 4.1 percent of Hispanic Americans had
access to the machines.  Those numbers narrowed a few
percentage points during the next decade as computer
technology began to spread among all groups, but the
technology gap remained conspicuous.  One can assume that
minorities continue to lag behind whites in their access to
computer equipment.

     Even more striking is the disparate presence of whites
and most minorities on the Internet.  A 1995 survey of
Internet users reported that nearly 86 percent of U.S.
Internet respondents were white, while only 1.84 percent
were Hispanic and 1.47 percent were African American (Pitkow
& Kehoe, 1995).  These results clearly do not reflect the
racial make-up of the United States as a whole.  The study
showed similar disparities with regard to gender (67.5
percent male), age (( = 32.7) and income (( = $63,000).
Moreover, few members of traditionally blue collar
occupations were found to use the Internet.  Not
surprisingly, four traditionally white collar job categories
constitute approximately 90 percent of the Internet
population:  computers (29.1), educational (30.9),
professional (19.9) and management (10.2).  We believe it is
no coincidence that these also comprise the upper SES groups
noted in the knowledge gap hypothesis.

     The Internet, particularly the World Wide Web, thus
serves as a new focus for debate about the dissemination of
information in a democratic society as well as the nature of
future knowledge gaps.  The Web undoubtedly has the
potential to act as a source of information for political
decision-making, but one wonders whether that potential is
being realized.  In short, do the Web sites of political
candidates offer real information, or are they simply
virtual burlesque-- a hollow imitation of political
intelligence?  This is especially of concern, in light of
findings that show that the Internet population differs
markedly from the U.S. population as a whole.  Thus, as the
title of our paper suggests, we ask whether this represents
a knowledge gap in the making.

     Before any such conclusion can be made, one must
analyze in depth the types of political information
available on the Net.  We argue that, before an Internet
knowledge gap can be determined, researchers must examine
the channel itself in order to assess its capacity for
transmitting information that would contribute to such a
gap.  Obvious loci to begin such a project would be the
sites of the candidates themselves.  Accordingly, our
instrument taps into two basic concerns of the knowledge gap
program of research, relevance and complexity, in order to
examine candidate Web sites.  We assess these issues through
an instrument that outlines ten categories of channel
effectiveness:  1) timeliness, 2) proximity, 3) prominence,
4) source diversity, 5) message diversity, 6) presentation
diversity, 7) volume, 8) accessibility, 9) image consistency
and 10) issue consistency.  Timeliness, proximity,
prominence, source diversity and message diversity tap into
the issue of relevance, while presentation diversity,
volume, accessibility, image consistency and issue
consistency address the issue of complexity.

                        Methodology

     If one wants to study U.S. politics on the Web there
are many sites that profess to offer information.  However,
perhaps the most obvious sources of information on political
candidates are the sites of the candidates themselves.
Ideally containing the full text of speeches, press
releases, issue stances and the like, these serve as the
primary sources on the Net, as opposed to sites that offer
secondary analysis of this data.  Thus, in preparation for
this study the researchers visited the "official" home pages
of all Presidential candidates.  One can find lists of such
sites in a variety of places on the Internet; however, we
selected those pages listed as official by the Yahoo!
subject-oriented Web index (Yahoo, 1996).

     Although it would be worthwhile to perform an in-depth
investigation of the Web sites of all political candidates,
the volume of the data and the exploratory nature of this
study made it more practical to limit the number of sites
examined.  Given the amount of public attention devoted to
the Republican candidates and the relative importance of
those candidates in a two-party system, we determined to
limit our analysis to these candidates.  The information
contained on Web sites is not fixed.  It can change on a
weekly, daily or even hourly basis.  Accordingly, this study
examined sites over time in order to gauge the flux and
consistency of information.  Sites were visited repeatedly
during a period from February 24 to March 12 inclusive.
Visits occurred at randomly selected dates and times, with
no more than a five day gap between visits.  During the
course of this study the Republican field narrowed to three
major candidates-- Bob Dole, Patrick Buchanan and Steve
Forbes.  These constituted the purposive sample for this
investigation.  Final coding occurred on March 12, the day
Forbes withdrew from the race, beginning at 5 p.m.

     Two primary coders examined each site using the
Netscape Navigator 2.0 Web browser over a 56 K per second
Internet connection.  Netscape Navigator is one of many
programs available for browsing the World Wide Web; however,
it has several distinct advantages for those conducting
Internet research.  An estimated 82.5 percent of those on
the Internet currently use Netscape Navigator to view Web
documents (Jennings Communications, 1995).  The second most
popular browser, Mosaic, claims only 3.1 percent of the
market.  Netscape Navigator also possesses the ability to
handle certain document features, known as "Netscape
Navigator extensions," and computer applications, called
"plug-ins," that other browsers have been slow to adopt
(Netscape Communications, 1996).

     The coders appraised and printed hard copies of every
document in every section or "room" at each site; in other
words, this study evaluates all data contained at each site,
not merely a sample of that data.  While variations among
equipment and software make it difficult to render a precise
estimate of the volume of data examined, one can roughly
compare one site to another simply by counting the number of
printed 8-1/2 by 11 inch pages.  Sites were found to grow in
volume rather than change indiscriminately, as older data
were archived and retained for public inspection.  Final
analysis involved 1,679 pages divided among the candidates
as follows:  Dole, 469; Buchanan, 974; and Forbes, 236.

     The study employs an instrument designed to gauge
channel effectiveness -- the degree to which candidates take
advantage of the new medium's unique capabilities, such as
immediacy, interactivity, sourcing and multimedia.  To this
end we have created an instrument that measures 10
categories of channel effectiveness:  1) timeliness, 2)
proximity, 3) prominence, 4) source diversity, 5) message
diversity, 6) presentation diversity, 7) volume, 8)
accessibility, 9) image consistency and 10) issue
consistency.  Using the Spearman-Brown Prophesy Formula,
intercoder reliability achieved a score of near 100 percent
in categories based upon purely objective data.  [1]
However, in the areas of image consistency and issue
consistency some decrease in reliabiltiy was noted.  In the
former category, the formula yielded a reliability
coefficient of 0.953.  In the latter, the formula yielded a
coefficient of 0.958.  The high level of inter-coder
agreement was due to the existence of overwhelmingly
dominent themes permeating each site.  One assumes that
reliability would drop off in the event coders attempted to
include secondary themes.

                         Relevance

     Timeliness refers to the perceivable gap between the
moment a user views a message or document and the time when
that message was last updated.  Most basic media writing
texts recognize that recent information possesses greater
news value than dated information (Brooks, et al., 1992;
Mencher, 1996).  We can therefore assume that timely
information will generally be more relevant that dated
information.

     This study assessed timeliness on several fronts.
Coders examined each site for an "update notice," a note on
the front page of the document indicating when the site was
last updated.  If no notice was found, they looked through
the site to verify when the site was last updated by
locating dated material.  A score of 1-10 was given for each
site, with a 10 for sites that had been updated on the same
day the coders visited that site and a 0 for sites that had
never been updated.  Coders also assessed the frequency with
which dated items such as press releases were updated.  Ten
dates were selected at random from January 1, 1996 to March
12, 1996.  The coders examined each room in the site looking
for material bearing that date.  "Hits" were recorded,
resulting in a score of 0-10, which was averaged with the
overall site update to generate a timeliness score.

     Journalists and end users also place greater value upon
information that is geographically proximate to them
(Brooks, et al., 1992; Mencher, 1996).  As a British press
lord once said, "One Englishman is a story.  Ten Frenchmen
is a story.  One hundred Germans is a story.  And nothing
ever happens in Chile" (in Romano, 1986).  Geographic
proximity remains important on the Internet.  However, the
interactive nature of the new medium provides a new kind of
"virtual proximity", a feeling that the person or persons
communicating with you are somehow near.

     This study gauged the geographic proximity by examining
the amount of information aimed at audiences in individual
states and voting provinces.  Given the nature of the
election, good sites have specific areas devoted to
state-level information.  Coders looked for a
geographic-specific room at each site and measured the
largest such room in terms of how many states and voting
provinces were addressed.  A score of 0-10 was given, with a
10 referring to a site that covered all states and voting
provinces and a 1 referring to a site that had only one
state in a geographic-specific room.

     Virtual proximity is measured by an examination of
personal/interactive devices.  For example, sites that offer
electronic discussion groups or e-mail forms were determined
to possess greater virtual proximity than sites lacking such
community-oriented tools.  Coders looked for 10 specific
tools on each site, giving one point for each tool found.
[2] The resulting count of 0-10 was averaged with the
geographic score in order to represent overall proximity.

     Media professionals generally recognize that prominent
people and organizations have greater news value than others
(Brooks, et al. 1992; Mencher 1996).  It follows that the
more prominent someone is, the more relevant their message
will be to end users.  To measure the prominence of a site,
we examined those whose endorsements appeared on a
candidate's site.  "Household names," what might in legal
terms be called general purpose public figures (Associated
Press, 1994) such as Mel Gibson and Barry Goldwater were
deemed Rank #1 endorsements.  Regionally or nationally
prominent individuals/groups who did not attain Rank #1,
such as most governors or U.S. senators, were labeled Rank
#2.  All other endorsements were placed in Rank #3.

     Our instrument assigned a score to that site based upon
the number of endorsements received in each category, with
Rank #1 endorsements receiving the 10 points, Rank #2
getting 5 points and Rank #3 endorsements receiving 1 point
each.  The resulting scores were added and divided by 10.
Scores were limited to a maximum of 10, resulting in a 0-10
point scale.

     Journalists also recognize that information is more
believable if it comes from more than one source.  A greater
diversity of sources produces a "replication effect" that
increases the amount of information retained (Jamieson
1992).  We examined each site for information from 10
different source types:  news articles, think
tanks/academics, government publications or organizations,
press releases, speeches, op-ed pieces/editorial cartoons,
politicians, social figures (movie stars, sports figures,
etc.), groups/organizations and unclassified individuals.  A
source could be either a group/person cited in a press
release or a link to another Web site.  A site received one
point for each source type represented, resulting in an
overall score of 0-10.

     A diversity of messages types also increases the
likelihood that end users will deem information relevant.
Candidates need to do more than simply present a
one-dimensional image of themselves.  For example, they need
to discuss the issues, provide personal and professional
information, talk about their political philosophy, and let
people know what's happening in their campaign.

     We examined each site for a ten message types:
candidate personal information, candidate professional
information, candidate financial information, party
information, site information, issue information, a
philosophical statement, a calendar/schedule, symbolic
information and "other" information.  A site received one
point for each message type represented, resulting in a
score of 0-10.

                         Complexity

     Communication scholars have long recognized that
information with a strong visual component tends to gain and
hold one's attention.  Media design specialists recognize
this and attempt to achieve a proper mix of textual and
graphical components.  Where appropriate, they also include
audio and video.  A design that balances text and graphical
images helps the end user mediate information complexity.
The Internet is capable of handling text, graphics, photos,
audio and video.  Therefore, an effective Web site should
provide a blend of these elements to better convey
information.

     We conducted a preliminary evaluation of graphical
elements on political Web sites with two coders.  After
reviewing all available candidate Web sites, the coders
concluded that two graphics per room appeared to be the most
effective.  This usually consisted of a graphical lede
element and a iconographic menu bar at the base of the
document.  In a pending study, the researchers presented
select Web sites to 60 students in three mass communication
courses.  Preliminary analysis of the data indicates that
the subjects seemed to find the 2:1 ratio of images to pages
of text most effective.  Therefore, in this study sites that
provided a 2:1 ratio of graphical images to text, together
with some audio and video, received a higher score than
either those that relied overmuch upon raw text or those
that appear to be a mere slide show of pretty pictures.

     The volume of information contained upon a site
presents several problems to the researcher.  A plenitude of
information, by its nature, has neither a positive nor a
negative impact upon its effectiveness.  On the one hand, an
end user can be buried by megabytes of worthless data.  Then
again, one would certainly be underwhelmed by a candidate
who offered only a single 8-1/2 by 11 inch page of
information about his campaign.  This study looks at volume
as one potential component of complexity.  Rather than
attempting to measure bytes or megabytes of information,
this study simply gauges volume by the number of 8-1/2 by 11
inch pieces of paper that the site itself generates when
printed in total.  When links pointed to off-site areas such
as the Library of Congress or the Republican Party's home
page, those sites were not included in the page count.
Printed pages from each site were counted and given a score
based upon a 0-10 variable interval scale as follows.

     Accessibility mediates volume.  More data has a
positive effect only if it is well organized and easily
accessible.  A telephone book is of little value in locating
a person's number if not organized alphabetically.  This
study measures both the logical and technical accessibility
of political Web sites.  In order to achieve a high
accessibility score, a site must be organized hierarchically
with a clear and complete table of contents.  A
computer-assisted method for searching the site should be
present, preferably one that permits the end user to search
for words of his or her choosing.  The site must be linear,
but also permit users to traverse rooms in three dimensions.
It also must be open to users of limited technical ability,
with limited equipment.  Sites were awarded points based
upon the existence or absence of these features.

     Consistency helps combat complexity.  The consistent
use of specific images -- symbols, visual cues and thematic
concepts -- helps the end user make sense of a specific
site.  They can lend it form.  Advertising researchers
advocate the consistent use of certain symbols (e.g., Golden
Arches, Helping Hands) to create a positive emotional
response.  On a political Web site, image consistency
supports a candidate's effort to provide information about
himself and his campaign.  In this study, coders examined
each site in its entirety and identified recurring symbols,
visual cues and thematic or "candidate image" concepts.
Each image type was then given a score reflecting the
percentage of rooms in which it appeared, receiving a score
from 0 (none) to 10 (90-100 percent of the rooms).  The
scores were then averaged to provide an overall image
consistency score.

     In addition to consistency of image, a political
candidate must be consistent with regard to issues.  One can
address every issue that arises during a campaign, but a
wise candidate selects a few specific issues to make his
own.  For example, Bill Clinton in his 1992 Presidential
campaign chose the economy as one issue that would help
define his campaign.  Coders in this study visited every
issue-oriented room in a candidate's Web site in order to
assess the site's issue consistency.  In most cases this
included rooms containing press releases, speeches, issue
statements and philosophical statements.  Where applicable,
it also included candidate writings, newsletters and weekly
updates.  Coders identified possible issue themes running
through a site.  Each issue theme was then given a score
reflecting the percentage of rooms in which it appeared,
receiving a score from 0 (none) to 10 (90-100 percent of the
rooms) to provide an overall issue consistency score.

     Each candidate's Web site was given a score ranging
from 0-10 in each category of channel effectiveness,
resulting in a total score of 0-100.  In each case, higher
scores reflect a site that reduces channel noise (Schramm,
1955) and helps communicate information effectively.  Lower
scores indicate that the site remains an ineffective source
of potential information.

     To summarize, a political candidate's Web site will be
deemed an effective channel for communication if the site
helps make the candidate's issues relevant and clear to the
end user.  Sites help make issues relevant if they offer
timely information that is both geographically and virtually
close to that user.  The messages sent should be diverse
enough that end users will get a full picture of the
candidate, and they should come from a variety of preferable
prominent sources.  Sites reduce the complexity of
information if they provide a balanced presentation with
consistent images and issues that offers a great deal of
data in a logically and technically accessible manner.

                          Results

     The Web sites of each of the three top candidates in
the 1996 Republican primary race took advantage of the
Internet's ability to provide timely information on the
campaign.  Both the Forbes and Dole sites had been updated
in some form on the day coders performed their analysis,
while Buchanan's had been updated one day earlier (Mean =
9.67).  However, none of the sites updated their press
releases, speeches, activities or other issues on a daily
basis resulting a low update frequency score (Mean = 2.33).
Forbes received the highest score for overall Timeliness
(7.0), with Dole second (6.0) and Buchanan last (5.0).

     The candidates exhibited far greater variance with
regard to proximity.  The Dole site possessed a very strong
geographic-specific room offering a clickable map that
provided information about organizers and supporters in all
50 states, but no voting provinces (9.0).  The site also
offered a variety of interactive material, including unique
opportunities for end users to create a Dole "poster" or
take a quiz about the site (6.0).  Buchanan's site featured
a geographic-centered room listing Buchanan Brigade members
in 47 states (8.0).  The site also boasted a wide variety of
interactive possibilities, including an open board where end
users could post messages to one another (8.0).  On the
other end of the spectrum, the Forbes site failed to take
advantage of either geographic or virtual proximity.  The
site offered only one room offering geographic-specific
information for less than nine states (2.0) and just three
interactive tools (3.0).  Buchanan received the highest
score for overall proximity (8.0), with Dole second (7.5)
and Forbes last (2.5).

     As might be expected, Dole easily claimed the top score
among the various candidates with regard to prominence.  He
communicated the endorsement of one Rank #1 source (Barry
Goldwater), 27 Rank #2 sources and 21 Rank #3 sources for a
prominence score of 10.0.  Buchanan fell a distant second
with one Rank #1 endorsement (Mel Gibson), 4 Rank #2 sources
and 12 Rank #3 sources (4.2).  Forbes listed only three
endorsements:  one Rank #1 (Charles Barkley), three Rank #2
sources and one Rank #2 (2.6).

     All candidate sites performed reasonably well with
regard to source diversity.  The Buchanan site offered
information from all 10 sources coded as part of this study
(10.0).  Forbes' site left out information from think
tanks/academics and government publications (8.0), while the
Dole site failed to use think tanks/academics, news articles
and op-ed pieces/editorial cartoons (7.0).

     Interestingly, none of the candidates took full
advantage of the Web's ability to provide a variety of
message types.  All three candidates failed to disclose
financial information about themselves or their campaign.
Unlike candidates from the Libertarian or Natural Law
parties, all three Republican primary candidates also failed
to include information about the Republican party or links
to GOP sites.  Only Dole provided serious information about
the Web site itself, even to the point of including pictures
of its opening day festivities.  The Dole and Buchanan sites
received identical scores for message diversity (7.0), with
Forbes' site falling further down the scale (5.0).

     The Dole site scored extraordinarily well with regard
to presentation diversity.  With 511 graphical images, 48
photos and 244 rooms, the site maintained a graphics to text
(page) ratio of 2.29:1.  By contrast, the Buchanan site
offered 247 graphical images and 35 photos in 487 rooms for
an adequate graphics to text ratio of 1:1.73.  The Forbes
site represented the online equivalent of gray space, with
only 10 graphical images and 8 photos in 114 rooms -- a
graphics to text ratio of 1:6.  Both the Buchanan and Dole
pages offered users the opportunity to access audio and
video, while these elements were absent from the Forbes
page.  Given the preferred ratio of 2:1, Dole achieved a
perfect 10.0 in presentation diversity.  Buchanan fell
second scoring 9.0, with Forbes a distant third at 3.0.

     In terms of sheer volume, the Buchanan site predictably
achieved the highest score (10.0).  A printout of the entire
site resulted in 974 sheets.  The Dole site achieved a
respectable 469 printed pages (8.0), while Forbes again made
a relatively poor showing with 236 pages (5.0).

     Interestingly, although both the Dole and Buchanan
sites had a large volume of information, they did not
perform equally well at making that information accessible.
With a clear and complete table of contents, as well as a
limited search mechanism, the Dole site was well-organized
and easy to navigate, lacking only a choice of text-only for
users with slow Internet connections and a more robust
search mechanism.  Meanwhile, Buchanan's site lacked both
logical structure and technical accessibility.  Although the
site claimed to offer a search mechanism, the program was
"down temporarily" during the entire study period.  The
opening page was organized haphazardly, without a clear
hierarchical structure.  Although interconnected in several
ways, the site offered no sense of linearity, causing coders
to sometimes lose their way in the site.  By contrast, the
Forbes site offered a clear and complete table of contents,
as well as linear access.  However, it lacked the
interconnectedness that provides depth in a Web site, as
well as an adequate search mechanism to locate documents
quickly and easily.  As at the Buchanan site, non-Netscape
Navigator users were left looking at a site that was
sometimes confusing to their browser.  Thus, Dole received
the highest score for overall Accessibility (8.0), with
Forbes second (4.0) and Buchanan last (3.0).

     Image consistency proved to be another fascinating
category in studying the candidates' Web sites.  Given the
lack of graphical images on Forbes site, it came as little
surprise that he lacked both symbolic and thematic
consistency.  Aside from a yellow bar that added a degree
visual consistency (9.6), the site lacked a consistent
image.  There was no symbolic consistency (0.0) and limited
thematic/candidate image consistency.  The idea of Forbes as
a "political outsider with an answer" emerged, but only in
approximately one quarter of the rooms (2.5).  By contrast,
the Dole site maintained a strongly consistent image.  The
flag symbol appeared in some form in every room on the site
(10.0), together with a pictographic menu bar at the bottom
and an iconographic lede at the top (10.0).  The site also
reinforced the idea of Dole as "the Republican party's
choice" (9.1) and, to a lesser extent, as "an American
Hero."  Buchanan's site, meanwhile, achieved some symbolic
consistency with flag-like images (4.97), but only limited
success with visual consistency in his parchment-like
backgrounds (0.68).  However, the site firmly reinforced the
Buchanan's image as an "American patriot" and a "Defender of
the Unborn" (10.0).  Thus, Dole received the highest score
for overall Image Consistency (9.7), with Buchanan second
(5.2) and Forbes last (4.0).

     While the Dole site clearly puts forth a strong image
of its candidate, it lacks power in addressing the issues.
Coders identified only three recurring issues on the Dole
site:  family values, welfare reform and a balanced budget,
and only the rather nebulous issues of family values
achieved a score above 50 percent.  Buchanan's site, on the
other hand, presented several issues forcefully and
consistently.  Issues appearing in more than 50 percent of
the applicable rooms were:  pro-life, anti-NAFTA,
anti-immigration and America First/pro-worker.  Forbes,
though more limited in scope than Buchanan, was just as
consistent in presenting his issues:  a flat tax and term
limits.  In terms of issue consistency then, Buchanan and
Forbes each achieve a 10.0, with Dole straggling at 5.4.

              Conclusions and Further Research

     The goals of this study were limited, and the results
should be considered provisional.  We began by noting that
the World Wide Web is a potentially diverse, rich and
powerful source of information about the political process.
We noted that the knowledge gap hypothesis predicts that
this new medium may follow a discernible pattern where the
information rich, the technologically literate and the
politically aware gain great advantage over those who lack
the ability, means and motivation to access the new
technology.  We further argued that the first step in
assessing the presence of a knowlege gap is to ask a set of
basic questions including whether or not anything there is
worth having.  In answering this question, the particular
and largely unexplored channel characteristics of the new
medium cannot be ignored.  The limited goal of this paper,
therefore, involved the exploration of these
characteristics.

     We tried to establish 10 dimensions of what we called
channel effectiveness -- the degree to which candidates take
advantage of the new medium's unique capabilities such as
immediacy, interactivity, sourcing and multimedia.  We began
by noting that some pundits have dismissed political Web
sites as mere billboards on the information highway.
However, such a characterization fails to take into account
the unique characteristics of the Web.  We believe that
there is strong evidence, based upon our analysis, that
channel effectiveness is an important consideration in
evaluating the potential flow of information on Web sites.
Sites that offer timely information that is both
geographically and virtually close to end users help render
that information more relevant to those individuals,
especially if the information is reinforced by a variety of
messages from prominent groups and individuals.  Sites that
provide a balanced presentation with consistent images and
issues and that offer a great deal of data in a logically
and technically accessible manner help reduce the complexity
of their information.  In short, Web sites that serve as
effective channels help individuals gather, retain and
potentially make use of information.

     Channel effectiveness is a useful concept not only for
candidates trying to create a persuasive Web site and
citizens trying to participate more fully in the democratic
process, but also for journalists covering a candidate's
campaign.  In addition to gathering copies of speeches and
facts about a candidate's stance on various issues,
reporters can also use Web sites to assess a candidate's
political strategy and, perhaps most interesting of all, his
or her overall style.  This style, which comes across
through the use of slogans, key words, iconic symbols, and
culturally significant themes, represents part of what is
called the "political spectacle" (Edelman, 1973; Schmuhl,
1990).  Decoding these strategies of mass persuasion and
personal presentation remains crucial to undertanding the
candidates themselves.

     Examining the data gathered for this study yields some
engaging insights into the candidates themselves.  For
example, the Dole site exhibits many of the strengths one
might expect of the GOP frontrunner:  prominent
endorsements, effective presentation style and the strong
thematic symbols and imagery of an experienced politician.
However, it is interesting to note that the site exhibits a
notable lack of attention to the issues.  In fact, an
examination of the Dole site shows it to be much more
concerned with image than issue.  Most of Dole's press
releases deal with the various endorsements he has received
during the campaign.  His geographic-specific room lists
prominent Republican supporters throughout the United
States.  Even his issues area appears designed to reinforce
the image of Dole as an American hero and rightful heir to
the GOP nomination.

     By contrast, the Buchanan site bursts with his position
on the various issues.  However, as is often the case, the
candidate's great strength might also be his great weakness.
Addressing 40 separate issues throughout the site -- more
than twice as many as either Forbes or Dole -- Buchanan runs
the risk of diluting the issues he truly holds dear and
alienating many of his supporters.  The Buchanan site
mitigates the damage by consistently reinforcing the
candidate's campaign themes of pro-life, anti-NAFTA,
anti-immigration and America First.  He addresses both NAFTA
and the abortion issue throughout his press releases, issue
reports, speeches, and his overall philosophical statement.

     It is interesting that Buchanan brings a diverse group
of sources to support his position on the various issues,
much more so than either Dole or Forbes.  Clearly he is more
accustomed to making an argument in print than either of his
two major opponents, and this is reflected on his Web site.
His Buchanan Brigades have a strong presence on the site as
well; however, they are perhaps partially responsible for
the confusing organization of the site.  Still, by eschewing
Dole's tactic of concentrating upon prominent endorsements,
Buchanan's site retains a populist feel.

     Meanwhile, the Forbes site illustrates that candidate's
single-mindedness in pushing for a flat tax and term limits.
These issues appear consistently throughout his site, as do
others calling for reform in the political system
(anti-Clinton, anti-Dole, anti-political class, and pro-tax
reform), with one exception.  Forbes, and even more so
Buchanan, appear to long for a return to the Reagan days.

     The World Wide Web represents a new form of
communication, but this is not to say that it constitutes a
radically different form of communication.  Rather, the Web
incorporates characteristics of mass communication and
interpersonal communication.  The average Web navigator
visiting a candidate's site today accesses what is
essentially a form of mass communication or "wholesale
politics," the candidate broadcasting the same message to
millions of viewers.  However, many intriguing interpersonal
elements have already begun to appear on these sites
foreshadowing a ressurgence of "retail politics" whereby
candidates or their avatars (virtual selves) interact
directly with potential voters.  For example, the Dole site
encourages visitors to create a customized poster of the
candidate or send a personalized Dole postcard to a friend.
The Buchanan site allows users to communicate with one
another either on the site's message board or through an
electronic listserv of Buchanan supporters.  It is currently
possible, though no candidate has done this, to host an
online talk show where the candidate fields questions from
users throughout the nation.

     One can imagine a time in the not-to-distant future
when users visiting a site will be welcomed personally by a
video image of the candidate that refers to them by name and
can answer any question asked, calling upon examples that
allude to items of specific interest to that user.  For
example, asking a question about an upcoming tax hike could
draw a different response whether the end user is an
educator, an environmentalist or a member of the military.
Simply reading data from the user's own e-mail address can
already provide Webmasters with a best guess as to their
affiliations.  Considering that very little personal
information remains private today, it is not inconceivable
that a person logging into a site might soon be
cross-referenced with a database containing vast amounts of
data about their likes, dislikes, group memberships and
financial situation.

     No study is without limitation.  We recognize a certain
degree of arbitariness in measuring particular categories
such as volume and presentation diversity.  We also realize
that future research on political Web sites would do well to
include more candidates, candidates from more than one
party, and play closer attention to the accuracy of
information on various Web sites by checking data provided
by candidates against other sources.  The Internet is a
perfect laboratory for studying such questions.  Web sites
of political candidates are compact, easily accessible and
contain most of the data deemed important by those
interested in closely following a campaign.  They are, in a
sense, an intensified version of an entire campaign equally
available to reporters, scholars and the general public.


                           Notes

     [1] /pxx( = (2 ( /pxx() / (1+/pxx()

     [2] Specifically, coders looked for the existence of an
interactive listserv, a guest survey, a non-interactive
mailing list, a mailto:  protocol, a FAQ, a guestbook or
other registration form, a quiz or test, request forms,
off-site e-mail capabilities, and a other (any other sources
of interacctive material).


                         References

     Aristotle.  Attributed.  In Diogenes Laertius. 1942 ed.
Lives of Eminent Philosophers.  R.D.  Hicks, trans, vol 1,
book 5, sec. 19, p. 463.

     Associated Press.  (1994).  Stylebook and libel manual.
Reading, MA:  Addison Wesley.

     Ball, S., & Bogatz, G. A. (1970).  The first year of
Sesame Street:  An evaluation.  Princeton, NJ:  Educational
Testing Service.

     Beckel, B. (1996).  "Websites are disinformation not
much more than 30 second spots."  CNN Sunday Journal 2/11/96

     Berg, L. C. (1984).  Use of an extension computer
decision-aid program by home vegetable gardeners.
Unpublished doctoral dissertation, University of Wisconsin,
Madison.

     Brooks, B., Kennedy, G., Moen, D., and Ranly, D.
(1992).  News Reporting and Writing.  New York:  St.
Martin's Press.

     Compaine, B. M. (1986).  Information gaps:  Myth or
reality.  Telecommunications Policy, 10, 5-12.

     Delli Carpini, M. X., Keeter, S., & Kennamer, J. D.
(1994).  Effects of the news media environment on citizen
knowldege of state politics and government.  Journalism
Quarterly, 71, 443-456.

     Dervin, Brenda. 1980.  Communication gaps and
inequities.  In B. Bervin & M.J. Voigt (Eds.)., Progress in
communication sciences (Vol. 2, pp. 73-112).  Norwood, NJ
Ablex.

     Donohue, G. A., Olien, C. N., & Tichenor, P. J. (1987).
Media access and knowledge gaps.  Critical Studies in Mass
Communication, 4, 87-92.

     Donohue, G. A., Tichenor, P. J., & Olien, C. N. (1975).
Mass media and the knowledge gap:  A hypothesis
reconsidered.  Communication Research, 2, 3-23.

     Edelman, M. (1973).  The construction of the political
spectacle.  Chicago:  University of Chicago Press.

     Ettema, J. S. (1984).  Three phases in the creation of
information inequities:  An empirical assessment of a
prototype videotex system.  Journal of Broadcasting, 28,
383-395.

     Ettema, J. S., Brown, J.W., & Luepker, R. V. (1983).
Knowledge gap effects in a health information campaign.
Public Opinion Quarterly, 47, 516-517.

     Ettema, J. S., & Kline, F. G. (1977).  Deficits,
differences, and ceilings:  Contingent conditions for
understanding the knowledge gap.  Communication Research, 4,
179-202.

     Finnegan, J. R., Viswanath, K., & Loken, B. (1988).
Predictors of cardiovascular health knowledge among suburban
cable TV subscribers and non-subscribers.  Health Education
Research:  Theory & Practice, 3, 141-151.

     Galloway, J. J. (1977).  The analysis and significance
of communication effects gaps.  Communications Research, 4,
363-386.

     Gandy, O., Jr., & El Waylly, M. (1985).  The knowledge
gap and foreign affairs:  The Palestinian-Israeli conflict.
Journalism Quarterly, 62, 777-783.

     Gates, B. (1996).  The Road Ahead.  Bergenfield, NJ:
Viking.

     Gaziano, C. (1983).  Knowledge gap:  An analytical
review of media effects.  Communication Research, 10,
447-486.

     Gaziano, C. (1984).  Neighborhood newspapers, citizen
groups and public affairs knowledge gaps.  Journalism
Quarterly, 16, 556-566, 599.

     Griffin, R. (1990).  Energy in the eighties:
Education, communication and the knowledge gap.  Journalism
Quarterly, 67, 554-566.

     Hornik, R. C. (1989).  The knowledge-behavior gap.  In
C. T. Salmon (Ed.), Information campaigns:  Balancing social
values with social change (pp. 113-138) Newbury Park, CA:
Sage.

     Hyman, H. H., & Sheatsley, P. B. (1947).  Some reasons
why information campaigns fail.  Public Opinion Quarterly,
11, 412-423.

     Hyman, H. H., Wright, C. R., & Reed, J. S. (1975).  The
enduring effects of education.  Chicago:  University of
Chicago Press.

     Jamieson, K. (1992).  Dirty Politics:  Deception,
distraction and democracy.  New York:  Oxford University
Press.

     Jefferson, T. (1899a).  Letter to P.S.  DuPont de
Nemours, April 24, 1816.  In Paul L. Ford, (Ed.).  The
writings of Thomas Jefferson, Vol. 10, p. 25.

     Jefferson, T. (1899b).  Letter to Charles William
Jarvis, Sept. 28, 1820.  In Paul L. Ford, (Ed.).  The
writings of Thomas Jefferson, Vol. 10, p. 161.

     Jennings Communication Inc.  (1995).  "Browser
Statistics."  Available
http://emporium.turnpike.net/J/jc/public_html/stats.html.

     Katzman, N. (1974).  The impact of communication
technology:  Promises and prospects.  Journal of
Communication, 24(4), 47-58.

     Lakatos, I. (1968).  Criticism and the methodology of
scientific research programmes.  Proceedings of the
Aristolelian Society, 69, 149-186.

     Lang, G. E., & Lang, K. (1984).  Politics and
television revisited.  Beverly Hills, CA:  Sage.

     Loges, W. E., & Ball-Rokeach, S. J. (1993).
Dependency relations and newspaper readership.  Journalism
Quarterly, 70, 602-614.

     Madison, J. (1910 ed.)  Letter to W.T.  Berry, August
4, 1822.  Works of James Madison.  Galliard Hunt, (Ed.),
vol. 9, p. 105.

     McLeod, D. M., & Perse, E. M. (1994).  Direct and
indirect effect of socioeconomic status on public affairs
knowledge.  Journalism Quarterly, 71, 433-442.

     Mencher, M. (1996).  Basic Media Writing.  Madison, WI:
Brown & Benchmark.

     Moore, W. E., & Tumin, M.M. (1949).  Some social
functions of ignornance.  American Sociological Review, 14,
787-795.

     Netscape Communications.  (1996).  Available at
http://www.Netscape Navigator.com.

     Nielsen Media Research.  (1995).  CommerceNet/Nielsen
Internet Demographic Survey Executive Summary.  Available
http://www.nielsenmedia.com/whatsnew/execsum2.htm.

     Olien, C.N.  Donohue, G.A., & P. Tichenor.  (1983).
Structure, communication, and social power:  Evolution of
the knowledge gap hypothesis.  In Wartella & D.C.  Whitney,
(Eds.), Mass communication review yearbook 4 (pp. 455-462).
Beverly Hills:  Sage.

     Pan , Z. (1990).  Inequalities in knowledge acquisition
from mass media:  Cross generational changes and
maintenance.  Unpublished doctoral dissertation, University
of Wisconsin--Madison.

     Parker, E. B., & Dunn, D. A. (1972).  Information
technology:  Its social potential.  Science, 176, 1392-1398.

     Pitkow, J., & Kehoe, C. (1995).  GVU's 4th WWW User
Survey.  Avaialable
http://www.cc.gatech.edu/gvu/user_surveys/survey-10-1995/.

     Price, V., & Zaller J. (1993).  Who gets the news?
Alternative measures of news perceptions and their
implications for research.  Public Opinion Quarterly, 57,
133-164.

     Robinson, J. P. (1972).  Mass communication and
information diffusion.  In F. G. Kline & P. J. Tichenor
(Eds.), Current perspectives in mass communication research
(pp. 71-93).  Beverly Hills, CA:  Sage.

     Robinson, J. P., & Levy, M. R. (1986).  The main
source:  Learning from televison news.  Beverly Hills, CA:
Sage.

     Rogers, E. M. (1976).  Communication and national
development:  The passing of the dominant paradigm.
Communication Research, 3, 213-240.

     Romano, C. (1986).  The grisly truth about bare facts.
In R. K. Manoff and M. Schudson, Reading the News.  New
York:  Pantheon Books.

     Rosen, J. and Merritt, B. (1994).  Public Journalism:
Theory and Practice.  Dayton:  The Kettering Foundation.

     Scherer, C. W. (1989).  The videocassette recorder and
the information inequity.  Journal of Communication, 39(3),
94-109.

     Schramm, W. (1955).  Information theory and mass
communication.  Journalism Quarterly 32:  131-146.

     Schmuhl, R. (1990).  Statecraft and stage craft.
American political life in the age of personality.  Notre
Dame:  Notre Dame University Press.

     Sears, D., & Freedman, J. (1967).  Selective exposure
to information:  A critical review.  Public Opinion
Quarterly, 31, 194-213.

     Simmons, R. E., & Garda, E. C. (1982).  Dogmatism and
the "knowledge gap" among Brazilian mass media users.
Gazette, 30, 121-33.

     Tichenor, P. J., Donohue, G. A., & Olien, C. N. (1970).
Mass media flow and differential growth in knowledge.
Public Opinion Quarter, 34, 159-170.

     Tomita, M. R. (1989).  The role of cable television in
providing information on world news:  A test of the
knowledge gap hypothesis.  Unpublished doctoral
dissertation, University of Minnesota, Minneapolis.

     Viswanath, K. (1990).  Knowledge gap effects in a
cardiovascular disease prevention campaign:  A longitudinal
study of two community pairs.  Unpublished doctoral,
University of Minnesota, Minneapolis.

     Viswanath, K. & Finnegan, J.R.  (1996).  The knowledge
gap hypothesis:  Twenty-five years later.  Communication
Yearbook, 19, 187-227.

     Viswanath, K., Finnegan, J. R., & Kahn, E. (1993, May).
Community pluralism and knowledge gaps.  A longitudinal
study of campaign effects in three community pairs.  Paper
presented at the annual meeting of the International
Communication Association, Washington, DC.

     Viswanath, K., Finnegan, J. R., Hertog, J., Pirie, P.,
& (1994).  Community type and the diffusion of campaign
information.  Gazette, 54, 39-59.

     Viswanath, K., Kosicki, G. M., Park, E., & Fredin, E.
(1993, November).  Community ties and knowledge gaps.  Paper
presented at the annual meeting of the Midwest association
for Public Opinion Research, Chicago.

     Wade, S., & Schramm, W. L. (1969).  The mass media as
sources of public affairs, science and health knowledge.
Public Opinion Quarterly, 33, 197-209.

     Watkinson, J. D. (1990).  Useful knowledge?  Concepts,
values, and access in American education, 1776-1840.
History of Education Quarterly, 30, 351-370.

     Webber, M. (1963).  Order in diversity:  community
without propinquity.  In L. Wingo, Jr.  (Ed.); Cities and
space:  The future use of urban land (pp. 23-54).
Baltimore, MD:  Johns Hopkins University Press.

     Yahoo!  (1996).  Available http://www.yahoo.com.

     Zandpour, F., & Fellow, A. R. (1992).  Knowledge gap
effects:  Audience and media factors in alcohol-related
health communication.  Mass Communication Review, 19 (3),
34-41.

***********************************************************

Author Information:  Matthew M. Reavy
                     Manship School of Mass Communication
                     Louisiana State University
                     Baton Rouge, LA 70803
                     mreavy@unix1.sncc.lsu.edu
                     (504)388-2095/(504)388-2125 (fax)

                     David D. Perlmutter
                     Manship School of Mass Communication
                     Louisiana State University
                     Baton Rouge, LA 70803
                     dperlmu@unix1.sncc.lsu.edu
                     (504)388-2055/(504)388-2125(fax)

     This article is based upon a paper the authors
presented at the annual meeting of the Association for
Education in Journalism and Mass Communication, Anaheim,
Calif., August, 1996.
***********************************************************
                       Copyright 1997
  Communication Institute for Online Scholarship, Inc.

     This file may not be publicly distributed or reproduced
without written permission of the Communication Institute
for Online Scholarship, P.O.  Box 57, Rotterdam Jct., NY
12150 USA (phone:  518-887-2443).