_ Mun ,‘fl'n .'
www- > '
.V“»“t¢a"'{\ - my;

«3.4" n
n
l _,M.

Fl;

«“1... ,
9513*

WW
n

1‘;- ’
141*.

I

91‘}

3
l
i
g.
i

2 . 5. $35:- ;
gfiisiflfié

3%“: #3:

I;

L
g
.I

g
{3123:}

3*! .

'5‘ " 3::
I"? '15-". .55
rfizfifijgev

I: ‘3
i5 _
:5 .1

2,3352
'1'

.‘
,.
9'

my; 4.

5‘ -
'5 £351!

33

g .
mu
‘ R4

’—
’f.

'5
i

.,
'5

’r ; "'1‘ n:
5513

 

"’S’TS

0
(L)

This is to certify that the

dissertation entitled

A DUAL CODING THEORY PERSPECTIVE
ON THE EFFECTIVENESS OF MULTIMEDIA
PRESENTATION MODES ON WEB NEWS SITES

presented by

David Weinstock

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Journalism]Telecommunicaticn

Mogaawull WM

Major professor

Date 6/10/02

 

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuep65—p. 15

A DUAL CODING THEORY PERSPECTIVE ON THE EFFECTIVENESS OF
MULTIMEDIA PRESENTATION MODES ON WEB NEWS SITES

By

David Weinstock

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASS MEDIA DOCTOR OF PHILOSOPHY
College of Communication Arts and Sciences

2002

ABSTRACT

A DUAL CODING PERSPECTIVE ON THE EFFECTIVENESS OF MULTIMEDIA
PRESENTATION MODES ON WEB NEWS SITES

By

David Weinstock

This work has two significant parts. The first part evaluates dual coding
theory (DCT) as a means to examine the efficacy of online news sites. It was
found to be especially useful, owing to its assertion that memories are more
concrete when stimulated by both text and images. Using DCT as a guide, the
second portion of this work compared the efficicacy of four multimedia
presentation modes based on subjects’ recall of news items presented in a
format mimicking an online news site. The experiment successfully proved most
of the hypotheses, suggesting that online content providers who use all of the

Web’s multimedia functions will achieve the highest recall with their audiences.

Copyright by
David Weinstock
2002

This work is dedicated to the Weinstock Family--past, present, extended and
future members—who through their constant support, love, encouragement and
no small amount of nagging, made this work not only possible but also complete.
And especially to Mom, who has always been my Muse.

w .-

ACKNOWLEDGMENTS

Many thanks to the following folks for each of their individual and collective
contributions to this work: to my dissertation committee—Lucinda Davenport,
Howard Bossen, Tom Muth, Jim Lloyd, and Todd Simon—and Dad, who kept the
faith; to the Thesis Naggers: Danny, Sarah, Art, Rachel, Jon, Mike, Josh, Bee
and Dave, Nat, Jeremiah, Dad, Leif, Chris and Frances; to Ethan, Ton and
Janeen, who sacrificed more than they ever should have; to Pluta and Patty, Rick
Perry and Brian Johnson, who will always be at the core; to Nancy Creed, my
silent partner of many years; to Cathie Blumer, for subject recruitment above and
beyond the call of love, friendship or duty; to Murph, my librarian; to my two ex-
ofiicio committee members, Bob LaRose and Joe Straubhaar; to Nathan Butki,
who was right that it was only a paper; to my brother Mike, for a timely copy of
SPSS; to Andy, Kyle, Kurt and Richard forjust-in-time Web help; to David Harley,
who taught me the lesson of the snails; to Bill Cote and Curt Harler, who returned
me to journalism; to Ken Allshouse and Brad Wurfel, who kept me fishing; to Kipp
Vemer, for technical consultancy and E0; to the Clan of the Wulf, who cheered
the loudest when it was done; to Jay, Ashley, Teri and Robert, my brothers and
sisters in arms; to Connie Lawson, for being there—always; to Chuck Salmon
and David Wright, who will always be examples of the highest order of academic
integrity; to Gretchen and Melissa, who will not; to Guy Meiss, who hired me; to
Carole Eberly, who remembered me; to Tim Boudro for his gift of the right paper;
and, to all the rest of my colleagues at CMU, who took a chance on me and to all

those bums at the other universities where I interviewed who didn’t.

TABLE OF CONTENTS

List of Tables ....................................................................................... viii
List of Figures ....................................................................................... ix
List of Abbreviations .............................................................................. xi
Introduction .......................................................................................... 1
Good Theory ...................................................................................... 2
Upstart Business ................................................................................. 7
Summary ......................................................................................... 41
Dual Coding Theoretical Perspective ........................................................ 42
Introduction ...................................................................................... 43
Theoretical Framework ....................................................................... 49
Alternate Theoretical Explanations ........................................................ 68
Conclusions ..................................................................................... 76
Literature Review ................................................................................. 68
Introduction ...................................................................................... 68
Instrument Design ............................................................................. 75
Reading and Publishing Online ............................................................ 87
Recall and the News .......................................................................... 94
Effects of Content and Illustration ......................................................... 98
Visual Processing ............................................................................ 101
Research Questions ........................................................................ 104
Conclusions .................................................................................... 107

vi

Methods ........................................................................................... 1 i0

Sample .......................................................................................... 1 10
Research Instrument ........................................................................ 1 12
Viewing the News ............................................................................ 115
Recall Questionnaire ........................................................................ 1 17
Data Analysis ................................................................................. 120
Results ............................................................................................. 123
Discussion ........................................................................................ 130
Introduction .................................................................................... 130
Mixed Results............ ..................................................................... 131
Conclusions .................................................................................... 135
Conclusion ........................................................................................ 142
Introduction .................................................................................... 142
DCT: A Mass Media Research Tool ..................................................... 144
Memory Creation: News Recall .......................................................... 152
The Experiment .............................................................................. 161
What of the Web? ............................................................................ 166
Conclusion ..................................................................................... 179
Appendices ....................................................................................... 183
Bibliography ...................................................................................... 196

vii

LIST OF TABLES

Table 1: Recall Score Means and Standard Deviations .............................. 124

Table 2: Comparison of Means .............................................................. 125

viii

LIST OF FIGURES

Figure 1: Dual Coding Theory ................................................................. 59

Figure 2: Estimate of Required Sample Size Formula ................................ 111

LIST OF ABBREVIATIONS

Replace this text and begin typing your List of Abbreviations here.

INTRODUCTION

This work will attempt to serve two audiences: the news media industry—
those who might create online content for news audiences—and mass media
researchers—those who might study Web news sites in an effort to discover how
to best deliver news content to audiences.

While the goals of both are the same—to find the means to best serve the
news audience—the means by which they achieve that goal are far different.
Commercial operations, driven by profit, test various iterations of their products
against audience responses in efforts to elicit the highest response rate, the
greatest reader approval, the best indicators of reader interest. Mass media
researchers, on the other hand, use theories to explain human behavior and to
guide their research. The experiments they conduct test the validity of theories as
tools for explanation of behavior, while at the same time attempting to prove or
disprove hypotheses about the results of experiments they conduct.

This work will attempt to jointly serve these two related and yet distinct
audiences. For the news media, it will attempt to show that the best way to

present online news is to use the World Wide Web’s1 full multimedia2 capabilities

 

' The World Wide Web is a global network of computer networks that allows people all over the world to
view visual and text information at computer stations as long as they can gain access to that
network.

2 Multimedia is a term that describes the use of text, photographs or other graphic material, audio and video

files together on a single Web page to present content to audiences. In the strictest sense, the simple
combination of text and photography might be construed as multimedia. But when used to describe
content presented on the World Wide Web, it refers to content that represents a convergence of
electronic media content with print media content presented within a single context.

because content presented in this manner will engage readers’ attention
sufficiently to create the greatest recall.

For academics, this work will show that a cognitive psychological theory,
called Dual Coding Theory,3 created in the early 19703, and since used almost
exclusively within the realm of educational research, shows great potential in
evaluating this new media channel called the Web. it will deliver a lengthy
evaluation of this theory against competing theories, it will match the theory
against the practice of reading news online and it will use the theory as a
foundation for an experiment in multimedia news recall.

Since this work is purposed for an academic endeavor, the first order of
business in this chapter will be to introduce the theory and its proposed
applicability to the problem. What follows is an introductory discussion of the
channel and how media are attempting to create Web media. This chapter also
contains a short history of the development of the lntemet, email and the World
Wide Web that serves as both a technological and historical backdrop for the
work.

Good theory

Dual coding theory is a good theory.

What makes a theory good? In the social sciences, many researchers say a
good theory is one that explains a large amount of human behavior. Another way
to determine if a theory is good theory is the number of people who seem to

believe in it, how many are willing to accept it as a paradigm for prediction of

 

3 Paivio, Allan. Imagery and Verbal Processes, New York: Holt, Rinehart and Winston, 1971.

outcomes. One measure of this might be how often a particular theory is cited in
peer-reviewed journals. Academic researchers use theories as a roadmap or
guide for experimentation. Theories provide a central assumption upon which
researchers base hypotheses or predicted outcomes that will result from their
tests or experiments. In this sense, they are charts or maps whose most-traveled
routes are well documented, each twist and turn marked and tabulated to
eliminate confusion future travelers down these same paths might have. Each
new bit of research momentarily clouds the map until the researcher, using the
new knowledge their work created, adds more clarity to the map than it originally
afforded. Following this line of thinking, then, the goal of research should be the
creation of additional theoretical knowledge to further refine the “map’s” clarity.

As time passes, the world changes and good theories only get better with
time. They continue to explain as much as they did when they were first
formulated. Good theory bends and does not break in the face of changing times
and an evolving society. Instead, it grows with change to afford continuing
opportunities for knowledge creation. Good theory should seem to resemble the
very fabric of change both before and after it occurs. Dr. Steven Lacy of the
Michigan State University School of Journalism perhaps put it best when he told
an undergraduate mass media and society class the test of a good theory was
that it could be used to explain the past, analyze the present and predict the

future.4

 

4 Lacy, Stephen, Lecture, “Introduction to Mass Media,” Michigan State University, College of
Communication Arts and Sciences, Fall 1995.

Dual coding theory (DCT) is a good theory because it does each of those
things. Developed in 1971 by Allan Paivio, a Canadian cognitive psychologist, it
has been argued, challenged, tested but never dismissed as one of the major
explanations for how humans create memory from things they have experienced.
These “things” might be sights, sounds, touches, odors, or events that are
comprised of each of these single things or combinations of several or all of them
experienced together. Once our minds have collected, processed and stored
these experiences, we recall them and then associate them with other similar
experiences to construct meaning. DCT explains how we collect, process, store,
associate, re-associate and learn from these memories.

Like many theories, DCT has a number of assumptions, but its central one
is that there are two systems at work in memory: imagery systems and verbal
systems.5 We use these two systems to encode all we experience and learn into
memory—hence, the name: dual coding theory. All the things we can recall can
be categorized within these two systems, depending on how we conjure that
information in our minds. Nearly 30 years of research has shown how we use
non-verbal images to create more concrete, less abstract knowledge than we can
with verbal information alone. In addition, DCT tells us when verbal information is
accompanied by non-verbal information, learning is improved; it is improved to
the point that there is an additive effect on recall when non-verbal information is

presented with verbal information.6

 

5 Paivio, Allan. "Dual Coding Theory: Retrospect and Current Status, Canadian Journal of Psychology
45.3 (1991): 258.
6 Paivio, “Retrospect” 264-265.

DCT explains much of how we humans build our internal experiential
databases. As time passes, the means by which we acquire information
changes. Sometimes content has changed our tastes so that we are perhaps
more interested in a particular genre or medium than we were before. Or we,
ourselves, might have changed: perhaps we have experienced some kind of
hearing loss or our sight may have diminished, causing us to prefer a different
kind of sensory stimulus. But unless we have experienced some measure of
brain damage, the elemental process of thought does not change; and that is
what DCT attempts to explain. DCT argues that environment—that which we
experience (process) in the present—and memory—that which we have already
experienced, helps us create a way to construct meaning from reality.7
Therefore, it survives the first test of what a good theory is by being able to
explain a large percentage of human behaviors.

DCT is one of the major cognitive psychological learning theories in use
today. Anyone conducting a search of bibliographies and library sources would
find hundreds of references to it. It began in the late 19605 and eariy 1970s8 with
researchers matching nouns and verbs with pictures. Today, those studies
continue, as does applied research on teaching techniques using pictures, audio-
visual aids and multimedia software teaching products.9 Thus, as we consider

the amount of citations this theory accrues and the research continuum this

 

7 Paivio, “Retrospect” 258.

3 e.g., Paivio, Allan. “Paired-Associate of Learning and Free Recall of Nouns as a Function of
Concreteness, Specificity, Imagery and Meaningfulness,” Psychological Reports 20.1 (1967): 239-
245.

9 e. g., Najj at, Lawrence J. “Multimedia Information and Learning,” Journal of Educational Multimedia
and Hymedia 5 (1996):129-150.

theory has been used to review, we can see how it has withstood time’s test and
remained relevant, satisfying the last two criteria of being a good theory.

This paper will take DCT beyond its traditional boundaries in cognitive
psychology and education into mass media research. Currently, this theory has
found itself well used within the bounds of memory study,1o teaching methods,11
reading12 and classroom learning performance.13 It has seen little exposure in
mass media research. Much of that is because a large portion of mass media
research does not concern itself with the results of verbal and non-verbal news
element interaction. A great deal of attention is focused on the effects or
outcomes of combined text and graphic messages delivered to the audience14
and how content is created.15 Indeed, much of today’s literature in visual
communication evaluates impact,16 is part of some kind of content analysis or is
used to measure some societal outcome.17 Few researchers concern themselves
with the cognitive processes readers and viewers bring into play as they

read/view the news. Eye-trac18 studies,19 in which researchers tracked where and

 

‘0 e.g., Marschack, ML. and Hunt, R.R., “A Re-Examination of the Role of Imagery in Learning and
Memory,” Journal of Experimental Psychology: Learning, Memog and Cogtaition, 15 (1989): 710-
720.

” e.g., Baggett, P. and Ehrenfeucht, Andrezj. “Encoding and Retaining Information in the Visuals and
Verbals of an Educational Movie,” Educational Communication and Technology Jou_rn_al, 31 (1983):
23-32.

'2 e.g., Sadoski, Michael. “An Exploratory Study of the Relationship Between Reported Imagery and the
Comprehension and Recall of a Story,” Rflding Research Quarterly, 19.2 (1983): 110-123.

'3 e.g., Najjar, Lawrence. “Dual Coding as a Possible Explanation for the Effects of Multimedia on
Learning,” Georgia Tech Graphics Visualization & Usability Center Technical Report, 1995.

‘4 e. g. Lowe, Phillip J. “The Effect of Flashback on Children’s Understanding of Television Crime
Content,” Journal of Broadcasting and Electronic Media, 43 (1999): 83-98.

'5 e.g., Bielby, Denise B. “Whose Stories Are They? Fans Engagement with Soap Opera Narratives in

Three Sites of Fan Activity,” Journal of Broadcasting and Electronic Media, 43 (1999): 35-52.

'6 e.g., Middlestadt, Susan E. and Bamhurst, Kevin G. “The Influence of Layout on the Perceived Tone of
Newspaper Articles,” Journalism & Mass Communication Quarterly, 76.3 (1999): 264-276.

'7 e. g., Perry, Stephen D. “Inhibiting Speech Through Exemplar Distribution: Can We Predict a Spiral of

Silence?” Journal of Broadcasting and ELectrom'aMedg, 44 (2000): 268-282.
'8 Eye trac research uses low-powered lasers directed at subjects’ pupils to track eye movement.

 

how peoples’ eyes flowed across pages of print and illustration, were done and
duplicated to the point that editors and designers learned how to construct pages
to lead reader’s eyes to wherever they wished. But with the advent of a new
channel—the Web—study of the basic processes must begin again so that we
might assure ourselves that what we hold true about information processing in
traditional print and broadcast media is also true (or not) for this new medium.
Upstart business

Online publishing is an upstart business. Born in the early part of this
decade, by 1997 there were approximately 1,600 online newspapers on the
World Wide Web.20 By 2001, this number has blossomed considerably to more
than 13,000.21 Precious little research has been published about how these
products should be constructed. This paper proposes to create a new research
framework and then test the new framework in an experiment to determine which
kinds of information presentation modes—text alone, text with still photography,
text with streamed multimedia video22 or text with multimedia video and
embedded audi023—will best facilitate reader recall of the news. An experimental

study, it will feature a single issue of a fictitious online news site called the

 

'9 e. g., Stark, Pegie and Garcia, Mario. Eyes on the News, St. Petersburg: Poynter Institute of Media
Studies, (1991).

20 “Newspapers Worldwide: North America,” NewsDirectonacom. 1997,

(http://www.ecolacom/news/press/na/usl).

2' Editor & Publisher Medialnfo Links Online Media Directog, 23 Oct. 2001,
(http://www.editorandpublisher.com/editorandpublisher/business_resources/mediastats. j sp).

22 Streamed video is video that is transmitted to computer stations so that a buffer of unviewed video or
audio is created before viewing or listening begins. If network connections lag or computers slow in
the processing of that video or audio, the buffer of streamed video/audio will allow viewers to
watch/listen to content uninterrupted while either machines or the networks catch up to refill the
buffer.

23 Embedded audio refers to the practice of creating a link to a sound file and placing that link within text.

Thus, a reader might read an article and encounter a link within its text; when they click on that link,
the computer would access and play the audio file.

Lansing Eagle. This site will contain four articles—each using the different visual
modes mentioned above—to depict the news of the day for subjects to read.
Immediately following their reading, subjects will take a short test to determine
their recall of the news and answer a few questions regarding their ethnicity and
computer skills to identify potential interaction effects. The results will be
analyzed using one-way analysis of variance to determine the differences in
news recall the various presentation modes might create.

This research will use DCT as its guide. A cognitive psychological learning
theory, it predicts readers retain more concrete images from text accompanied by
illustration than by text alone. This project hopes to show that use of more
sophisticated, illustrative multimedia on online news sites will facilitate even
higher rates of recall than text accompanied by simple photographs or even text
alone.

This is a key point. In fact, it is the next logical step to launch from a 1992
study that showed recall for news text presented on computers was, in fact,
better than it was in newspapers, television and radio.24 Yet this research hopes
to prove this is not true for younger audiences who will someday step up to be
the mainstay of the news audience. If successful, this research will underscore
the importance of what all online news providers are currently using on their
sites, while at the same time, making the case that all online news sites—not just
the larger news sites hosted by larger media corporations—should take the next

step to incorporate multimedia video and embedded audio into their publications.

 

2‘ DeFleur, Melvin, Davenport, Lucinda, “Audience Recall of News Stories Presented by Newspaper,
Computer, Television and Radio,” Journalism Merly, 69.4, 1992 Winter: 1010-1022.

In doing so, they will complete the technological iteration of media convergence.
Online news sites will no longer be online newspaper, or Web TV or Web radio
sites. Rather, they will be online news sites that are perhaps owned by radio, TV
or print media enterprises but are simply Web media, in their own right. For the
Web is a channel, much like a newspaper edition or a broadcast signal. It is a
means by which information is transmitted from senders to receivers and, as
such, is a new medium unto itself.

More and more, as news media organizations acclimate to the multimedia
possibilities of the World Wide Web, the traditional boundaries between print and
broadcast media are blurring. Multiple daily deadlines are no longer the exclusive
province of electronic media. Deadlines at sites owned by print media firms and
corporations are limited only by the number of times their Webmasters choose to
add material during any given day.

Therefore, convergence—the integration of digital, audio, visual and text
information into all-purpose data networks—has created a new medium within
the WWW channel that at once merges technology into a common base and also
eliminates rigid distinctions between media systems.25 It is important to
conceptualize online news sites as a whole, rather than as online newspapers,
Web TV news or Web radio news sites. It is wrong to think of the online medium
as Web radio or Web TV. For, in truth, conceptualization like this would have

categorized news being read at a radio station as “audio newspapers” or

 

2’ Straubhaar, Joseph, and LaRose, Robert “The Changing Communications Media Environment,”
Communications Media in the Information Socieg, Belmont: Wadsworth, 1997: 40.

television news as “video newspapers,” which we know today to be irrelevant
categorizations.

We must cease categorizing online media according to the look of the
traditional media their Web designers attempt to emulate in the design of the site.
We must also not categorize online sites based upon the offline businesses of
their owners. This is not a huge leap in habitual thinking. Large media
corporations own various kinds of media outlets. For example, The Tribune
Company, Chicago, 111., owns newspapers, television and radio stations. It also
owns and operates several new media ventures. This thinking just needs to be
extended to all Web media enterprises, not just the majors. What works for
national and international news will work for local news, too. If there are local
television and radio stations within the community covering the news, and local
news media think the market will sustain an lntemet news enterprise, then this
work’s results will prove that even these operations must build partnerships in
either the print or the electronic media to assure maximum recall rates.

A good example of how convergence is already serving to eliminate
traditional media boundaries are the Web sites of two large news media
concerns, The New York Times and CNN. The New York Times Web site26
features headlines, banners, photos and cutlines, just as one might expect from
any newspaper. And yet, so, too, does CNN Interactive.27 A television news
operation, its Web site also features headlines, banners, photos and cutlines.

Consider, too, that news video clips are as likely to appear on the New York

 

26 New York Times on the Web, 1998: http://www.nytimes.com/.
27 CAN}, 1998: http://www.cnn.com/.

1O

Times on the Web as they are to appear on CNN Interactive. Therefore, based
on appearances and content, both these ostensibly different Web media have
converged in form, substance and design on the Web.

Truly, the channel has shaped the media it carries and not vice-versa.
Web developers can easily digitize news content,28 be it film, audio or text, and
make it available for access. Just as in the radio and television businesses, as
long as the audience possesses the appropriate technology, in this case, a
computer capable of lntemet access, it can consume any or all of that content.
Pages that feature too much text get very little traffic through them, no matter
how compelling the text is. Why? Because computer screens are unlike books in
one crucial way: they are a light source; people cannot focus their eyes on
material that requires too much eye contact, owing to the eye-strain it will cause,
over extended periods of time. Again, the channel acts to shape the medium.
Designers cannot impose traditional news media design imperatives on this new
and different channel and hope to be successful with audiences. Previously,
each channel carried a singular kind of content. Newsprint conveyed a
newspaper finn’s message to its audience. The newsprint allowed ink to be used
to imprint images of pictures and letters upon it to create one-way communication
from a single source to many recipients. The broadcast media, similarly create
specific signals to carry encoded images or audio signals to be decoded by

signal-specific receivers, again, to large numbers of people.

 

28 A taped interview, for example, is easily converted from an analog video signal into a digital file using a
comparatively inexpensive software program, such as Adobe Premiere. The operation involves
playing a VCR tape into a computer, then saving the analog signal as a binary file.

11

But the Web is a medium capable of distributing a wide variety of content
forms either together or separately. On the other hand, the lntemet is a channel
capable of transmitting both interpersonal and mass communication. It is
alternately capable of creating one-to-one (e.g., email) and one-to-many (e.g.,
email, Web pages) communication. Another of its characteristics is that
communication on this channel is facilitated with a tool. This makes it part of
family of communication called mediated communication. Just as television and
radio both require receivers, lntemet communication requires computers,
software and network access before communication may occur. Properly
configured, this array of equipment imparts a particular attribute upon this
mediated communication, called interactivity. Previously limited to interpersonal
communication, this characteristic is the means by which users influence the
form or content of the media experience with their own inputs into the mediated
presentation.29 It differs from the response generated from mechanical
equipment, such as vending machines, because computers make various
responses to input based on previous input users made. Mechanical equipment,
on the other hand, responds to users with the same action each time a particular
switch is activated; previous activity will, in no way, modify that response. Thus,
online media consumers can determine when they might choose to consume
their news as well as what kind of content—text, graphic or multimedia—all within
the same channel. In some cases, online news consumers can even engage in

interpersonal communication with the creators or even the subjects of the news

 

29 Steuer, J ., “Defining Virtual Reality: Dimensions Determining Telepresence,” eds: Biocca, Frank and
Levy, Mark R., Communication in the Age of Virtual Reality HillsdalezLawrence Erlbaum

12

content they have just accessed, via email or lntemet chat. The lntemet makes
all this possible. A significant number of the world's citizenry owns or works with
a computer equipped with a modem30 or a direct network connection. This
combination of technology, used in concert with software, opens a gateway to
whoever else in the world is also connected to the same network.

But it takes more than a network to create a new medium. The hardware,
the software and the ingenuity must be in place first. What is so fascinating about
the online news evolution is that its genesis was within recent memory, in the
mid-1970s. Over a five-year span, from the mid-1970s to the early 1980s,
personal computers, those that could be distinguished from the mainframes31
and mini-computers,"2 went from a dream to a reality. It all began in January
1975 when Popular Mechanics announced the availability of the Altair.33
Available only in kit-form, it had no keyboard, monitor or software and no

graphics capability. Users loaded programs by toggling switches on the front of

 

Associates 1995: 33-56.

30 The word, “modem” is actually an abbreviation for the operation that computers perform on data when it
is transmitted over a telephone line. The reason this conversion must occur is because most
telephone lines cannot transmit anything but analog signals. Therefore, the first thing that happens
to data about to be transmitted from one computer to another is that at the transmission’s
source, a modem, modulates digital information into analog information so that it can be
transmitted through telephone lines. Then, at its destination, data is demodulated from its analog
form back into its original digital form so that it can be used and interpreted by the computer
receiving the information. Thus, the “MOdulation-DEModulation” process this equipment
performs was abbreviated into the word, “modem.”

3 ‘ Mainframe computers are large computers capable of high-speed processing for multiple, simultaneous
users. These computers were originally designed to handle the computing needs of businesses and
large institutions. Their design is obsolete, though some might argue that so-called supercomputers,
used to calculate simulations and perform high-end physics and engineering calculations, are based
on mainframe designs. Today, most mainframes are used primarily for data storage.

32 A computer built between 1963 and 1987, smaller and less powerful than a mainframe, typically about
the size and shape of a chest of drawers, that was mounted on a single tall rack. Mini-computers
were characterized by limited hardware and software facilities and small physical size. Their low
cost made them suitable for a wide variety of applications, such as industrial control, where a small,
dedicated computer, which was permanently assigned to one application, was needed.

33 Straubhaar and LaRose. Communications Media 292.

13

the box. Initially, its only function was to allow users to imitate flashing light
patterns on its control panel. Subsequent models expanded its functionality. The
following year, the Altair offered video capability and users were able to purchase
a Polymorphics Systems video board that allowed them to view text and low-
resolution graphics on the same screen. Much of the video board’s memory was
spent in determining whether the individual cells on the screen would display text
or graphics. The screen had 64 cells in 16 lines; graphic cells were two wide and
three high, with a 128- X 48-pier screen resolution;34 today, most monitors
default to 640 X 480 per screen, and are capable of offering resolutions upwards
of 1280 X 1024. Also during this time, a hacker from Seattle, named Bill Gates,
then a freshman at Harvard, sold Altair a computer language to allow users to
program advanced functions. His success with this venture prompted him to
found Microsoft Corporation with a friend.35

In 1976, pre-assembled, mass-produced microcomputers were born. The
inventors, Steve Jobs and Steven Wozniak, called it the Apple.36 A year later, the
upgraded Apple II offered a keyboard, a built-in power supply and a color
monitor. In 1980, a mainframe computer company downsized its products into
desktop machines and dubbed them PCs, an acronym for "personal computer.”
At the time, most American consumers conceived of computers as room-sized
machines that computed fantastically complex calculations. All of a sudden,

anyone could own a box that fit on a desktop for about the cost of a used car and

 

3‘ Senzig, Don, CYHISTCommunig Memogy: Discussion List on the History of
CMrspace, Posted Message. 7 Nov. 1998.

35 Straubhaar and LaRose. Commanicatioaa Media 292-293.

3" Straubhaar and LaRose. Communicationertfla 293.

14

use it to write letters, play games, do taxes and a host of other everyday
activities. It was the beginning of the information revolution.

By the mid-19805, IBM and a burgeoning host of new companies
convinced enough people they wanted these new desktop models that the
personal computer industry was born. Other industries and businesses embraced
them as they offered such huge advantages over the far-more-costly mainframe
computers already on the market to say nothing of giving them more direct
control over the more sensitive operations of their companies, such as
accounting and inventory systems. These new machines brought such things as
low-cost, instantaneous communication, in the form of email; easy-to-use,
inexpensive accounting systems, so-called spreadsheet packages; and smaller
database programs offering the same data analysis capabilities previously
available only on room-sized mainframes.37

Personal computer manufacturers priced their products so that this
technology could be exported out of the office and into US. households.
Business people, already familiar with this technology in the workplace,
welcomed it into their homes. These same manufacturers also captured the
imagination of teachers, who saw personal computers as tools to make their jobs
easier, and children, who, once exposed to them, found them useful, playful and
exciting. By the late 1980s, it seemed a large part of the US. population either
had personal computers or wanted one.

Computer networks developed in a parallel fashion, but, interestingly

enough, on a much faster and earlier track than PCs did. As the Space Age

15

commenced in the late 19503 and in its wake, nuclear détente, the US.
Department of Defense became intrigued with the notion of a network of
computers that might communicate with each other, and continue to do so in the
event of a nuclear conflict. Research in the 1960s using telephone technology
yielded a 1,200 bit per second modem.38 lt communicated data well enough, but
not within a secure or interference-free channel.

Thus, in 1969, the US. Department of Defense commissioned the design
and construction of the Advanced Research Projects Agency Network
(ARPANET),39 the wide area network (WAN)40 that one day would become the
lntemet. It officially funded the project in 1973. Interestingly enough, email was
actually invented in 1971 by Bolt, Beranek and Newman, a computer network-
consulting firm, two years before the lntemet’s design was even funded.41 The
first lntemet network connected four host computers at four U.S. universities,
allowing them to share information and resources. By 1976, the network
consisted of host computers at 37 US. universities; the following year, it went
international, making connections in England and Norway.”

One of the ARPANET's distinctive features was a set of networking

43

communication rules, called a transmission protocol The first set was called

 

37 Straubhaar and LaRose. Communication Media 291-292.
38 "ARPANET and Beyond," The World-Wide Web Virtual Library: The Magnet
Studies Page, Tappendorf, Sean. 1995
<http://www.ciolek.com/WWWVLPages/QltyPages/NetStudies.html>.
3" ARPANET’s original purpose was to create a secure computer network for the Pentagon to communicate
with its university-based research teams.
4° A wide area network is a computer network that extends more than two miles in distance.
‘1 Cerf, Vinton. CYHIST Communig Memogy: Discussion List of the History of Cyberspace,
12 Dec. 1998: posted message.
‘2 Tappendorf, "ARPANET" <” http://www.ciolelccom/WWWVLPages/QltyPages/NetStudies.html>.
43Transmission protocols are sets of rules computers use when they transmit data on networks of any kind.

16

network control protocol (NCP). Implemented in 1971-72, NCP’s limitation was
that it could only communicate with other computers on the same network.44 But
ARPANET's designers knew all kinds of computer networks needed to be able to
communicate with each other to create a network that might one day circle the
globe. Therefore, in 1973, ARPA commissioned Vinton Cerf of the Stanford
Research Institute to develop a new protocol capable of intemetworking, creating
communication among different computer networks. Five years later, Cerf and
Robert Kahn of ARPA completed work on Transmission Control Protocol/lntemet
Protocol (TCP/IP), which is still in use today.45 Not only does this protocol allow
various networks to communicate with each other, it also allows users of any kind
of computer to communicate with each other on those networks. Both NCP and
TCP/IP used a process called packet switching to transmit information through
the lntemet from source to destination. “Packet switching is the breaking down of
data into datagrams or packets that are labeled to indicate the origin and the
destination of the information and the forwarding of these packets from one
computer to another computer until the information arrives at its final destination
computer. This was crucial to the realization of a computer network. If packets
are lost at any given point, the message can be resent by the originator."46 Still

in use today, it is the primary reason why email seems to be a near-

 

“ The Int_emet Socim “A Brief History of the lntemet” Leiner, Barry M., Cerf, Vinton, 0., Clark, David
D., Kahn, Robert E., Kleinrock, Leonard, Lynch, Daniel C., Postel, Jon, Roberts, Larry G. and
Wolfl‘, Stephen. 2001 <http://www.isoc.org/internet/history/brief.html#0rigins>.

‘5 W111: History of the lntemet,” Kristula, David. <http://www.davesite.com/webstation/net-history.shtml>

1997.

46 The 'Lectric Learning Web Cong ”ARPANET--The U.S. DOD-Sponsored Network," Baran, Paul. 1995

<http://www.chaos.com/learn/History.html.>

17

instantaneous means of communication. When email is sent from one host47 to
another, the message is broken into small data packets. Computers attach
labels that include the destination address to each packet and send them out into
the lntemet. These packets, instead of following a direct line route to their
destination, flow through the part of the network that, at that particular
nanosecond, has the least amount of network traffic on it. Upon reaching the
email’s destination, each packet is then reassembled into the original message,
ready for reading.48 Thus, packet switching eliminates the need for dedicated
lines to connect one computer to another. Instead, many users can share a
single high-speed transmission system with fewer channels.49

University researchers involved in defense research used ARPANET to
communicate with each other. As time passed, and they became accustomed to
this new instantaneous means of communication, the universities they worked for
saw the potential of using a similar communication network for all faculty. In
1986, the National Science Foundation “initiated the development of the
NSFNET, which, today, provides a major backbone communication service for
the lntemet.”50

At roughly the same time US. DoD launched ARPANET, the Xerox

Corporation introduced the first local area network51 (LAN) standard. Xerox called

 

‘7 A host is a computer connected to a network. It is a term used to differentiate them from other network
equipment, such as routers and printers.

‘8 Free Online Dictionaryof Computitm. “Packet Switching” 2001 <http://www.foldoc.org>.

‘9 Straubhaar and LaRose, Communications Media 304.

5° The 1mm Society. Cerf, Vinton G., "A Brief History of the lntemet and related Networks” 2001:
<http://www.isoc.org/intemet/history/cerfihtmb.

5 ' A data communications network that is geographically limited (typically to a 1 mile radius, though many
times within a single building or room) that allows easy interconnection of terminals,
microprocessors and computers. The main reason for creating LAN 3 is to allow users to share

18

it the Ethemet.52 It, too, created a network, but unlike ARPANET, it was designed
to connect fewer computers together within a more limited space. The distinction
between a LAN and a WAN is largely how much geography it covers. LANs tend
to connect computers located within space limited to single rooms or even a
single building. WANs tend to connect computers located within a larger area,
such as a city or college campus. From a technical standpoint, the difference
between a LAN and a WAN is one of transmission speed and the physical
location of data links.53

The diffusion of computer network use to home-users was very similar to
that of PC diffusion from offices to home. Companies began to discover that
computer networks allowed them to share hardware and software resources and
accomplish tasks more cost-effectively. As this technology diffused through
industry and commerce, networks became an everyday appendage of corporate
culture. By the eariy 19903, when word of the lntemet began to emerge in the
popular press, networks were a familiar concept, making home adoption of
technology necessary to access a network a relatively easy step to take,
assuming the proper technology and knowledge was in place. With network
access, workers could communicate and, with direct access to computer data
from home, extending workplace productivity into their homes.

It wasn't until 1992 that the World Wide Web evolved out of the computer

network known as the lntemet. Prior to this time, the lntemet was a network

 

resources, such as software; instead of buying individual packages for each workstation,
enterprises can buy site licenses for software that charge a reduced rate, on a per station basis, for
individual network users.

52 Ethernet is a local area networking protocol now recognized as the industry standard.

19

seeded with repositories of data existing in binary (digital)54 and text information
alone. There was no formal organization of the lntemet and hence, no means to
map where information was stored.

By and large, lntemet users in this early era simply knew where the
information they needed was and went there to get it. But as time passed and
information storage capability grew, it became more apparent there was far more
data available on the lntemet than even its most expert users could catalogue. At
that point, the lntemet changed from a shining example of modern information
storage technology to a really good example of a system falling into chaos. Users
opposed a central authority to impose order, fearing regulation would ruin the
unfettered nature of this environment. Yet they also recognized the need for
some attempt at mapping this space.

The first effort to create a map of the lntemet was embodied in a program
called Archie. Created by programmers at McGill University, it automatically
gathered the locations of stored information on the lntemetf“5 The original
program did nothing more than create a list of domains,56 where anonymous
FTP57 sites were located. FTP (file transfer protocol) sites are actually computers
that allow users access to them to store and download text or binary files
(programs or images) to their own personal computers via the lntemet. The

reason they are called "anonymous FTP sites" is because users are not required

 

53 Vemer, Kipp. personal interview 1997 Aug. 8.

5‘ Non-text computer data is referred to as binary, or digital, because when viewed in machine language, it
appears to be nothing more than 1’3 and 0’s.

5’ Free Online Dictiong of Computer Terms, "Archie," 2001 <http://www.foldoc.org>.

56 Domains are the names applied to the group of computers that store the content of a particular Web site.

57 FTP is a transmission/communication protocol that allows a user on one computer to transfer files to and
from another computer over a TCP/IP network.

20

to login with specified user names or passwords. Access is gained to these sites
simply by entering the word "anonymous” into the space for the usemame and
leaving the password space blank, though later, designers requested users enter
their email account names for use as a password. Improved versions of the
Archie program, dubbed Jughead (Jonzy's Universal Gopher Hierarchy
Excavation And Display)58 and Veronica (the Very Easy Rodent Oriented Net-
wide Index to Computerized Archives, created at the University of Nevada),59
improved the sought-after lntemet mapping by returning lists of sites along with
the title of the documents and files stored there. Subsequent versions of these
programs featured limited search functions of the FTP sites they listed. Archie's,
Jughead’s and Veronica’s chief limitation was that they merely found documents
and listed their existence.60 Once found, people used other programs, such as
complicated DOSsI-based Fl' P programs, or an easy-to—use Apple FTP tool
developed at Dartmouth University, called Fetch, to retrieve documents and
programs from FTP sites.

The next generation of Web navigation came in the form of a

client/server62 program called Gopher.63 Created at the University of Minnesota

 

58 Free Online, “Jughead,” (http://wwwfoldooorg).

’9 Enzer, Matisse. “Veronica,” Glossary of Inyernet Term_s_, (http://www.matisse.net/files/glossary.html):
2001.

60 “Stroud’s Reviews: Gopher/Archie/Finger,” Hedgedus, Peter. 2001,
<http://www.intemet.com/cws/gopher.html>.

6' DOS is the acronym used to denote “disk operating system,” the name of the system Microsofi invented
to drive personal computers.

62 A common computing system in which software is split between server (computer) tasks and client (user)
tasks. A client sends requests to a server, according to some protocol, asking for information or
action, and the server responds. This operation could be likened to a customer sending an order on
an order form to a supplier, who dispatches the goods and an invoice. The order form and invoice
are part of the ”protocol" used to communicate, in this case. There may be either one centralized
server or several in various locations. This model allows clients and servers to be placed
independently on networks and hosts.

21

as a campus information retrieval system in 1991, it did more than list locations
of documents; it actually retrieved the documents it found. Users accessed
Gopher servers located at various firms and universities around the world via
Gopher client software. These interconnected servers catalogued information
stored on computers within certain geographic regions. Gopher client software
programmers designed it to organize Gopher server information into folders
representing the file directories of various lntemet hosts, thereby making it easier
to find and retrieve lntemet information.

The same year Gopher was released, Timothy Bemers-Lee, a researcher
at a Swiss particle physics laboratory called CERN, created W3, based on a
paper he wrote in 1980,64 what would become known as the World Wide Web.
Aimed at the high-energy physics community, W3 was a networked information
system designed to allow internationally dispersed scientists to communicate
with each other and affiliated support groups. Bemers-Lee created a structure of
documents and links using a new communication protocol called hypertext
transmission protocol (HTTP). A new kind of software, called a "browser,” used
HTTP to allow users to either search document archives using a keyword or to
use their mouses to click on a link to access documents or other information on
the network.65

Ironically, it was the failure of the phone modem that caused the US.

Department of Defense to create the lntemet as a more secure alternative to

 

63 Gopher is a program designed as a “document retrieval system.” Users access servers that have lists of
documents on them with software clients.

6‘ W_3C. “A Little History of the World Wide Web,” 2000. < http://www.w3.org/History.html.>.

65 World Wide Web Consortium Web Page. ”World Wide Web-Summary," 1992

22

telephone data communication. Yet it was this same instrument that caused
consumers to flock to access the lntemet in droves. Connected, as it were, with
an inexpensive modem, personal computers allowed users to access the lntemet
and the world. Once phone modems diffused into about 10 percent of US.
households, America ”discovered" online communication. Among the first online
products confronting consumers were online services. In 1979, two of the first
information services—CompuServe lnfonnation Services (CIS) and The
Source—were born.66 Called videotex services, these businesses were a kind of
commercial library, meeting place and shopping mall, all rolled into one. Their
existence endowed people who owned personal computers with the information
and communication potential the prophets of the information revolution had
promised. Initially, access and connection fees were high. ClS charged $6 per
hour for access and targeted businesses as primary customers, hoping to attract
them with business news, stock market reports and demographic databases.
Free access to data similar to that being offered by videotex services began with
the creation of BITNET (Because It’s Time Network) in 1981.67 Launched as a
cooperative network at NYU, its first connection was with Yale University. This
network offered email, listservs and FTP. Its audience was almost entirely
academic professionals.

But as diffusion of personal computers led into consumer markets,

services like 018““8 eyed consumers as well. Competition for consumer audiences

 

<http://www.w3.org/Summary.html>.
66 My Home Page. Theys, Mitchell. 1998 <http://garage.ecn.purdue.edu/~theys/stuff7misc/comp_hist>.
67 Matisse, “BITNET,” <http://www.matisse.net/files/glossary.html>.
68 CompuServe Information Services, is an early videotex service headquartered in Columbus, Ohio.

23

grew when other services—America Online, Delphi and Prodigy—were formed in
the mid-1980s. These services differed from original videotex offerings providing
reference materials, meeting places and electronic shopping malls aimed
specifically at capturing the fancies of the American family. Each could be
accessed for monthly fees similar to those charged for cable TV subscriptions.
For a time, these businesses flourished. But it quickly became apparent the main
reason why people joined these services was to get an email account.‘59 Other
offered services languished as customers discovered many of these same
offerings could be found on the lntemet at much more competitive prices. Today,
only America Online (AOL) remains as an independent company. Much of their
traffic centers on providing access to the lntemet and interpersonal
communication via email and instant messaging rather than the host of services
offered when it entered the market.

The evolution of the lntemet into a worldwide network of computer
networks has taken about 20 years. According to the Computer Industry Almanac
Inc. there were over 400 million lntemet users worldwide at year-end 2000-up
from less than 200 million lntemet users at year-end 1998, with 134 million of
those users living in the United States. Users were defined as people aged 16
years or older who used the lntemet on a regular or occasional basis. Usage is
also high in Japan, Germany, Canada, the United Kingdom, South Korea and

China were the countries with the highest populations of lntemet users.7o

 

69 Vonder Haar, Steven. “AOL Slows Growth as Membership Chums,” Inlefialctive Week, May 20, 1996:
15.

7° Computer Ingram! Almanac. “U.S. Has 33% Share of lntemet Users Worldwide Year-End 2000
According to the Computer Industry Almanac,” 2001 <http://www.c-i-acom/>.

24

Interestingly, the Georgia Visualization Unit's 9th lntemet User Survey, the last of
these surveys to done in recent years, points to lntemet usage in Europe as
being concentrated in more youthful demographic groups because the primary
means of lntemet access is still centered in universities.71 Here in the United
States that survey indicated 62.5 percent of users were aged 16-40.72 Overall,
online gender ratios are balancing themselves closer to population non'ns.73
Data, commerce and communication stream everywhere around the
globe, at speeds we can only imagine. Growth in electronic commerce is
considerable. In 1997, online purchases totaled $191 annually, the average
income of an lntemet shopper was $65,000 per year and the average percentage
of repeat customers at lntemet sites in 1998 was 33 percent.74 According to an
Ernst & Young report released January 15, 2001, average income of online
shoppers had fallen to $52,300 per year, and the profile of an average shopper
was a 42-year-old, married female with some or no college education.75 The
combination of age and buyer demographic information bodes well for Web
media content providers. According to 1998 FlND/SVP market data,76 62 percent
of all adults who logged onto the World Wide Web clicked on an ad. Yet care
should be exercised in how they interpret this information. News content
providers need to recognize this data is extremely useful as a profile to use to

market ad space to advertisers; yet it is of limited value as a profile for a potential

 

7' GVU‘s Ninth WWW User Sm.”General Demographics Summary-Primary Place of WW Access,",
1998. <http://www.gvu.gatech.edu/user_surveys/survey-1998-04/reports/ l 998-04—General.html>.

72 Gig, 1998. <http://www.gvu.gatech.edu/user_surveys/survey-l998-O4/graphs/general/q34.htm>.

’3 FIND/SVP. “Free Data,” 1998. <http://www.cyberdialogue.com/free_data/free_data.html>.

7‘ Cyberstats Online "Profile of a Net Buyer," 1997.
<http://www.zdnetcom/zdimag/cyberstats/ 1997/06/users.html>.

’5 Ernst & Young, “Global Online Retailing: Key us. Findings,” 15 January 2001.

25

paying customer for news. The reason for this lies in the existence of so-called
”push services". These lntemet businesses derive their name from the fact that
instead of users pulling information from their sites, they deliver content without
synchronous user request. These successful services77 provide news information
purchased from a variety of sources to customers free of charge. They take their
profits from advertisers who want to market their products to an audience drawn
in by the lure of free news content served up as a screensaver. In effect, this
presentation mode devalues perceived monetary value of news content to the
consumer, who does not pay for software, the service or the news. Indeed,
today, most news firms have discovered that unless their audience consists of
subscribers to highly specialized newsletters, the only real opportunity they have
to “sell the news” on the lntemet is to charge fees to users to access news
archives.78 The normal mn of lntemet news consumer knows they need not
purchase the news when they can get it for free elsewhere.

The news industry discovered computer delivery at about the same time
videotex services entered the market. The first online newspaper service was the
Fort Worth Star-Telegram StarText bulletin board service (BBS) in 1982; it
moved to the Web in 1995.79 BBSes were the precursor to Web servers.
Computer enthusiasts loaded their computers’ hard drives with BBS software.

This software created message archives, file libraries, chat rooms and online

 

7‘ FIND/SVP, 1998. <http://www.findsvp.com>.

77 e.g., Infogate. 2001. <http://www.infogate.com>.

78 Indeed, this is the pricing practice for documents at NYTimes.com. The “price” of the day’s news is
nothing; to access documents in its archive, users are charged a fee.

79 Harper, Christopher, And That’s The Way It Will Be, New York: New York University Press, 1998:
202.

26

gaming areas on PC hard drives. Computer owners then offered other users
access to their BBSes via phone modems.8° By the beginning of 1994, about 20
newspaper online services were available worldwide; at year’s end there are
nearly 100 either operating or under development, including online editions of the
Atlanta JoumaI-Constitution and the Los Angeles Times available on the
lBM/Sears-owned Prodigy videotex service. A year later, there are 700 online
newspapers, CNN had created a 24-hour-a-day Web news site called CNN
Interactive and USA Today launched its online edition, first as a fee-based
publication, then later offered it free of charge. At the end of 1996, the number of
online news sites had stabilized at a little more than 1,600, pretty much where it
remains today.81

Despite the large amount of participation in Web news product creation,
the world’s news industry has shown neither a great deal of sophistication nor
much understanding of this new media channel. A good example would be the
rather checkered online record of Knight-Ridder. This chain was the first large
newspaper firm to recognize the potential of redistributing newspaper content on
a new channel. In 1984, it launched an experimental news delivery service called
Viewtron. This service delivered newspaper stories via cable TV for display on
television screens. Three years and $50 million later, Knight-Ridder shut the
Viewtron project down. America was not ready for a new news medium yet.82

Undaunted, Knight-Ridder opened its lnfonnation Design Laboratory in Boulder,

 

8" FOLDOC, “BBS.” 2001. <http://www.foldoc.org>.

8' Harper, The Way, 201-206.

‘2 E&P Online, Outing, Steve. “Why Web TV is Not Viewtron Revisited“ Dec. 9-10, 1996
<htm://www.mediainfo.com>.

 

27

Colorado, in 1992. Headed by Roger Fidler, its mission quickly coalesced around
a small computer known as a “video tablet.” Touted as the hand-held computer
newspaper of the next century, the device resembled a larger, heavier computer
notebook that was to have been capable of downloading news from a variety of
sources with all the readability and portability of a newsprint newspaper. Knight-
Ridder abandoned the project and closed the lab in 1995.83 Today, Knight-Ridder
has one company-wide Web site, RealCities.com, which acts as a portal to city-
based Web sites around the United States. These sites resemble search engines
in appearance, providing free email accounts, and daily, regional news produced
by its newspaper operations and local lntemet advertising.

Over the years since mass media began in mid-19th century New York
City with the penny press, to the present, the news industry has demonstrated a
mastery of going out, getting the story and telling it in a compelling enough
fashion to attract the attention of its audiences. . .until now. Knight-Ridder’s
experience mirrors that of the rest of the industry. On the lntemet, the only thing
news media have consistently demonstrated with their online products is how
prone their management is to parochial thinking. Site after site attempts to
present news content in a manner similar to that of the core newsprint product,
which categorically ignores the multimedia presentation capability the Web
channel offers. Web radio, TV and newspaper operations stream local content to
an audience that is international and cares little about local council meetings and

school board candidates.

 

83 E&P Online. Outing, Steven. “Fidler Moves to Kent State,”, 9 Sept. 1996. <http://www.mediainfo.com>.

28

Content questions aside, how much do we really know about this
medium? It is an amalgam of old and new media, but what do we really know
about how people read text, look at illustrations and process information from
Web news sites? Is it different or does it occur in the same fashion when a
person reads a newspaper or watches a television news program? How does the
addition of a computer and the lntemet affect this process?

Dual coding theory suggests the multimedia potential of the Web channel
must be used to facilitate audiences getting the most out of what they process.
But most news sites offer little in the way of moving pictures, sound and
animation, favoring the usual run of text, photography, cutlines, pulled quotes
and headlines. Web TV and radio sites offer little more than daily programming
schedules. Some have begun to offer lntemet programming, but as yet, any
advertising seems to be the province of the companies that create the client
software that delivers content to lntemet users. Not surprisingly, advertisers lead
the charge in exploring the limits and possibilities the Web and its audience
offers. Animation, video and even sound are frequent features of Web
advertisements. Indeed, lntemet advertising revenues indicate the success of
this particular kind of online content. In 1998, lntemet advertising revenue
increased twofold over 1997 levels. lntemet ad revenues outstripped outdoor
advertising revenue, posting a $1.92 billion year-end total in 1998; also quite
significant is the fact that the leading category of online advertisements was for
consumer goods. According to an independent study conducted for the lntemet

Advertising Agency by Price-Waterhouse-Coopers, “the categories [that] led

29

online spending during the 1998 fourth quarter were consumer-related (29%),
computing (20%), financial services (19%), telecom (8%) and new media (7%).
The report also found that the ovenlvhelming number of revenue transactions,
93% continue to be cash-based with barter/trade and packaged deals accounting
for 6% and 1% of total revenues respectively. Banner advertisements continue to
be reported as the predominate type of advertising, accounting for 56%, with
sponsorships (30%), interstitials (5%), email (1%) and other (8%) rounding out
the category.“4 This success has continued through the 21“ Century. Third
quarter 2001 statistics peg lntemet advertising revenues at $1.795 billion and for
the first nine months of that year at $5.55 billion.85

The news business has perhaps asked the right questions but not found
the answers to what the convergence of computers, networks and news media
bring to the table. In the face of this convergence, then, is the question, “How do
people process news found on Web sites?” and additionally, “Do the reading and
recognition patterns traditional media research has revealed still hold true for
Web media consumers?”

These are the two larger questions this work will attempt to address. It will
also evaluate which individual news presentation modes—embedded film and
audio clips and photographs—will best create what the DCT literature refers to as

an “additive, superior recall effect” when presented with text-based news stories.

 

8‘ lntemet Advertising Bur_ea_u. “Internet Advertising Revenues More Than Double in 1998,”
1998 <http://www.iab.net/>.

85 lntemet Advertising Bugaap. “IAB lntemet Advertising Revenue Report Pegs Total at $1.792 Billion For
30 2001,” 2001 <http://www.iab.net/>.

30

Summary

This dissertation opened with an introduction that briefly described the project,
presented the history of the lntemet as a technological backdrop to provide
context to how media is using this new channel to create new products. In
addition, the introduction contains its justification and a brief introduction to the
theory that will guide its construction. The introduction also contained a history of
the development of the tools necessary to building an online new industry:
personal computers, networks, the lntemet and the World Wide Web.

Chapter 1 is a theoretical review of DCT that covers its models,
constructs and assumptions and also argues why this theory is appropriate to
use as a guide to new media study. In this process, it will then examine the DCT
literature about how people read, add to their knowledge and memories and
discern meaning from information they either already have in their possession or
have just acquired. It will also discuss comparative news media recall research
and examine the state of knowledge in human-computer interfaces and how
people read and process information in computer environments versus the same
material on paper. It also will evaluate relevant conflicting theories. Lastly, it will
review educational literature that used dual coding theory to evaluate the efficacy
of multimedia classroom tools. Conclusions drawn from this literature review will
be used to create the theoretical framework that will guide an experiment
conducted as part of this dissertation to determine which news elements best aid

in news recall.

31

Chapter 2 is a literature review directed at evaluating scholarly pursuits in
areas relevant to the construction of the experiment done using DCT as its guide.
Chapters 3 and 4 outline the experiment’s methods and results.

Chapters 5 and 6 cover the discussion of the results and conclusions
drawn from the results of the experiment and the theoretical review. It is worth it
to point out to readers that approximately five years passed from the time the
experiment was performed, data was collected and computed to the time in

which the data was evaluated, and the discussion and conclusions were derived.

32

Chapter 1

DUAL CODING THEORETICAL PERSPECTIVE

Introduction

This dissertation uses dual coding theory (DCT) to guide a project to
compare recall of news items in online publications using various kinds of audio
and pictorial representation. This chapter will examine the theoretical literature
with the goal of how this cognitive psychological learning theory can be used to
create a new theoretical framework for use in evaluating the effectiveness of
online news products.

There is little mass media research that uses this theory to explain how
people remember what they read, hear or view in the news media, and what few
examples exist are atheoretical.86 To date, the greatest body of work using this
theory is in educational literature, where it is used extensively to test and
evaluate the impact of various teaching methods on memory.

Allan Paivio’s dual coding theory87 will be used as a theoretical framework
for this project. It will be used to explain how the use of pictures and, ultimately,
multimedia, in concert with text improves younger readers' abilities to form
concrete mental images about the information they process from online news.
One of the reasons this theory presents itself as a good tool for this work is its

history of usage in classroom effectiveness studies.88 A classroom is a cutout

 

8" e. g., Kenney, Keith, “Memory and Comprehension of TV News Visuals,” News Photoggipher, August
1992: 53-54.

87 Paivio, Allan. Imagery and Verbal Processes , Toronto: Holt, Rinehart & Winston, Inc. (1971).
88 The collected work of Lawrence J. Najjar, then a graduate research assistant at the Georgia Tech

33

model of a mass media channel in that its message (the lesson or lecture)
originates from a single sender (a teacher or lecturer) and is delivered to a large
or small group of receivers (students) who provide feedback on that message
(test or quiz performance).

A second reason why this theory is a good tool for this project is that its
foundation lies in experimentation designed to test recall of younger
populations.” Young, online newsreaders, for the purposes of the experimental
portion of this work, are undergraduate students who were used to populate this
study. Aside from their immediate availability in a university setting, young people
represent an important audience segment to all media. Possessing a young
audience segment implies a channel has a future, that if young people are
accessing/reading/viewing/Iistening to that channel, as they grow older, they will
continue their use of the channel.

Lacking a significant young audience segment is cause for concern
because as older audiences age and move to fixed incomes in retirement, their
ability to spend money on both acquiring media as well as the consumer goods
advertised on them diminishes. By contrast, as younger audiences age, their
buying power tends to increase, which generates a great deal of advertiser and
media interest. Older audiences are also limited by their lifespan, in terms, of
how long they will consume media products. As people age, their ability to
consume media products diminishes (e.g., eyesight and hearing weaken or fail).

This view is fostered somewhat by evidence in demographic studies of the

 

Research Institute, Atlanta, Georgia, is a good example of this research in this area. This work is
cited in Chapters 1 and 3 of this dissertation.

34

dwindling reach of newspapers to the 18- to 24-year-old audience90 and more
recently, a clip in the number of young people who watch television news.91 In
contrast, lntemet demographic studies show large increases in audience
numbers and unlike their traditional media counterparts, a growth trend.“Z A
larger question that has not been studied is: “Do these numbers, when
considered as a whole, represent dwindling and increasing interest in various
media types or do they represent a kind of chum—with audiences moving back
and forth among the various media as their needs, preferences or ages change?”
This is a question for another work.

Allan Paivio, a psychological researcher at the University of Western
Ontario, proposed DCT, a cognitive psychological Ieaming theory, a little more
than 20 years ago.93 It has stood the test of time and competing doctrines,
including, but not limited to significant theoretical challenges from supporters of
elaboration and schema theories of memory.“

Much of the mainstream theoretical research Involves testing subject
recall of text and image pairings. Effects of vowel and consonant pairs, different

kinds of imagery’s (illustrations, graphics, pictures, paintings, etc.) effects on both

 

”9 e.g., Paivio, “Paired” 239-245,
9° Tolley, B. Stuart. “Newspaper Effectiveness: The Data,” Journal of Adverti_sing Research. 31.1
Feb/March 1991: 12.
9' Pew Research Cenjg for the People and the Pres_s. “TV Viewership Falls: Fall Off Greater for Younger
Adults and Computer Users,” 18 March 2001. <http://www.people-press.org/mediarpt.html>.

Nua lntemet Surveys “How Many Online?”, 18 March 2001. <http://www.nua.com.ie>.

9’ Paivio, Imagm.

9‘ e.g., Sadoski, Mark, Paivio, Allan, and Goetz, Ernest T. ”A Critique of Schema
Theory in Reading and a Dual Coding Alternative,” Reading Research Quarterly, 26.3, Fall 1991:
463; Wilson, Brent, and Cole, Peggy, “A Critical Review of Elaboration Theory,” Educational

Technology Research and Development, 40.1, 1992: 63-69.

92

35

long- and short-tenn memory seem to make up the general run of its basic
theoretical research.95

This theory is also well accepted in educational research for studying the
use of graphic and pictorial imagery in teaching material and presentations.96
Some research using this theory evaluates uses and methods for teaching
various kinds of technical and non-technical material to various student
audiences.97 Again, researchers test both long- and short-temi recall of subject
matter in an effort to prove the general superiority of the use of illustrative
material (as opposed to text-only) material for teaching purposes.

Since multimedia has entered the classroom as an educational material
delivery tool, educational researchers have employed DCT to examine the
benefits of its use, both in the classroom and in distance education venues.98
Some of this literature applies dual coding theory to how computer imagery aids
in memory creation in the Ieaming process.99 There have also been papers
written that evaluate hypermedia from a dual-coding perspective.1°° In all cases,
this research is aimed at evaluating how these new tools either add or detract

from the classroom goal of educating students.

 

95 e. g., Kosslyn, Stephen M. and Pomerantz, James R. “Imagery, Propositions and the Form of Internal
Representations,” Cogpi'tive Psychology, 9, 1977: 52-76.
96 e.g., Najjar, Lawrence J. “A Review of the Fundamental Effects of Multimedia Information Presentation
on Learning,” Technical Report GIT-GVU-95-20, June 1995.
97 e.g., Najjar, Lawrence J. “Dual Coding as a Possible Explanation for the Effects of Multimedia on
Learning,” Technical Report GIT-GVU-95-29, September 1995.
9’ Jerram, Peter, “Who is Using Multimedia?” New Media, October 1994: 48-58.
99 Rieber, L.P., Tzeng, S., Tribble, K., & Chu, 6., “Feedback and Elaboration within a Computer-Based
Simulation: A Dual Coding Perspective,” American Educational Research Association annual
meeting. New York, New York April 1996.
'00 e.g., Najjar, "Multimedia" 129-150.

36

Mass media research can be summarized with the single question: “Who
sends/receives what message, in which channel, to what effect?” Mass media
researchers seem to be less interested in how audiences process information,
doing more results-oriented work about the effects of the messages. Surprisingly
few journalistic research projects have used DCT as a map for work.101 This is
especially surprising considering the huge body of research testing audience
recall of news information, as well as the equally large body of research that
compares recall among various media types. The mass media literature is rife
with publications applauding the prospects of online journalism, yet aside from a
few advertising researchers, no one has examined the process receivers would
employ to consume the message, despite a growing number of studies looking at
access questions from a information gap hypothesis perspective.102 In terms of
visual communications, there is a profound amount of literature on Web page
design principles?” but again, there is nothing about the audience recall process
of this material. DCT’s presence in this literature is nonexistent. It is as if mass
media researchers take the memory process for granted, assuming that memory
works the same, no matter what news or commercial messages the audience
consumes. Interestingly enough, there has been a little audience recall work
done with Web advertisements, but, owing to the newness of the media channel,

it focuses on the early stages advertising/editorial packaging and on audience

 

'0' e.g., Kenney, “Memory” 53-54.

'02 Pluck, GO, and Brown, R.G. "Curiosity is Stimulated by Incomplete Information," Proceedings of
the Britialfi’sychological Socm, 7, 1999: 5.

'03 e. g., Coyle, James, and Brill, Ann, “Examining Metaphor Use in On-Line Advertising,” Association of
Educators in Journalism and Mass Communication annual meeting, Baltimore, MD. 1998.

37

recall of advertising bundled with varying Web content.104 But like its traditional
mass media counterpart, it is more concerned with if memory occurs and how
much memory occurs rather than how memory occurs—and rightfully so. For if it
is to grow beyond a “new" channel into a viable one, its commercial success will
only be assured once participants have the answers to those questions.

There have been some projects that compare reading news on computers
to other news media channels.105 Yet there is little, if any, experimental literature
evaluating the efficacy of online news products. Such as there is has been of a
comparative nature, likening news products produced by newspaper firms to
offline, paper products or Web radio or Web TV products to their broadcast,
offline counterparts.106

These comparisons serve well enough in establishing media research for
a new channel. They serve to ground the research community with ties from
institutions they understand to those they don’t. However, for Web research to
legitimize itself, it must move past this stage of comparing one channel to
another and begin to evaluate the channel with a ruler whose purpose is to
measure individual channel efficacy.

This work proposes to do that. It will propose a theoretical framework for
evaluating online news, which is a wholly separate and different channel than its
segmented, offline counterparts. The method it proposes evaluates the channel’s

efficacy based on the Web audience’s ability to recall news messages presented

 

'04 Lindlof, Thomas R., and Schatzer, Milton J., “Media Ethnography in Virtual Space: Strategies, Limits
and Possibilities,” Journal of Broadcasting and Electronic Media, 42.2, 1998: 170-189.

'05 Lee, Seungwhan, “The Convergence of the Web and Television: Current Technological Situation and Its
Future,” Association for Education in Journalism and Mass Communication. Baltimore, MD. 1998.

38

within the channel. It does not engage in comparative recall but, instead,
evaluates Web news delivery based upon its own merits. This work will also test
this proposed framework with an experiment using DCT as a roadmap to guide
its findings to prove its usefulness as a tool in evaluating Web media
effectiveness.

Theoretical framework

Dual coding theory (DCT) is a cognitive psychological Ieaming theory that
has been around for a little more than 30 years.”7 It explains how people
remember information they have processed into memory.

DCT is based on the notion that cognition—the process of knowing things
in the broadest sense, including perception, memory and judgment‘OB—consists
of the manner in which symbolic representation systems, such as memory of
environmental information, serve functional or adaptive behavioral goals. These
systems process behavioral, perceptual and emotional information. Human
communication is unique because language, as far as we have been able to
determine, is the only communication system that is capable of creating both
verbal and non-verbal imagery from the same inputs and outputs. Thus it creates
both linguistic input and output while serving a symbolic function for non-verbal

objects, events and behaviors.109

 

'06 Lee, “Convergence.”

'07 Paivio, Imageg.

‘08 Webster’s New World Dictionag of American English, 3rd College Edition, ed.: Neufeldt, Victoria,
New York: Prentice Hall, 1994, 271.

'09 Paivio, Allan, Mental Rapresentations: A Dual Coding Approacp, New York: Oxford University Press,
1990. 53.

39

According to DCT, there are two processes the human mind uses to deal
with sensory stimuli. One, the symbolic subsystem, is specialized for handling
information derived from nonverbal objects and events; the second deals directly
with linguistic inputs. Paivio refers to the symbolic subsystem as the “imagery"
system, owing to its critical function of analyzing scenes and generating mental
images from a variety of sensory modalities, such as speech, sound, smell or
touch. He refers to the linguistic subsystem as the “verbal” system. Both the
imagery and verbal systems are structurally and functionally distinct; however,
they are not totally independent. These two systems are independent in that
either system can operate without the other, or both can operate at the same
time. Yet they are interconnected so that activity in one system can stimulate
activity in the other. Paivio asserts that verbal systems symbolically represent the
structural and functional properties of language and that non-verbal systems do
the same for the non-linguistic world.110 His theory assumes an orthogonal
relationship between symbolic systems and specific sensorimotor systems. Yet
both text and non-text systems are manifested within different modalities (e.g.,
visual systems can be triggered by printed words or visual objects).111 For
example, the smell of a lemon might evoke the word, “lemon”; the word, “lemon,”
might invoke the memory of the color, yellow, or it might trigger a memory of a
sour taste.

Thus we can begin to see DCT’s theoretical relevance to testing online

news environments. DCT assumes the existence of verbal and non-verbal

 

"0 Paivio, Allan, "Dual Coding Theory: Retrospect and Current Status,” Canadian Jour_'n_al of PsychologY.
45.3, Summer 1991: 257.

40

systems within a larger symbolic system. Online media products are constructed
similarly. Individual stories have both verbal and non-verbal subsystems that are
comprised of verbal and non-verbal elements. When considered together, these
two subsystems can either function independently to create meaning, as a
multimedia video might stand alone on a page, unaccompanied by text, to tell its
story, or it might be one of several independent but related elements to tell
several aspects of a larger story.

“Structural representations” of DCT refer to stable, long-ten'n memory
information that corresponds to verbal and non-verbal perceptually identifiable
objects and activities. “Processing” is the activity that surrounds the verbal and
imagery systems, including activation either by stimuli, called “encoding,”
activation of one by the other, called “recoding,” organization and development of
information within each system, as well as information transformation,
manipulation and retrieval from either system.112

Conceptually, the theory is organized into a hierarchical structure. At its
topmost level, DCT is about systems that serve either symbolic or
representational functions. These then are divided into verbal and imagery
symbolic subsystems, which can be expanded into sensorimotor subsystems. At
the lowest level, there are two representational units known as logogens—word
generators—and imagens, image generators. Operative cognitive mechanisms

occur at the lowest level of this hierarchy.

 

”' Pavio, “Retrospect,” 257.
“2 Paivio, Mental, 54.

41

Here again, we see a similarity in structure between the theory and the
objects it might be used to study. Online media products have a certain
hierarchical structure as well. At their topmost levels, they are collections of
stories used to create a representation of the day’s events within a specified
geographic area. These stories deliver either a verbal or non-verbal message or
both. At their lowest level, these stories consist of words or pictures or words and
pictures that, when presented together in a coherent fashion, create meaning for
those who process them.

DCT assumes mental representations are developed from perceptual,
motor and emotional experiences and that they retain those roots in memory.113
Therefore, representational structures and processes are modality-specific,
depending on the individual. For this reason then, no mental representation can
be completely abstract, nor does it deny the possibility of instinctual memories—
those programmed into the human brain after conception. For example, the
capacity to conjure images may not be uniquely human, yet it has been proven to
have a hereditary base.114 But, according to DCT, whether an individual is
inherently more comfortable processing verbal or imaged information depends
entirely on that person’s experiential dominance (i.e., whether their Ieaming more
acclimated to verbal or imagery inputs).115

Dual coding theory asserts that verbal and non-verbal systems organize

and process information in distinctly different ways. Information in the verbal

system is organized within a sequential or hierarchical structure, while non-verbal

 

”3 Paivio, Mental, 56.
”‘ Paivio, Mental, 56.

 

42

systems are organized within nested structures.116 In the verbal system, an
object would be described by arranging text-based descriptions of its individual
components into words, sentences and paragraphs until the whole was
completely described. Any new information that may be acquired is inserted into
the sequential frame, augmenting or replacing its various elements. In a non-
verbal system, there is no sequence or hierarchy of components. Rather,
because the complete object can be visualized at once, each component can be
recalled individually, but is conceptualized as part of a whole. For example, the
image of a fish includes the intuitive inclusion of scales, fins, a tail, a mouth, gills
and eyes that together comprise a unique and specific creature. These objects
can be manipulated in memory according to dimension, sensory change
(whether real or imagined) or movement. “Thus, imagens correspond to natural
objects, holistic parts of objects and natural groupings of objects. The
represented information includes, not only static appearance but dynamic and
variable properties as well.”117
Here, too, we can see how DCT theory parallels the media this project will
study. Words organized into a story must be processed sequentially or their
meaning will escape the audience. Stories that use video, embedded audio and
multimedia video as elements to accompany the written text have nested objects

that comprise a whole story that will impart meaning to the audience. When

readers/viewers/listeners recall facts from stories, they visualize the complete

 

“5 Paivio Mental, 59-61.

”6 Sadoski, M. and Paivio, A. and Goetz, E., "A Critique of Schema Theory in Reading and a Dual Coding
Alternative," Reading Research Quarterly, Fall 1999: 463-484.

“7 Paivio Mental, 60.

 

43

object in their minds, while searching their memories for information gathered
from story components. lnfonnation from each component can conceivably be
recalled, but when conceptualized together, the photos, audio tracks and text will
be remembered as a single story.

In a verbal system, each component would have to be described in detail
and in list-fashion for the mental representation of the whole creature to be
complete and accurate. Both verbal and non-verbal systems are assumed to be
functionally independent; that is, each may operate without the other or may
operate in parallel.118 When they do operate together, the theory assumes an
additive effect on recall.119

Structural lnfonnation can also be organized in a simultaneous way. 12° In
other words, when people visualize parts that make up a whole, they can also
visualize all those parts assembled and created as a whole. Consider an online
news article with audio, multimedia video and text components. From a DCT
perspective, individuals who read these articles will remember each component
of the total article in response to being asked to recall the article. But they will
first have to call up the memory of all the material contained in each of the
elements simultaneously. To do this, then, they will call up the parts available for
processing and fit them together in memory as a whole article. Logogens, too,

can vary in size, but their structure is different than imagens in that they have

components—letters—that make up parts of syllables, which are organized

 

“8 Paivio Mental, 218-222.

”9 Kobayashi, Susumu, “Theoretical Issues Concerning Superiority of Pictures Over Words and Sentences
in Memory,” Perception and Mot_or Skills, 3, 1986: 783-792.

'20 Paivio, Allan, “Imagery and Synchronic Thinking,” Canadian Psychological Review, 1975, 147-163.

 

44

sequentially into words. Taken further, writing is the sequential ordering of words
into thoughts and thoughts, when considered as parts of a whole, constitute
memory. Thus, logogens can be nested but are always sequential in nature.
Therefore, the sheer volume of sequential lnfonnation that must be recalled limits
the amount of exact verbal lnfonnation that can be recalled. Visual logogens are
a good example. Research has shown individuals can image letters and short
words, but this function is limited, in general, to words of three to four letters.121
Early DCT research had subjects make “word” associations with pictures, even
though the words were nonsensical letter combinations that were not actual
words. Thus, subjects Ieamed new word associations with familiar visual images
and allowed researchers to measure new memory creation. This research
showed that subjects visualized words in their minds, not just the images they
associated with the word combinations. As such, these visualized letter
combinations are “visual logogens”.

As stated earlier, verbal and imagery systems are functionally
independent in the sense that one system need not be active for the other to be
active or that both can be active in a parallel manner. Still, it is possible that
activity in one system can catalyze or trigger activity in the other, but it is
important to remember this activity is optional. This suggests interconnection
between the two systems. The most direct evidence of these interconnections

occurs when people make references to objects with words by pointing to named

 

'Z'Weber, D. and Hamish, J.D. “Visual Imagery for Words: The Hebb Test,” Journal of Experimental
Psychology, 54, 1974: 409-414.

45

objects (e.g., pointing to the word, “dog,” in response to seeing a picture of a
dog)

Dual coding theory specifies two kinds of interconnections between
imagens and logogens. The first is called a referential interconnection.122 It is a
between-system link that connects imagens and logogens, allowing objects to be
named and names to evoke images. In this example, most US. citizens can
easily put the correct name to a picture of President George W. Bush, when seen
on a Web news site. Similarly, when confronted with President Bush’s name in
print, these same people can easily conjure an image of George W. Bush.

The second interconnection is called an associative interconnection.123
This is a within-system link that allows logogens to be associated with other
logogens and imagens to be mentally associated with other imagens. Both verbal
and imagery systems are bi—directionally interconnected; interconnection is
assumed to be one-to-many.124 For example, an image is capable of evoking any
number of different verbal or non-verbal representations. Which images or verbal
descriptions will be activated will vary according to individual experience. Mass
media does not leave this process to chance, however. Recognizing that bi-
directional interconnection exists, text and visual components of news are
seldom presented alone to the audience. Film has voiceovers, text is usually
presented together with illustration or photography and that photography is

usually accompanied by text outlines to further explain what readers/viewers see.

 

‘22 Paivio Mental, 63.
'23 Paivio Mental, 69-70.
124 Paivio “Retrospect,” 259.

 

 

46

Associative processing in the verbal system is assumed to be hierarchical.
Thus the word, “fish,” would be a category label for such words as “trout” or
“bass”. These hierarchies are based in the cultural organization of our language
structures. Media depend heavily upon cultural organization of language
structure to deliver messages to their audiences. This is especially true for those
media that attempt to appeal to some kind of localized audience. “Speaking the
same language” is very important to local media in terms of reaching their
audiences. This point does not serve to make the study of Web media products
difficult with this theory. While it is true Web media attempt to appeal to mass
audiences, the Web’s audience is, in general, younger, hipper, has higher
income levels and is more educated. Therefore, cultural organization of language
structure for Web audiences is just as critical as it might be for more traditional,
localized media, despite the global geographic diffusion of the audience.

Thus far, this theoretical discussion has primarily dealt with how memory,
or as Paivio refers to it, mental representation, is created. Another portion of this
process specific to DCT is how mental representations are activated. According
to the theory, activation of both verbal and imagery representations is a joint
function of stimuli and individual differences.125 Put more simply, memory is
activated jointly by the conditions that call for memory to be activated and the
individual’s experiences (affective, behavioral and sensorimotor).

Two types of stimuli are part of this process. One is called target stimuli,
which can be something presented for processing, such as various online news

articles with different levels of multimedia in them. The second is contextual

47

stimuli—those stimuli that constitute an instruction set within which subjects put
target stimuli to use to accomplish specified goals.126 An example here might be
a recall test on the editorial material contained in articles that experimental
subjects just finished reading. Or it might be a business decision that calls for
knowledge of the day’s news.

Experimental data support the theoretical assumption that the activation of
nonverbal representations in imagery processing creates more concrete mental
representations than verbal stimuli.127 Similarly, the use of more concrete verbal
stimuli is more likely to create concrete mental representations than the use of
abstract verbal stimuli.128 DCT suggests, then that verbal representation
activation is maximized when the words that serve as stimuli have a great
capacity to arouse verbal associations, or if the task demands verbal processing
and instructions to perform the task are delivered verbally to the subjects.129

Still, DCT theorists recognize individual differences will also play a role in
representational activation. Different people have different experiences or even
different preferences for thinking verbally or visually. These differences are
generally expressed in probabilistic tenns—that they will influence the probability
of how well verbal and nonverbal representations will be activated and put to use

for the proposed task.

 

"5 Paivio Mental, 67-69.

‘26 Paivio Mental, 67-68.

‘27 op. cit., Paivio, Allan, “Coding Distinctions and Repetition Effects in Memory,” ed: Bower, G.H. _T_li_e
Psychology of Learning and Motivatjg, New York: Academic Press:, 1975.

‘28 Paivio “Retrospect,” 261.

'29 Paivio Mental, 68.

 

48

DCT describes three kinds of processing people use to access mental

representations using these two interconnections (See Figure 1).

Chart 1: DUAL CODING THEORY

 

 

S BDSOI'VStIITUII

 

 

 

 

 

(, Representational Processing v

 

 

Referential

 

V

 

 

 

 

Verbal
Systems

Processing

 

 

 

 

Associative 1

Processing Non-Verbal
Systems

The first, representational processing,130 refers to direct activation of

logogens by verbal stimuli and imagens by non-verbal stimuli. Researchers have

shown activation using referential processing takes less time and is more direct

than the two other kinds of processing, referential and associative processing.

Referential processing131 is the processing required for the cross-system

activation in imaging to words or conversely, how a word might evoke an image.

From a theoretical standpoint, this kind of processing is believed to be more

indirect because it involves a mental crossover from the verbal to imagery

systems or vice versa. Paivio notes one reason for this is that image generation

 

‘30 Paivio Mental, 70.

49

and object naming tasks often involve other kinds of processing as well as use of
both verbal and imagery systems. Associative processing132 consists of the
within-system activation of one Iogogen by another or of one imagen by another.
All cognitive tasks require representational processing, but some may involve all
three?33

At this time, it is appropriate to note that one function of mental
representation is that of active organization or reorganization. As we assimilate
new lnfonnation, we must reconcile the new with the old. Theoretically, DCT
suggests that memory is dichotomized along verbal and imagery lines and so,
too, is memory organization?34 Thus, in reorganization, the verbal system
rearranges representational processing according to a sequential process while
imagery systems generate synchronous and spatial structures in memory?35

Over time, research has shown that both imagery and verbal systems
have important functions in encoding, storage and retrieval of new information.
Paivio asserts the evaluative function of mental representation is rooted in the
verbal system when we engage in an attempt to assign some relative or absolute
value to information?“6 The reason, he says, is because we give ourselves
verbal, unspoken or unconscious instructions to trigger the evaluative process,
which is ultimately quantitative?37 Even when we evaluate the pros and cons of

abstract concepts, we still would place one concept in a superior position and the

 

'3' Paivio Mental, 69.

”2 Paivio Mental, 69.

133 Paivio "Retrospect," 259-260.
‘3‘ Paivio Mental, 141.

'35 Paivio Mental, 142.

'36 Paivio Mental, 74-75.

'37 Paivio Mental, 74.

 

5O

other in an inferior rank position. In effect, we have sequenced this information,
which, as noted earlier, is a function tied to the verbal system.

Another function served by representational systems is the mnemonic
function?“8 This function serves to explain how imagery and verbal systems aid
in creating memory. One assumption that governs this part of DCT is that the
pathway used to create a particular memory is modality specific?39 In other
words, if the memory was encoded verbally, it will be retrieved verbally and
similarly, if it was encoded by means of the image system, it will be recalled as
an image. Representational coding—words evoking word memories, images
evoking imaged memories—is usually sufficient for some level of accuracy in
recognition and recall tests. Referential encoding is also possible, as long as the
contexts in which the input items are presented are not confusing. For example,
viewing a multimedia news video on an online news Web site would be a
contextually appropriate way to view news material. Therefore, it would be
possible that subjects would be able to referentially encode information gathered
by reading/viewing news material on a Web site. DCT also notes that the
probability for successful referential encoding will be much higher if subjects are
given instructions as to the nature of the tasks prior to processing the
information?40 Objects, such as sliders, radio buttons and links fulfill this function
on Web pages. Since experienced Web users intuitively know how to use these
devices, their function serves as instruction on how to access material and also

provides clues as to the nature of the material they will access in using these

 

'38 Paivio Mental, 71.
'39 Paivio Mental, 149.

51

A!

‘1“ .....—- 3“!" '#

features prior to processing this information. For example, a slider button
underneath a framed visual image cues viewers they are confronted with a video
frame on a Web page. It tells them the information they are about to process is
moving and not a still photo and will likely, but not always be accompanied by a
sound track. Thus, the user is “instructed” to receive lnfonnation in a particular
way by the frame surrounding the text or images they are about to process.
Associative encoding is the activation of associated representations in the same
symbolic modality as the input item?41 Thus, a subject who recalled the subject
matter in a video as a mental representation of a film clip in which news was
imparted would have engaged in associative encoding to put that lnfonnation into
memory.

Mnemonic retrieval involves recoding, “jogging” a subject’s memory with
material related to the information input.142 An example of this might be an
objective question in a news recall test that sought a particular answer to a
particular event related in a news story via text, film or embedded audio. This
question would be subtly worded to remind viewers of the subject matter without
answering the question. Recoding during retrieval can occur if the query material
is similar in structure to the material when it was initially encountered.

But, following the theory, verbal representations make it more difficult to
recall verbal information. Subsequent research used to develop the theory has

shown that pictures are easier to remember than words due ”entirely to the

 

"0 Paivio Mental, 122-124.
'4' Paivio Mental, 76.
“2 Paivio Mental, 236.

52

mnemonic superiority of image code over the verbal code."143 As stated earlier,
the two codes are independent of each other but research indicates pictures are
additively encoded?44 In an experiment in which participants were first asked to
name or generate Images for pictures and for words, then unexpectedly asked to
recall the words or names of pictures, imaging to words doubled recall relative to
imaging for words that had only been pronounced.145 In addition, imaged words
were recalled about as well as named pictures. In another experiment, Paivio
used a repetition-lag model in which he again demonstrated non-verbal systems’
mnemonic strength over verbal systems. The lag effect occurs when recall of
repeated items increase with the number of other items that intervene between
repeated items?“6 Interestingly, zero-lag items, those that are successively
repeated, are "less than additive."147 But overall, in terms of mnemonic value,
Paivio notes that empirical evidence shows imagery codes to have a 2:1
advantage over verbal codes.148 The sum of experimental research shows that
imagery surpasses even concreteness as a predictor of recall?49
Pictures and words contribute unequally to recall with pictures

contributing approximately twice as much to their additive effect as words. This

advantage occurred in both massed repetitions and once-presented item pairs?s°

 

"3 Paivio "Retrospect," 265.

"4 Paivio, Allan and Csapo, K. "Picture Superiority in Free Recall: Imagery or Dual Coding?" Journal of
Cognitive Psychology, 1973: 176-206.

”5 This finding might well be tested again with recall of complementary embedded audio in online news
text.

”6 Op. cit., Paivio, "Retrospect,": Paivio, Allan, "Coding Distinctions and Repetition Effects in Memory,"
ed.: Bower, G.H., The Psychology of Learning and Motivation, New York: Academic Press, 1975.

”7 Paivio "Retrospect," 265.

"8 Paivio Mental, 76.

“9 Paivio Mental, 159.

'50 Paivio, Allan, “Spacing of Repetitions in the Incidental and Intentional Free Recall of Pictures and

53

This lag phenomenon was once identified but then mislabeled in mass
media research in a study on news audience recall literature. A researcher
performed a study testing audience recall of television news, in which he
incorrectly rejected dual coding theory in favor of a single-coding hypothesis?51
The reason, it seems, is that he was unaware of the lag phenomenon mentioned
above. His experiment showed how words are remembered as pictures and
pictures are remembered as words, which is consistent with referential
processing. He called this a “translation phenomenon,” but noted that it was
more likely to occur 48 hours after viewing a news story than immediately, which
also supports the lag phenomenon reported above and not a single-coding
hypothesis that suggests all messages are stored the same in memory.

As seen above, there is evidence of limited usage of dual coding theory in
mass media study. One reason might be that it is simply overlooked. Another
reason might be that, thus far, researchers have merely been concerned with the
end results—what actual recall was and how it compared to that for material in
other media—and less so with how it is stored in memory. Because publishers

b,152 researchers must

are now placing news products on the World Wide We
now concern themselves with discovering if visual communications design
principles that work well in the print and broadcast media will also work well

within the computer medium.

 

Words,” Journal of Verbal Learning and Verbal Behavior, 13, 1974: 497-511; op. cit., Paivio, Allan,
“Coding Distinctions and Repetition Effects in Memory,” Bower, G.H., ed., The Psychology and
Motivation, New York: Academic Press, 1975.

'5 ' Grimes, Tom, "Encoding TV News Messages into Memory," Journalism Quarterly, 67.4, Winter

1990: 757-766.

‘52 NYTimes.com, CNN.com, LATimes.com, MSN-NBC.com to name a few.

54

Psychologists divide memory into as many as six different categories.
DCT eschews this discussion in favor of addressing the difference between two
of these categories—those being semantic153 and episodic memory‘s“. Semantic
memory is the part of memory in which lnfonnation is manipulated, transformed
and compared to create meaning?55 Episodic memory is memory about specific
events at a particular time and place?56 Both semantic and episodic memories
are relevant to the study of news recall. News, in and of itself, is episodic in
nature; it is event-driven and as such, for those who do not witness the events
they depict, reading the news becomes the event for them. This is especially tme
for news presented on the Web. With multimedia capabilities,
readers/viewers/listeners may actually witness, hear about and view events all at
the same time. Semantic memory becomes important once news has been
experienced. Once mental representations are processed, some measure of
semantic processing occurs so that we might create meaning from what we have
just processed. As we process what we Ieamed from the news, we relate the
newly acquired facts to what is already stored in memory to create meaning and
meaningful memories.

DCT approaches semantic memory from a holistic perspective. As
mentioned earlier, the process recognizes that its objects are made of
connected, sequenced or nested units. Thus, the end of the process is not the

representation of the parts, but rather the representation of the whole. Paivio

 

‘53 Paivio Mental, 120-139.
'5‘ Paivio Mental, 140-176.
"5 Paivio Mental, 120-121.
'56 Tulving, Endel,. “Episodic and Semantic Memory,” Eds: Endel Tulving and Wayne Donaldson,

 

 

55

notes that DCT does not distinguish any structural or functional differences
between episodic and semantic memory?57 This, too, bodes well for study of a
channel that presents multimedia lnfonnation that must be recalled from both
sequenced and nested memory structures. Recall then must be measured
against memory of the whole rather than memory of lnfonnation imparted by
specific components.

The theory refers to the meaning creation process as memory trace?58 It
assumes memory creation processing is also modality specific and is derived
from two separate kinds of sources: external and internal?59 We derive external
sources directly from sensory stimulation and we derive internal sources,
associatively, from our experience. Paivio thinks external sources are more
reliable than internal sources because external sources are directly derived from
sensory stimuli?60 He reasons extemal memory sources are derived from
specifically recalled events while internal sources’ accuracy may be difficult to
assess “in the absence of ...objective (criteria).”161

The theory also has assumptions about how encoding storage is
organized. Essentially, it suggests encoding follows a dual-track, modal-specific
path (or trace)?62 The image system, as stated earlier, is best suited to encoding
spatial and synchronous lnfonnation while the verbal system is best suited for

sequential lnfonnation.

 

Organization of Memory. New York: Academic Press, 1972: 58-74.
"7 Paivio Mental, 14o.
"8 Paivio Mental, 76.
'59 Paivio Mental, 141.
'°° Paivio Mental, 141.
'6' Paivio Mental, 141.
'62 Paivio Mental, 143.

 

56

lea-m I

But it is important to note that while some information will be encoded into
memory by the verbal system, other information will be encoded via the image
system and that still other information will be encoded by both the verbal and
image systems. Here again, it is important to note that the theory assumes and
experimentation has proven that memory items that are encoded dually will be
remembered better than those encoded singly?63

In addition to following the general structural assumptions of DCT,
encoding also follows the model’s processing assumptions. Representational
encoding, for example, suggests information destined for memory creates
perceptual memory traces in the same manner in which they are received?64
Printed words, in part, create verbal traces, filmed events produce visual traces
and audio tracks create auditory traces. Additionally, these traces activate others
like them or related to them from within semantic memory. Completely new
information, lnfonnation that is totally outside an individual’s experience, is
encoded according to its input mode but rarely elaborated significantly.

Referential encoding involves memory items that are derived from verbal
and imagery memory traces. Interestingly, printed words are usually encoded, or,
more accurately, recoded into an auditory-motor form called the phonemic
form?65 In other words, memory derived from printed words comes not only from
the verbal structure of the input item, but also from the auditory structure of the
sound of the word reverberating within a person’s consciousness. Thus, memory

trace gathered from reading is referential encoding, involving both imagery and

 

“’3 Paivio, “Spacing,” 497-51 1.
‘6‘ Paivio Mental, 144.

 

57

verbal systems together. Taken further, than multimedia lnfonnation
viewed/readlheard on Web pages will use both systems as well, no matter if the
information appears in the form of printed words or multimedia video.

Imagery and verbal lnfonnation presented together are remembered
together, even if there is no prior association between the images and the verbal
information. Thus, an online news story about the Russian premier’s visit to Lima,
Peru, that presents lnfonnation in verbal, audio and video forms will be
referentially encoded from both the verbal and imagery forms in memory
together, even if the two countries were never associated together in memory
before.

Associative encoding involves the elaborative processing mentioned
earlier within either verbal or imagery processing systems. It can be based on the
material in its original form or it can be based upon previous knowledge retrieved
from semantic memory?66

To sum up encoding then, episodic memory performance depends on the
type of code involved—verbal or imagery—the number of codes or traces that
are involved in creating the memory and the number of relations individuals
generate between trace components.

Alternate Theoretical Explanations
There are a large number of psychological theories used to explain how people
learn information from text, audio, video and even multimedia materials. This

next section will identify a number of these theories, provide brief explanations of

 

'65 Paivio Mental, 143.
'66 Paivio Mental, 144-146.

58

 

them and discuss briefly why they are not as well suited to the task of explaining
how people process news from online news sites as DCT is.

There are also a large number of psychological theories used to explain
memory and Ieaming. Many of these presume the existence of prior experience
and suggest various ways in which what is being Ieamed can be compared to
what is in experience. Essentially, they attempt to explain recall based on the
audience’s receptivity to the message. Cognitive dissonance theory is a good
example of this kind of theory. This theory assumes people seek a consistency
among their beliefs and opinions.167 When inconsistency is encountered between
attitudes and behaviors, dissonance occurs and individuals attempt to create
consistency?68 Dissonance usually occurs when individuals must choose
between two incompatible beliefs or actions; the greatest dissonance usually
occurs when two alternatives appear equally attractive. The theory predicts
attitudes will be changed to accommodate behavior?69 The number of beliefs an
individual has and the importance attached to each belief determines the
strength of the dissonance. Thus, it is easy to see that cognitive dissonance
theory, and others like it, would not be useful in research like this because they
concern themselves more with persuasion than they do with cognitive function.

Experimental subjects in this work are viewing news presented within

several presentation modes from the same source,170 processing various kinds of

 

'67 DeFleur, Melvin L. and Ball-Rokeach, Sandra, “Theoretical Strategies for Persuasion,“ Theories of
MassCommunication, (Longman: New York, NY), 1989: 276-277.

'68 DeFleur Theories, 277.

'69 DeFleur Theories, 277.

”0 For example, a single news story might be delivered via text, embedded audio and multimedia video, all
presenting complementary information about the same subject matter from the same Web page.

59

lnfonnation through various sensory systems. Many of alternative cognitive
psychology theories fail to account for these different channels, though some
suggest subjects have media preferences or predispositions that enable them to
process information more efficiently. Aptitude-Treatment Interaction Theory is
one example. This theory suggests certain media are more or less effective for
particular individuals, depending upon their specific abilities?71 While this theory
sounds good, researchers in this area have found aptitude interactions to be
difficult to clearly demonstrate?72 Another problem they quote is that various
Ieaming styles can be linked to individual and aptitude variables but they also
tend to vary within individuals, depending on the task, which makes them very
difficult to operationalize?”

A number of cognitive psychological Ieaming theories have been created
to explain problem-solving outcomes. Anchored Instruction Theory is a good
representative theory in this genre. Similar to grounded theory in mass media, it
suggests Ieaming and teaching activities should be designed around an “anchor"
that is based on a real-life or case-study situations. Learners, without
supervision, then explore this material.174 This theory is used extensively in

research that evaluates videodisc and multimedia classroom equipment?75 In

 

171 Cronbach, L. and Snow, R., “Aptitude-Treatment Interaction,” The Theory in chtice Database, 1999,
<http://tip.psychology.org>.

172 Snow, R., “Aptitude-Treatment Interaction as a Framework for Research on Individual Differences in
Learning,” eds: Ackerman, P., Stemberg, R.J., Glaser, R. Learning and Individuflifferences
New York: W.H. Freeman, 1989.

”3 “Framework,” p. 5 l.

174 Bransford, John D., and Stein, Barry S., The Ideal Problem Solver: A Guide to Improvipg Thinfigg,
Learning and Creativig, New York: W.H. Freeman, 1984.

”5 e.g., Glaser, Candyce Williams and Prestidge, Linda K., “Using Technology to Support Special
Education Teachers’ Implementation of the Theory of Anchored Instruction,” American Educational
Research Association Annual Meeting, Washington, DC, 21 April 1995.

 

6O

this case, news might reasonably be considered anchored material and certainly
audiences do consume news without supervision. But as mentioned above, news
is not material that is customarily used in problem solving situations.

Gestalt theory offers some interesting possibilities for those who would
study Web page design. At the center of gestalt theory is the notion of grouping
stimulus characteristics that cause individuals to structure or interpret what they
see in a certain way.176 The theory describes four primary factors—proximity,
similarity, closure and simplicity—as the chief determinants of grouping?77
Proximity states elements are grouped according to their neamess; similarity,
that those with similar attributes are grouped together; closure, items that seem
to be parts of the whole of a particular entity; and simplicity, items are organized
according to symmetry, regularity, and smoothness?78 Gestalt’s failing, however,
is that it speaks only to perception and again, problem solving. It, too, falls short
in terms of providing explanations of how items are implanted into memory and
why visual ones seem to be more memorable than text.

Another theory that also shows promise for would-be design theorists is
lnfonnation pickup theory. Similar to gestalt, it suggests that perception depends
entirely on such factors as terrain and its variables—shadows, texture, color,
convergence, symmetry and layout—that determine what individuals perceive. "9

Unlike other cognitive theories, it discounts previous experience as a cognitive

 

'76 Kearsley, Greg, “Gestalt Theory,” The Theory in Practice Dam, 2001.
<http://tip.psychology.org>.

177 Kearsley, “Gestalt.”

”8 Wertheimer, Max, Productive Thinking, New York: Harper & Row, 1959. 8-13.

'79 Kearsley, Greg, “Information Pickup Theory,” The Theory in Practice Database, The Encyclopedia of
Psychology Society. 2001. <http://tip.psychology.org>.

 

61

if
m

function and instead declares perception to result directly from environmental
stimulus. 18° lnfonnation pickup suffers the same shortcoming gestalt theory
does. It describes only perception and does not delve deeply enough into
memory function.

Level of processing theoretical framework suggests that the more
processing a person performs on lnfonnation, the more it will be retained and
remembered. Information that is processed more deeply is more likely to be
remembered. Information that is more likely to be processed deeply includes
strong visual images or information that has an association with previous
memories. Processing is unconscious and automatic unless we pay attention to it
directly. 18‘ Thus this theory suggests strong visual material and material with
previous associations in memory facilitate memory processing.

Yet video and photographic material is not always strong; many times,
accompanying video and photos are merely supportive. In addition, news, by its
very nature is new information. Being able to make a connection between what is
being read and what is in memory may not always occur. Therefore, the level of
processing framework fails to fully explain recall of various multimedia news
presentation modes in online news stories.

Elaboration theory very nearly contends with DCT to explain how
readers/viewers/listeners might retain information from online news stories. It

proposes seven major presentation strategy components: (1) an elaborative

 

'80 Gibson, James Jerome, The Senaes Considered as Perceptual Systema, (Boston: Houghton Mifflin),
1966: 30-35.

'81 Craik, F. and Lockhart, R. “Levels of Processingon a Variety of Memory Tasks: New Findings and
Theoretical Implications,” Consciousness and Cogition, 5.1, 1996: 142.

62

sequence, (2) prerequisite sequences, (3) summary, (4) synthesis, (5) analogies,
(6) cognitive strategies, and (7) reader control.182 The elaborative sequence is
defined as a simple or complex sequence in which the first presentation
highlights the ideas and skills that follow. Elaborative sequencing should be
accomplished with a single type of content, although two or more types may be
elaborated simultaneously, and should present only a few ideas or skills?83
Headlines, decks and even lead sentences perform this same precise function.
Within a single content type, they present lnfonnation in summary or synthesis
fashion immediately prior to readers processing more text or multimedia that
elaborates on the initial presentations.

Elaboration theory suggests that material presented in an order of
increasing complexity is the best way to maximize recall.184 For example, when
presenting information, the simplest version of the lnfonnation should be
presented first. Subsequent presentation should provide additional versions of
the lnfonnation until the full range of information is presented. With each
presentation, the theory suggests, audience members should be reminded of all
versions presented thus far in summary or synthesis form?85

Ultimately, this theory is tailored to classroom presentations far better than
multimedia news presentations. Its required redundancy of presented information

is one reason why elaboration theory cannot be used to compare multimedia

 

'82 Reigeluth, Charles M., Iastructigral Theorie_s in Action: Lessons Illustrating Selected Theories an__d_
Models, Hillsdale: L.Erlbaum Associates, 1987: 24-25.

'83 Kearsley, Greg, “Elaboration Theory,” The Theory in Practice Database, The Encyclopedia of
Psychology Society. 2001. <http://tip.psychology.org>.

'84 Kearsley “Elaboration”.

'85 Kearsley “Elaboration”.

 

63

news presentation methods. The journalistic style for news writing strictly forbids
this kind of redundancy. In addition, this experiment randomized treatments and
each of the multimedia components of each story were carefully selected to
assure the messages each presented were not redundant with either other
multimedia material or material in the text.

A key idea of elaboration theory is that readers need to develop a
meaningful context into which subsequent ideas can be assimilated.186 News
stories with multiple multimedia information sources that are, in effect, nested
together to create a Web news site presents lnfonnation in the manner in which
this theory requires.

Theory proponents claim elaboration results in more stable cognitive
representations—fostering better retention and transfer—and increased reader
motivation through the creation of meaningful contexts?87 But it is the
successive, redundant nature of the required lnfonnation presentation that makes
elaboration a poor choice for this research.

Yet another theory that also shows potential in explaining how audiences
process multimedia news lnfonnation is symbol system theory?88 Much of the
work done with this theory has been on television and film, but like DCT, there
has been a significant amount of work done recently with computer content?89

The theory proposes that each medium is capable of conveying content via

 

'86 Kearsley “Elaboration”.
‘87 Kearsley “Elaboration”.

'88 Salomon, Gavriel, Interaction of Media, Cogpition, and Learning, Hillsdale: L Erlbaum Associates,
1994: 54.

'89 Kearsley, Greg, “Symbol Systems,” The Theory in Practice Database, The Encyclopedia of Psychology
Society. 2001. <http://tip.psychology.org/>.

nil

certain inherent symbol systems?9° For example, television requires less mental
processing than reading. Meaning secured from viewing television tends to be
less elaborate than those secured from reading. Meaning extraction from various
media is based on individual differences?91 Thus, a person may acquire
lnfonnation about a subject they are familiar with equally well from different
media but be significantly influenced by a single medium for certain information.
There are symbol systems associated with the various media that affect how
lnfonnation can be acquired. Some of these symbol systems relate directly to the
kind of content being presented?92 Others vary according to the relative ease
with which information is encoded into memory; for example, as noted above,
television takes less effort to imprint than does text. “Thus, symbol systems partly
determine who will acquire how much knowledge from what kinds of
messages.”193 Symbol system theorists argue schema play a major role in
detennlning how messages are perceived. Individuals create individual
anticipatory biases that influence what lnfonnation is selected and how it is
interpreted?94

Symbol system theory offers a plausible explanation as to how memory
might be created from reading news. But for the purposes of this work, the
reliance it places on individual differences makes it too difficult to control for all

individual variables that might affect their predispositions for relating to various

 

'90 Kearsley “Symbol.”
'9' Kearsley “Symbol.”
192 Kearsley “Symbol.”
‘93 Salomon Interaction, p. 227.
'9‘ Kearsley “Symbol.”

 

65

media. DCT, with its simple picture superiority effect on recall makes research of
this kind easier to explain and control for.
Conclusions

DCT has been used effectively in educational research to study the
efficacy of multimedia classroom tools (e.g., computers) as compared to
traditional teaching methods. Classrooms, by definition, are similar to mass
media. Messages are delivered from a single source to a large group. It has also
been used extensively to compare text and pictorial symbols’ advantages in
creating recall.195 Yet, for some reason, mass media researchers have not used
this theory to any great degree in comparative media recall studies or others for
that matter.

This is very surprising, especially considering the parallel structure of the
theory and many of the multimedia products this dissertation proposes to test.
DCT’s assumption of a hierarchical structure of language closely tracks the
hierarchical structure of media. The nested structure it assumes for imagery
systems tracks the nested organization that occurs when visual elements are
added to a Web site. The inter-relatedness of verbal and imagery systems,
especially during the process of meaning creation also tracks how Web pages
are organized to present news information.

Lastly, this theory, unlike many others concerns itself almost completely
with imprinting lnfonnation in memory. If news-editorial is indeed an end, then
recall of news is the way to measure it. DCT is the roadmap that maps the paths

that memory must take for news recall to occur.

66

 

'95 These studies will be cited in the literanu'e review portion of this dissertation, where appropriate.

67

a“

Chapter 2

LITERATURE REVIEW

Introduction
The World Wide Web is a media channel that is affecting its more traditional
competition in much the same way each traditional media channel did when it
was not traditional, but, in fact, “new” and entering the world media market for the
first time. New media traditionally emerge in the media market and erode others'
audiences, wooing first young people, then more mature audiences from more
traditional competitors. As advertisers watch their market audiences migrate
away, they, too, move their trade to these new media channels. Like their
advertising counterparts, media firms follow their audiences as they see them
migrate to other channels. They follow, using their capital to purchase access
and then create content in these new channels?96

Consider what happened when TV entered the media market in the late
19403. Heretofore, radio had enjoyed a golden age. Advertisers wishing to
efficaciously reach mass audiences, especially national mass audiences, had no
choice but to run ads on the radio networks of the day?97 But when television
entered the market, viewers flocked to this new medium, despite the relatively
high cost of the TV signal receivers necessary to capture the signals. The same

firms that enjoyed a market success in the radio broadcast business, migrated to

 

'96 This is evidenced by the creation of any of a number of Web sites, such as CNN .com, NYTimes.com or
LSJ.com.

'97 LaRose, Robert, and Straubhaar, Joseph, “Audio Media: Radio and Recorded Music,” Communications
Media in the Information Sociapy, Belmont: Wadsworth, 1997: 177-178.

68

television.198 Advertisers followed the firms and audiences to TV, with many
national accounts forsaking radio for TV and embracing the novelty of being able
to present their products visually and with sound?99

Similarly, the Web is beginning to steal audience members away from
mainstream media. A 1996 Pew Research Center study showed network news
audiences were losing audience and that the greatest audience losses were
among young people and computer users.200 According to a 1998 Pew Research
Center report, 46 percent of all lntemet users began using the channel that year.
A little more than 50 percent of these users were aged 30—49 and 25 percent
were aged 18-29. Almost 60 percent of this group earned $30,000 or more per
year, with 35 percent of those earned more than $50,000. More than half
accessed the lntemet for pleasure, as opposed to work purposes.201 These
trends speak to a promising future for the Web. The audience is both young—
promising longevity—and well moneyed, which is attractive to advertisers.

In a 1998 survey of more than 5,000 Web users, there is evidence the
Web is supplanting other media use. One of the questions the survey posed
dealt with the frequency of their use of the Web instead of watching TV. More

than 80 percent of respondents said they used the Web daily and weekly instead

 

'98 LaRose “Audio,” 180.

'99 LaRose “Audio,” 180.

20° Pew Research Center for the People and the Press. "TV News Viewership Declines: The lntemet News
Audience Goes Ordinary,” 2001. <http://www.peop1e—press.org/mediarpthtm>.

20' Pew Research Cent_er for the Peoplg and the Pres_s. "Online Newcomers More Middle-Brow, Less Work-
Oriented: The lntemet News Audience Goes Ordinary," 1999.
<http://www.peoplepress.org/tech98sum.htm>.

69

Wifll

of watching TV; similarly, nearly 50 percent said they used the Web daily and
weekly instead of reading newspapers, magazines or books.202

While audiences are seeking newer, more novel media channels,
advertisers have not yet made this same decision. Thus, even though an
audience exodus from television has clearly begun, that is precisely what it has
done: it has only just begun. The same is true for the commercial exodus that
must accompany this trend for it to repeat other historical audience exoduses.

Network programming, once the king of the prime time audience, has lost
considerable market share to cable TV. In 1979, 90 percent of television's prime-
time audience watched network programming. By 1997, networks drew only 50%
of the prime-time audience.203 This trend is very similar to one associated with
newspaper readership. From 1945 to 1970 newspaper circulation rose steadily in
this country. Since 1970, it has stayed roughly the same, despite population
increases. Today, despite flat circulation increases, those who do read
newspapers spend less time doing so.204

A top-rated TV show, such as ER, draws 30 million viewers per
episode.205 Cable TV claims a presence in only two-thirds of US. households.206

A hit cable program draws from two to three percent of the national cable

audience.207 Therefore, it is not surprising that network TV still attracts the lion's

 

202 gyp. "GVU's Tenth WWW User Survey Graphs" 2001.
<http://www. gvu. gatech.edu/user_surveys/survey- 1 998- 10/graphs/graphs.html#technology/>.

203 Barans, Stanley J. “"Media, Media Industries and Media Audiences: Television," Introduction to Ma_§§
Communication: Media Literacy and Culture, Mountain View: Mayfreld, 2001, 288.

20‘ Barans, Stanley J. "Media, Media Industries and Media Audiences: Newspapers," Introduction to Mag
Communication: Media Literacy and Culture, Mountain View: Mayfreld, 1999, 102.

205 Barans “Television,” 282.

20" LaRose “Multichannel Media,“ 251.

207 LaRose “Multichannel Media,” 251.

70

 

share of TV advertising revenues. In 1999, network TV pulled in a little more than
$21 billion in ad revenue, while cable TV earned only $13.6 billion.208

Globally, lntemet advertising totaled an estimated $2.1 billion in 1998,209
$4 billion in 1999210 and $8.2 billion in 2000.211 Thus, though the Web has not
yet surpassed its more traditional competitors' ad revenue numbers, it is gaining
on them.

Clearly, network TV has little to worry about from the lntemet in the near
term, but it cannot discount it either. Certainly, the mergers between Microsoft
and NBC and Time-Wamer and AOL presage a bold new online world. There
can be no mistaking the vision that went into these mergers. Wedding the world’s
foremost software publisher and videotex service with two of the world’s largest
media content distributors are classic examples of modern day vertical
integration, especially in light of Microsoft’s near-proprietary interest in the World
Wide Web. As the publisher of the leading Web browser product in use today,212
lntemet Explorer, Microsoft effectively controls how the majority of Web denizens
view its content. One needs look no further than its proprietary channel
architecture, now known as "Windows Media".213 The company has created an

archived content system that offers the most sophisticated and logically

 

2"" Barans “Television,” 283.
209 Nua lntemet Surveys. "Total Revenue From Advertising - 1998,"
<http ://www.nuaie/surveys/analysis/graphs_charts/ 1 998 graphs/ad_revenue.html>.
21° Interactive Advertising Bureap. Price-Waterhouse, "IAB lntemet Advertising Revenue Report: 1999
Third-Quarter Highlights,” March 2000, <http://www.iab.net/>.

Interactive Advertising Bureau. “Interactive Advertising Bureau (IAB) Reports $8.2 Billion Ad Revenue
in the United States for Year 2000; Fourth Quarter Totals $2.2 Billion”, 2001.
<http://www.iab.net/>.

2'2 Proffrtt, Brian. “BrowserWatch Stats Station,” 2001.

<http://browserwatch.intemet.com/stats/stats.html/>.

2 '3 Windowsmedia.com, 2001. <http://www.windowsmedia.com/>.

211

71

«mm-hit“ “1 '

organized offering of Web content available today. Users need not search, nor
laboriously enter URLs to access content. A simple point-and-click interface
instantly delivers a sophisticated offering of the Web's best original, multimedia
"programs“ ranging from The Discovery and the Arts and Entertainment
Channels to Warner Bros. cartoons.

All the elements are in place for traditional media to migrate news
products to the Web. Useful client software, such as RealPlayer Plus214 and
Windows Media Player,215 now enable users to conveniently access a wide
variety of online news content packaged in sound and video files, as well as
traditional content such as scanned images, static art and text.

Previously, many users shied away from video and audio files because
they were so large and took so long to download, then play. But the advent of
streaming technology has changed all that. Today, even those with 56.6-kilobaud
per second dialup connections can listen to lntemet radio shows or watch CNN
newscasts without waiting for them to first download, then convert, then play.
Streaming technology downloads content to create a buffer between that which is
being viewed and that which is downloaded and processed for viewing.216 As
long as the content continues to stream to the users computer—uninterrupted—
viewers can view content with a minimum of down time, making the viewing
experience much more similar to that of actually watching TV or listening to a

radio.

 

2” Real.com. RealPlayer Plus, <http://www.real.com>.
2'5 Microsofi Cogporation. Windows Media Player, <http://www.windowsmediacom/>.
2'6 Howe, Dennis. Free Online Dictiom of Computing, <http://foldoc.doc.ic.ac.uk/foldoc/>, 2001

72

lntemet 2, a consortium of universities and companies that have
networked together as part of a joint lntemet media research and development
effort, also promises new lntemet technology to aid multimedia development on
the Web. In existence for approximately two years now, this consortium of
“corporations, colleges and universities will use I2 to explore areas of instruction,
research, and public service, such as Ieaming ware (software designed for
classroom and distance Ieaming), digital libraries, tele-immersion, and virtual
laboratories. Eventually, the plan is that the technology will be available to
everyone.”217

There is also a vast array of access options available to consumers.
Cable companies offer high-speed lntemet access with lntemet-to-home-
computer connection speeds, known as downstream connections, ranging from
500 kilobits per second up to 1.54 megabits per second (referred to as T1).218
Regional Bell operating companies have been offering Digital Subscriber lines
(xDSL) to customers. This service, which, like cable modems, requires a special
modem to be installed, offers access speeds varying from a low of 256 kilobits
per second to a high of 8 megabits per second, depending on how far consumers
live from the area’s switching office. xDSL's chief advantage is that it provides
high-speed access via regular telephone lines.219 Lately, as AT&T has competed

mergers with a number of cable TV companies to become the largest cable TV

provider in the United States, the firm has begun offering its own “broadband”

 

2” Lundsten, Apryl and Doiel, Robert. "Internet 2; What is 12?", 27 April 1999.
<http://www.builder.com/Servers/IntemetZ/index.html>.

Cable Television Laboratories Inc. “Cablemodem.org,” 2001.<http://www.cablemodem.org>.

“9 Howe FOLDOC, <http://foldoc.doc.ic.ac.uk/foldoc/>.

218

 

 

73

services, featuring combinations of DSL, satellite access and cable modem
service. All these new access schemes are priced generally at $50 per month
and will likely be offered for less as more consumers seek faster access speeds.

Thus, from a technical standpoint, consumers who wish to access
multimedia content from the Web have some of the necessary tools available to
them and will have even more available soon. What remains is for networks to
develop further to allow higher speed access and for content providers to deliver
it. Why they have not begun to do so on a large scale seems mystifying.

One reason perhaps is that most news operations are in fact, single
medium rather than multimedia operations. They have contented themselves
with Web media products that scratch at the surface of the Web’s capability to
deliver news to mass audiences. They have not crept far beyond offering
rudimentary graphic presentations of text, pictures and some animated charts
and graphs. Only a few sites220 have attempted to meld video, audio, text and
scanned photographs together in an effort to present multimedia news. For
many, the reason is a lack of access to the media materials they do not directly
produce. Indeed, for such conglomerated media operations as Time-Wamer, Cox
and even the Tribune Co., securing content that might be converted to
multimedia online news presentations are relatively easy. For most, however, it
would be a significant hardship, involving negotiating complex licensing

agreements and royalty arrangements, before such content would be presented.

 

220 By late 2001, The New York Times, BBC, CNN and MSNBC were the major news firms that ventured
into multimedia presentation of Web news.

74

To be convinced such complicated and arduous undertakings are
necessary to media operations, it would be fruitful to explore if multimedia news
presentations are worth delivering to Web audiences. Media history is spotted
with many well-intentioned flops that promised new visual and sensual
experiences, only to fall flat in the face of audience ambivalence. Who can forget
such mass media debacles as Viewtron, 3-D movies and Smell-O-Vision?

This dissertation proposes to explore the value of multimedia presentation
of news by testing subject recall of news presented within a variety of
presentation modes: text alone; text and photography; text and multimedia video;
and text, multimedia video and embedded audio. This section of the dissertation
will review the experimental and theoretical literature that is related to the
experiment that tests the theoretical framework laid out in Chapter 2. As such, it
will serve as a comprehensive examination of all related literature not covered in
depth in the previous chapter and will focus itself primarily on the literature that
relates to the experimental portion of the dissertation.

Instrument Design

More and more media are entering the World Wide Web with news pages. This
phenomenon is due, in part, to the growing connectivity that is occurring in US.
households. But the same media firms that are putting the news on the Web are
still casting about looking for ways to best present news on this channel.

This study proposes to test and compare four modes of online news

presentation delivered to young audiences. Experimental subjects will be tested

75

err-“m" -

on their recall of the facts associated with the news stories they read and listened
to on the online news Web pages constructed for the experiment.

Each of the presentation modes to be studied in this work is a
representative of the media that comprise the multimedia environments available
to all who would build Web sites. Each component medium has its strengths, its
weaknesses, its audiences and its non-users. Individually, they represent a
segmented news audience whose preferences follow lines drawn by age,
income, education and even geographic location.221 But presented together, they
represent a powerful new combination that is a singular medium, just as its parts
are.

Owing to the relative novelty of this new media channel, very little
research has been done to evaluate it. Therefore, a more integral approach to
literature review was pursued to create this chapter. Among the research areas
to be examined in this work will be text and reading comprehension within
multimedia and computer environments, Web page design studies, graphical
user interface design studies and content recall research. Some review of dual
coding theory related work will be covered in this chapter, but only those studies
that are media related. For a detailed handling of dual coding theoretical
literature, please refer to Chapter 3.

At the outset of any discussion of computer media, some measure of
interactivity must be considered. Interactivity refers to the unique exchange of

information that occurs between a computer and a computer user.222 This

 

22' "GVU's WWW User Survey," <http://www.gvu.gatech.edu/user_surveys/survey- l998-lO/graphs/#use>.
222 Najjar, Lawrence. “Multimedia Information and Learning,” Journal of Educational Media and

76

exchange is mediated by the information provided by both user and machine and
will vary according to what is provided by both participants during the
exchange?23 This active participation by users may well create greater reader
attention than those who read text in non-interactive media, such as books,
newspapers or magazines. It may also lead to greater reader involvement, and
perhaps even comprehension, since they are actively manipulating this
information on the way to acquiring it.

By and large, reading text online is a task involving working memory and
content format. This experiment will look closely at content format. Working
memory is that part of memory that takes care of temporary lnfonnation storage
and processing when individuals are engaged in more complex tasks.224 Its
function is closely tied to the referential processing described in dual coding
theory.225 Working memory takes newly acquired lnfonnation and ties it to related
information stored in long-tenn memory.226 Referential processing is necessary
to text comprehension. Research has shown that less complex text tends to
diminish the gap in text comprehension between those individuals with greater
and lesser amounts of working memory.227 Therefore, news copy would be far
less working memory-intensive than other kinds of texts. Thus, the task of

balancing content acquisition skills (such as Web page navigation or on-site

 

Multimedia, 5, 1996: 129-150.

223 Sims, Roderick C. “Engagement, Control and Learner: A Theoretical Appraisal of Interactivity,”

Australian Society for Computers in Learning in Tertiary Education Conference, Adelaide, 1996.

22‘ Baddeley, Alan D. Short Term Memory. New York: Oxford University Press, 1986.

22’ Paivio, Mental 69.

22° Sentius Cogp_oration. “Guide to Online Publishing for Reading Comprehension — Section II,” 1997.
<http://Www.sentius.com/>.

227 Budd, D., Whitney, P., and Turley, K.J. "Individual Difi‘erences in Working Memory Strategies for
Reading Expository Text," Memogy and Cogg'tion, 23.6, 1995: 735 - 748.

77

‘mh a 1133‘. q

search engine operation) with lnfonnation related to the media content would
likely have less effect on recall. Part of this might also be attributed to the amount
of reader involvement in the act of processing information.

Roger Fidler spent a number of years working for Knight-Ridder
developing a portable "news tablet“. This device was designed to download news
information from remote locations to provide readers with the same portability
they experienced with paper news products. 228 Fidler contends digital media
forms must be as ”comfortable” to read as paper media.229 In some research he
was involved in, the findings indicated a preference among subjects for "portrait-
oriented” text (as opposed to landscape-oriented).23o Historically, PC screens
have been landscape-oriented (i.e., having a horizontal orientation). Original
designs, prior to the release of the IBM-PC in 1981, featured portrait-oriented
screens (screens that feature more vertical views) that were originally fashioned

to mimic a single page of 8-1/2 x 11-inch paper.”1

Both these facts point to a
recognition by computer hardware and Web designers that even though Web
material is presented via computer, it is still essentially viewed as a document.

Indeed, a study that compared reading speed and comprehension of screen and

 

m Maney, Kevin. “High Tech Tablets: Next Step for Newspapers,” USA Todgv Tech Report. 28 Feb.
1999. <http://www.usatoday.com>.

229 F idler, Roger. Mediamorphosis: Understanding New Media,_ (Thousand Oaks: Pine Forge Press, 1997:
236.

23° Wearden, Stanley "L, Fidler, Roger, Schierhorn, Ann 3., and Shierhorn, Carl. "Portrait vs. Landscape:
Potential Users' Preferences for Screen Orientation," Association for Education in Journalism and
Mass Communication annual convention, Newspaper Division, New Orleans, 1998.

23 ' Smith, Douglas, and Alexander, Robert Fumbling the Future: How Xerox Invented, Then Ignored the
First Personal Computer, New York: Morrow, 1988: 89.

 

78

paper texts found no differences between reading text on paper and reading text
online.232

Similarly, a case study dealing with how lntemational Monetary Fund
economists gathered and processed information concluded that documents that
did not rely on their linear structure and were used for search and retrieval tasks
on small sections of text and well-defined texts (such as manuals or dictionaries)
were better suited to digital presentation?33 Again, at least some of these
attributes are common to news copy.

The beginning of the PC revolution in the late 1970s brought computers
that were previously too large or too expensive into homes and offices.”4 These
computers could be dedicated to singular tasks, as opposed to larger, heavy-duty
computational tasks mainframes and mini-computers of the day were doing. But
one of the drawbacks to these earlier machines was that their interfaces—the
screen views that interpreted machine language into human language so that
people could effectively interact with their computers—were not very highly
developed. This relegated early PCs to those people who could Ieam the
extensive DOS and software package command languages, thus eliminating

much of the public from participating in the personal computer revolution.

 

232 Osborne, D.J., and Holton, D. "Reading from Screen versus Paper: There is No Difference,"
lntemational Journal of Man-Machine Studies, 28, 1988:1-9; O’Hara, Kenton and Sellen, Abigail.
“A Comparison of Reading Paper and Online Documents,” CHI 97, Atlanta, Georgia, 22-27
March 1997.

2” Sellen, Abigail, and Harper, Richard. "Paper as an Analytic Resource for the Design of New
Technologies," CHI 97, Atlanta, Georgia, 22-27 March 1997.

23‘ LaRose “The Computer Industry,” 293.

79

. ”IF-Ii- --71‘?!
‘A

But in 1984, Apple Computers introduced the Macintosh, and with it, the
first high-resolution graphical interface?35 It captured the imagination of many
potential computer users who were put off by the dark screen, the single carat
and the single, blinking block of the disk operating system (DOS) screen. This
interface even captured the imagination of Microsoft Corporation, which
incorporated the graphical user interface into its next generation operating
system, which would be called Windows.

Today, graphical user interfaces "contribute to the ease of use of
computers. Interfaces based on windows, icons, menus and pointing allow fairly
non-trivial operations to be performed with a few mouse clicks."236 Research on
graphical user interfaces has shown that graphic or animated visuals aid greatly
in computer user interfaces. For example, a 1995 research project done at the
University of Wisconsin found that still graphic and animated visuals in online
help instructions enabled users to perform tasks in less time and with fewer
errors than those who used online help that did not contain visual graphic
elements?37 This finding is in concert with dual coding theory principles that
suggest that the presence of images with instructional text enhances humans’
abilities to Ieam.238

The World Wide Web uses a graphical user interface to allow users to

navigate its sites. Apple Computers originally introduced the navigation scheme

 

2” LaRose “The Computer Industry,” 294.

2“ Worden, Aileen, Walker, Neff, Bharat, Krishna and Hudson, Scott. "Making Computers Easier for Older
Adults to Use: Area Cursors and Sticky Icons," CHI '97, Atlanta, Georgia, 22-27 March 1997.

237 Harrison, Susan M.. "A Comparison of Still, Animated or Non-Illustrated Online Help with Written or
Spoken Instructions in a Graphical User Interface," Proc. of Computer-Human Interfaces '95.

2” Paivio “Mental,” 206.

80

My}

the Web employs now in a 1987 programming package called HyperCard?39
This package enabled designers to create user interfaces whose purpose was to
deliver content to audiences. HyperCard enabled users to create "stacks" of
cards that were actually screens, much like today's Web pages. Another
similarity Web pages have to HyperCard stacks is that screen-to-screen
navigation, or in this case, card-to-card navigation, was enabled by buttons users
mouse-clicked, linked text and cards that featured both text and graphics on the
same page?40

Conceptually, HyperCard was the precursor to HTML editors and, in doing
so, was the precursor to Web page architecture. This architecture constitutes a
universal language among Web users woridwide that directs how they navigate
the space in which content is located. It also directs how they navigate within
web sites, leaving memory free for higher order tasks, such as processing
content and committing it to short-term or long-term memory.

A fair amount of DCT research on reading text on computer screens
exists, much of it dealing with developing graphical user interfaces (GUls)?"1
There is also research that deals with reading news text on computers, but these
studies presented news material on computers while failing to provide an

environment that mimicked current online news media appearance. This

occurred either because the studies were done when the only online news

 

239 Straubhaar, Joseph, and LaRose, Robert. "Computer Media and the lntemet," Media Now:
Communication Medi_a in the Information Agg, Belmont: Wadsworth. 2000: 21.
24° Howe F OLDOC, <http://www.foldoc.org>.
2‘" e.g., Lomicka, Lara L., “To Gloss or Not to Gloss: An Investigation of Reading Comprehension
Online,” Langt_1_age Learning and Technolggy, January 1998: 41-50; Chun, Dorothy M., and Plass,
Jan L. “Research on Text Comprehension in Multimedia Environments,” Langu_age, Learning and
Technology, July 1997: 60-81.

 

81

 

material were still being offered as part of videotex services, such as AOL and
CompuServe, or even prior to that, when newspapers were experimenting with
other services, such as Knight-Ridder's “news tablet” and Vievvtron projects.

One of the first considerations of this experiment was how do people read
and process online information. For the purposes of this experiment, since it
focuses on news presentation modes, it seemed wise to target research dealing
with multimedia information processing. For this kind of information, it was
necessary to turn to the educational literature. In that sector, researchers have
been evaluating the efficacy of multimedia as a classroom tool for a number of
years. Since this project is using dual coding theory (DCT) as its guide, it was
also necessary to look specifically at research that used DCT.

First, however, it is important to define what multimedia is. There are a
number of definitions throughout many genres of research literature. Most seem
to agree, however, that multimedia uses text, graphics, animation, pictures, video
and sound to present infom'ration?42 Some limit this definition to an attribute of
microcomputers?43 Still others note this content presentation mode is not limited
to one particular medium, such as computers, but instead represent the
convergence of several different media.244 Perhaps the best definition for the
purposes of this work is that it is the integration of audio, visual and text

Information that erases distinctions between media systems?45

 

2‘2 e.g., Najjar, Lawrence J. "Multimedia Information and Learning," Journal of Educational Multimedia
and Hypgrmedia, 5, 1996: 129-150.

243 Barans Introduction, 493.

2“ Wolfe, Mark R. “The Wired Lot: Lifestyle Innovation Diffusion and Industrial Networking in the Rise
of San Francisco’s Multimedia Gulch,” Urban Affairs Review, May 1999: 707-728.

2“ Straubhaar, Joseph and LaRose, Robert. "The Changing Communications Media Environment,"
Communications Medfl the Informgtion Society, Updated Edition, Belmont: Wadsworth. 1997: 20.

82

Thus, if media have converged within this channel to present a veritable
rainbow of content types, will the audience's ability to process this information
collectively match or even surpass its recall recorded from viewing/reading the
individual media components within their native environments? Will newspaper
recall be higher or lower than multimedia news recall? Similarly will broadcast
news recall be higher or lower than multimedia news recall?

These are natural questions we ask of any new media, especially if
previous historical patterns of media adoption by audiences hold true. Are so-
called “new media” necessarily better than what we have now? In diffusion of
innovation theory, this concept is known as relative advantage?46 How much
improvement will this new thing offer over what we are now using? Clearly this is
a question media firms must ask and answer before entering the Web with new
media products.

Educational research has generated large numbers of studies detailing the
benefits people believe can be derived from using multimedia to aid in
Ieaming?47 There is another old saw—unsupported by any research—bouncing
around journals, speeches and lectures: ". . .people generally remember 10
percent of what they read, 20 percent of what they hear, 30 percent of what they

see (and) 50 percent of what they hear and see. . ""248 All this folklore has

 

2‘6 Rogers, Everett M. Diffusion of Innovations, New York: Free Press. 1995: 73.

247 Op. cit., Najjar. "Multimedia," 129; Bosco, J. "An Analysis of Interactive Video," Educational
Technology, 25, 1986:7-16; Fletcher, D. "The Effectiveness and Cost of Interactive Videodisc
Instruction," Machine-Mediated Learning, 3, 1989: 361-385; Rigrry, J.W. and Lutz, K.A. "Effect of
Graphic Analogies of Concepts in Chemistry on Learning and Attitude," Journal of Educational
Psychology, 68, 1976: 305-311.

248 Treichler, D.G. "Are You Missing the Boat in Training Aid?" Film and A-V Communication, 1967: 14-
16.

83

combined to give people a general sense that multimedia is desirable if message
delivery is their goal.

But there is a fair amount of support for this feeling in the literature as well.
Georgia State University educational researcher Lawrence Najjar cited meta-
analyses that examined more than 200 studies comparing Ieaming information
presented in traditional classroom lectures to Ieaming the same information from
computer-based, multimedia instructions?"9 Subjects ranged from K-12 and
higher education students to people from industry and the military?50 Subject
matter varied over a wide variety of topics including the sciences, electronic
equipment operation and foreign language study.251 Control groups usually
Ieamed in classrooms or via classroom lectures combined with hands-on
experience.252 Comparison groups usually Ieamed information from interactive
videodiscs or some other kind of computer-based instruction?” Learning was
most often measured using tests of achievement or perfonnance?54 Over this
wide range of students and topics, Ieaming performance was greater when
lnfonnation was presented with computer-based multimedia systems?55 Other
studies showed that Ieaming using multimedia systems took less time than
conventional Ieaming methods?56

One of the reasons for this seems to be interactivity. Interactivity refers to

situations in which real-time feedback is collected from receivers of messages

 

2‘9 Najjar “Multimedia," 130.
25° Najjar “Multimedia,” 130.
25 ' Najjar “Multimedia,” 130.
”2 Najjar “Multimedia,” 130.
2” Najjar “Multimedia,” 130.
2" Najjar “Multimedia,” 130.
2” Najjar“Mu1timedia,” 130.

84

and used by the source to continually modify the message sent back to the
receiver.”7 For example, where a user clicks a mouse on a screen determines
what kind of feedback the Web site will send to the receiver. Clicking on different
parts of the screen will elicit different responses. Because users must click their
mouses, this kind of communication requires a measure of engagement and
involvement in content that is, at least somewhat higher than someone who is
merely reading a book and turning pages. Mouse clicks result in continued
delivery of content. Contrast this to a classroom lecture, in which involvement
does not result in continued feedback (the lecture will go on, whether or not the
receiver is involved in taking notes or listening). By contrast, educational
researchers view interactivity as mutual action between the learner, the Ieaming
system and the Ieaming material.258 From their perspective, multimedia-based
Ieaming system success is due to the interactivity the Ieaming tools deliver to the
information.

Najjar also reviewed nearly 40 multimedia studies that pointed to a
temporary, positive novelty effect on Ieaming. When compared to traditional
classroom lecture Ieaming, improvements were higher for groups that used
multimedia for four weeks or less but the Ieaming advantage tailed off

significantly after eight weeks?59

 

25" Najjar "Multimedia," 130.

257 Straubhaar Communications, 12.

258 Op. cit., Najjar “Multimedia,” 131: Fowler, B.T. “The Effectiveness of Computer-Controlled
Videodisc-Based Training,” diss., University of Iowa, 1980.

25" Najjar “Multimedia,” 132.

85

Perhaps more significant, as it relates to this work, is the meta-analysis of
redundant multimedia and "monomedia."260 Redundant multimedia is defined as
a Ieaming system in which the same verbal information is presented using audio
and printed text together; monomedia is a Ieaming system that presents
lnfonnation within a single media mode, such as audio alone, or text alone.
Research has shown redundant media improve Ieaming when compared to
monomedia presentations with older students but not with children. In fact, one

study,261

published in 1987, exposed college students to a series of voice-over
news stories that varied in their amounts of audio and visual channel
redundancy. The students' recall of both auditory and visual information was
measured; the researchers found higher auditory recall and story understanding
in a high redundancy condition. The reverse was true for visual information: recall
was higher in low redundancy conditions. Clearly, this experiment and the
previous paper show that recall is affected by redundancy and should therefore
be avoided in any experiment whose goal is to compare the efficacy of various
media presentation modes.

Neariy all the recall studies reviewed for this work analyzed the data
collected for it using Analysis of Variance (ANOVA). ANOVA is used to test
hypotheses about differences between two or more means.262 When there are

only two means to test, the T-test is used, but when there are more than two

means, as in the case for this experiment, which features recall tests of four

 

26° Najjar "Multimedia," 133.
26' Kenney “Memory and Comprehension of TV News Visuals,” News Photoggpher, August 1992:
52-53.

86

treatments, repeated use of T-tests creates a marked increase in the possibility
of a Type I error (in which a true null hypothesis might be incorrectly rejected)?63
It is more preferable to use a within-subjects design and to test those results for
significance?64 In this kind of design, each subject is tested at each level of the
experiment. In this case, an individual level is defined as each one of the media
presentation modes. Thus, each subject will read each of the stories that
represent each of the levels of the experiment and each of their mean recall
scores will be compared to other subjects”.
Reading and Publishing Online
People read text in a variety of ways:265
0 linear - sequential reading - novels
. hierarchical - navigating, using a table of contents or tree
a structure - dictionary, reference materials
. reading reference texts using focused (e.g., keyword)
retrieval - encyclopedias, indexes
. mixed reading with linear and nonlinear navigation -
textbooks
o nonlinear browsing - jumping around different topics to

form a chain of associations (path of ideas) - hypertext

 

’62 HyperStat Onlin_e. Lane, David M., “ANOVA,”
<http://davidmlane.com/hyperstat/intro_ANOVA.html/>.
263 HyperStat Online. Lane, David M., “Type I and Type II Errors,”
<http://davidmlane.comlhyperstat/Al8652.htrnl/>.
2“ Lane “ANOVA.”
2‘” Chignell, Mark H., and Valdez, J. Felix. "Methods for Assessing the Usage and Usability of
Documentation," Third Conference on Quality in Documentation, Waterloo, Ont., 1993.

87

When people read the news, they use several of these reading behaviors.
Whether on or offline, readers engage in some hierarchical reading. Online and
traditional print media incorporate both tables of contents and sections to
organize and package news in each edition. Those who use browsers practice
focused retrieval with the find function (Ctri-F, Apple-F) built into this software to
search out news of interest within online news systems. Linked text also
facilitates non-linear browsing.

The computing literature refers to reading behavior as text navigation
strategy,266 conceiving it as a way in which people move their attention through a
document. Computer interface researchers seem less concerned with Ieaming
outcomes and more concerned with reading process?67 They also differentiate
between reading computer documentation and printed matter and frequently
attempt to compare the two. The computer documentation most often studied in
this realm of research is called hypertext. Hypertext is defined as ”...documents
[in which] the user can follow paths of links through text and graphics by
selecting terms (anchors) marked by the author of the document in some way

(i.e., underlining, using a different color for the word, etc)"268

 

266 Harrison, Susan M.. “A Comparison of Still, Animated, or Non-Illustrated Online Help with Written or
Spoken Instructions in a Graphical User Interface,” ACM Special Interest Group Computer-Human
Interaction Conference, Denver, Colorado. 1995: 2.

267 e.g., Harrison, “Comparison”; Altman, Erik, Larkin, Jill H., and John, Bonnie E. “Display Navigation by
an Expert Programmer: A Preliminary Model of Memory,” ACM Special Interest Group, Computer-
Human Interaction Conference, Denver, Colorado. May 1995; Furnas, George W., “Effective View
Navigation, ACM Special Interest Group, Human-Computer Interaction Conference, Atlanta, Georgia.
1997.

268 Chignell, "Methods".

88

In the late 1980s and early 1990s, there was a large body of research
comparing hypertext to printed documentation.269 Principally, these studies
compared the "usability” of these two kinds of documentation. Usability studies
attempt to determine how well and easily users can use programs and, in this
case, documentation, in the manner in which its designers intended they be
used?70 The findings of many of the studies in this era were contradictory and
inconclusive in terms of awarding a clear advantage to either hypertext or printed
documentation, especially when they created experiments with linear
comparisons that used printed documentation as a control or objective standard

in the experiment?71

Yet when researchers compared the utility of hypertext
online against similar characteristics of printed documentation, they were able to
conclude that the best way to use hypertext was in ways in which its fullest
capabilities could be used?72 In other words, simply placing text online created
no value or advantage over printed documents. Using links and anchors within
text to encourage reader browsing and other non-linear reading behavior greatly
added to the utility of online documentation?73

In this way, then, the value of the online material is enhanced because users

can experience its full functional range with their computers. Text that does not

 

269 e.g.,Monk, A.F, Walsh, P., and Dix, A.J. "A Comparison of Hypertext, Scrolling and Folding
Mechanisms for Program Browsing," in Jones, D.M. and R. Winder, R., eds., People and Computers
fl, London: Cambridge University Press. 1988: 421-435; Egan, D.E., Remde, J.R., Landauer, T.K.,
Lochbaum, CC. and Gomez, L.M. "Behavioral Evaluation and Analysis of a Hypertext Browser,"
Proc. of CHI'89, Conference on Human Factors in Computing Systems, Association of
Computing Machines, New York. 1989; Walker, J .H., Young, E., and Marines, S. "A Case Study of
Using a Manual Online," Machine Mediated Ieaming, February 1990.

270 Karat, John and Dayton, Torn. “Practical Education for Improving Software Usability,” ACM Special
Interest Group, Human-Computer Interaction Conference, Denver, Colorado. 1995.

27' e.g., Harrison “Comparison,” 9.

272 Lomicka “Glossing,” 10.

89

feature functions that require computers may just as well be published as printed
documents. After all, a book can be easily placed next to a computer for ready
reference and be just as handy as online documentation. It is the addition of
computer-aided functionality—such as keyword searches, and task-sensitive
links—that creates perceived value in online text applications.

There are a number of measures document designers consider when
deciding if the material they wish to publish should be made available online or
as printed documentation: 2"

. Does the value of the online information exceed the cost
of putting it online?
. Does the information take advantage of the speed,
storage and display capabilities of the online medium?
0 Does it "communicate timely and critical lnfonnation?"
. Is it an evolving document, one that is updated or
reissued often?
0 Does it contain inter-related information?
. Is it a large, searchable document?
Online news meets or exceeds all of these measures. News has one
significant advantage over online documentation: it is generally accompanied by
advertising. Since advertising's purpose is to offset the cost of printing, paper and

delivery and to generate a prOfit, then publishing news on the Web that is

 

”3 O’Hara “Reading,” 9-15.
274 Horton, William. "What Should and Should Not Go Online," Desigg'ng & Writing Online
Documentation: Help Files to Hypgrtext, New York: John Wiley & Sons. 1990: 16-23.

90

accompanied by advertising surpasses the first standard of exceeding the cost of
putting it online.

News is a timely commodity, thereby satisfying the "timely and critical”
standard. The speed with which online news can be made available and updated
(satisfying yet another standard) to audiences adds a timeliness to news
products that is not possible with other media. Some media firms are using the
Web's full multimedia possibilities to create online news sites. Others, especially
print enterprises have discovered an after-market for archived news, selling
access to "old issues" via database access.275 Thus, the storage available via
servers can be used very efficiently and the documents are searchable, thereby
satisfying all the remaining standards for online use.

The timeliness and accessibility that having the news stored and available
via the Web actually creates value for some users. Because this information is
available on a 24-hour, seven day-a-week basis, and is constantly updated, the
value it accrues because it is available online is much greater than it would as a
printed product alone. Print products are not archive-able, not searchable, unless
filed in filing cabinets and therefore, are not as valuable.

Computer and Media Attitude Testing

Certainly one aspect common to many research projects dealing with any kind of
evaluation of computer use includes the use of computer attitude questions.
These questions are generally designed to discover if research subjects have
any attitudes that might create artifacts within the data gathered in the

experiment.

91

They are especially necessary when researchers use analysis of variance
(ANOVA) to determine the accuracy of averaged data?76 Questions dealing with
media preference and computer attitudes are crucial to identifying "error term."
Error term refers to the differences each subject has with each other?77 In the
case of this experiment, it refers to varying attitudes subjects have about
computer use and media. Once identified, the ANOVA results can be tested
against their presence to determine if these attitudes had a significant influence
on the responses subjects made, as a group, to study questions.

As varying computer attitude categories were unearthed in research,
researchers sought to test them for validity. There are a number of these studies
in the literature today that define valid scales for testing computer attitudes, such
as computer anxiety, computer confidence, computer liking and computer
usefulness.278 Since this study will use undergraduate students from a major
midwestem university, it is believed the relevance of these scales may have
passed. More and more, undergraduates attending universities are well schooled
in the use of computers and use them with the same skill they might a typewriter
and view them as necessary to their education as notebooks, pens and pencils.

Still, it would be foolhardy in the extreme to assume this would be true for all

 

275 365-Dav Archive. <http://archives.nytimes.com/archives/>.

m Hyperstat Online. Lane, David M., “Interaction,” <http://davidmlane.com/hyperstat/glossary.html>.

”7 HyperStat Online, Lane, David, “Within-Subjects,”
<http://davidmlane.com/hyperstat/within-subjects.htrnl>.

273 Loyd, Brenda H. and Loyd, Douglas E. "The Reliability and Validity of an Instrument for the
Assessment of Computer Attitudes," Educational and Psychological Measurement, 1985: 903-908;
Loyd, Brenda H. and Gressard, Clarice. "The Effects of Sex, Age and Computer Experience on
Computer Attitudes," AEDS Journal, 1984: 67-77; Loyd and Gressard. "Reliability and Factorial
Validity of Computer Attitude Scales," Educational and Psychological Measurement, 1984: 501-
505; Koslowsky, Meni, Lazar, Aryeh and Hoffman, Michael. "Validating an Attitude Toward
Computer Scale," Educational and Psychological Measurement, 1988: 517-520.

92

students, especially those who come from more rural areas of the region, where
computer use in classrooms is not as prevalent as it might be in larger, more
urban school districts. For this reason, then, valid and reliable survey questions
drawn from this genre of research were used to assess research subject
computer attitudes.

Another area likely to create artifacts that might adversely affect the
research's ability to discern advantages to various news presentation modes
would be attitudes towards various media. Research has shown that various
people have varying biases about certain media. Audience research has shown,
for example, that fewer than 30 percent of 18-29 year-olds in this country read
newspapers?79 A 1996 media-use survey of US. citizens showed nightly
network news audiences had slipped to 42 percent of households—from 48
percent in 1995 and 60 percent in 1993—that regularly watched one of three of
the nightly news broadcasts. Numbers were higher for local news content
viewing (65 percent), but this segment was slipping, too (72 percent in 1995).
Like newspaper circulation, radio news audience numbers have flattened
recently (a little more than 40 percent)?80

According to a recent study performed by Knowledge Networks/Statistical
Research, “households with lntemet access, and one or more PCs, are more
likely to rent VHS tapes and DVDs, buy pay-per-view programming and pay $50

or more per month for cable or satellite services.”281 While this data does not

 

279 Barans Introduction, 116.

28° Pew Research Cther for the People and the Pres_s_. “Falloff Greater for Young Adults and Computer
Users: TV News Viewership Declines,” I996. <http://www.people-press.org/mediarpt.htm>.

28' Nua lntemet Surveys. “Net and TV Are Not Mutually Exclusive,” Dec. 2001.

93

mun-i1?! '-

speak to media credibility, it does speak to media preference. Stating a specific
preference for one media over another speaks to the potential existence of media
bias. Certainly a study that would compare various media presentations might
very well cause people to evaluate or consume different kinds of media content
with varying attention or even dispensation, based on personal biases. This, too,
had the possibility of creating error term, making it necessary to ascertain its
existence with a number of questions dealing with media preference.

Thus, the two error terms this experiment elected to control for were

media bias and computer attitudes.

Recall and the News

Dual coding research has discovered what it refers to as a lag
phenomenon?‘32 When subjects are presented with repetitive memory items,
recall is better with more items between the two repetitive items than it is when
the items are presented successively.

This lag phenomenon was once identified but then mislabeled in mass
media research in a study on news audience recall literature. A researcher
performed a study testing audience recall of television news, in which he
incorrectly rejected dual coding theory in favor of a single-coding hypothesis?83

The reason, it seems, is that both he and his reviewers were unaware of the lag

 

<http://www.nua.ie/surveys/>.

282 Paivio, Allan, “Episodic Memory,” Mental Rgpresentations: A Dual Coding Approgc_h, New York.
Oxford University Press. 1990: 161; Paivio, Allan, “Dual Coding Theory: Retrospect and Current
Status,” Canadian Journal of Psychology, 45, 1991: 265.

283 Grimes, Tom, "Encoding TV News Messages into Memory," Journalism Mrly, 67.4, Winter

94

phenomenon discovered in dual coding theory research. His experiment showed
how words are remembered as pictures and pictures are remembered as words,
which is consistent with referential processing. He called this a "translation
phenomenon," but noted that it was more likely to occur 48 hours after viewing a
news story than immediately, which also supports the lag phenomenon reported
above and not a single-coding hypothesis that suggests all messages are stored
the same in memory.

As seen above, there is evidence of limited usage of dual coding theory in
mass media study. One reason might be that it is simply overlooked. Another
reason might be that, thus far, researchers have merely been concerned with the
end results—what actual recall was and how it compared to that for material in
other media—and less so with how it is stored in memory. Because publishers are
now placing news products on the Worid Wide Web, researchers must now
concern themselves with discovering if visual communication design principles
that effectively communicate news in the print and broadcast media will be able
to do the same thing within the computer medium or if new design principles are
needed.

First, though a few words that deal with the difference between recall and
recognition would be in order. News recognition is a person’s simple
acknowledgement, when presented with a basic description of a news story, that
they read or heard about the event or topic. By contrast, news recall is a person’s

ability to remember, when cued, the details of a particular story. 28"

 

1990:757-766; Kenney “Memory,” 52-53.
234 Price, Vincent, and Czilli, Edward J. “Modeling Patterns of News Recognition and Recall, Journal of

 

95

In an experiment in which fragments of drawings were used to help
subjects recreate drawings from memory, researchers effectively demonstrated
how components could be used to aid in recall of the whole.285 While it must be
granted this experiment dealt with imagery components, any subject will also
engage in referential processing in cued recall tasks. Retrieval from memory,
according to DCT, assumes that in either cued recall or recognition tasks, a
similarity match must be constructed between the evoked cue and the memory
trace?“5 Therefore, questions created for the purpose of discerning subjects’
news recall of recently read items should be derived as direme as possible from
the text. This suggests that access to the correct answer by means of cued recall
or recognition is probabilistic because the memory trace that is activated by the
stimulus may or may not be the one that might satisfy the condition of
correctness. Therefore, it behooves researchers to assure cues are as direct as
possible toward the required response. Accurate retrieval depends, to some
extent on associative factors, but more so on the similarity between the encoding
pattern that occurs when the event transpires and the retrieval cue offered when
a correct response sought.

Some news is so inherently memorable that care must be made to assure
it does not skew the results of this proposed study. For this reason, when
selecting stories to be used for study, researchers should err to the cautious by

selecting follow-up news items instead of breaking news. In addition, it shows the

 

Communication, Spring 1996: 55-78.
285 Bower, G.H., and Glass, A.L. “Structural Units and Redintegrative Power of Picture Fragments,”

Journal of Experimental Psychology: Human Learning and Memog, 2, 456-466.
2“ Paivio Mental, I46.

96

need for complementary visual information accompanying text and the dangers
of what might happen if it is too independent of supporting text.

The first example speaking to the need for care in visual selection is a
study that used dual coding theory to determine the role of verbal imagery in
recall of deviant news.287 The author defined deviant news as news stories that
were "schema-incongruent of events that are unexpected."288 Schemas are
knowledge structures based on abstractions or generalizations of our perception
of the world around us. We use these structures to evaluate new information.
Parts of these are our expectations of people and how things happen. Deviant
news, then, is news that violates expectations.289

A test of referential interconnection—the ability of words to evoke
pictures—the study sought to discover the power of the printed word. By singling
out deviant news and the powerful imagistic language that seems to go hand-in-
hand with these kinds of stories, the researcher tested reader recall of deviant
versus non-deviant news leads. But in doing so, the author recognized that
deviance itself might certainly affect recall, as would the degree of deviance
within various stories. To this end, he created two story categories that
juxtaposed imagery and deviance: high deviance/low imagery and low
deviance/high imagery.290 The advantage to using dual coding theory is that it

predicts a main effect only for imagery (both verbal and non-verbal) but not for

 

287 David, Prabu. "Role of Imagery in Recall of Deviant News," Journalism & Mass Communication
Qpa_rper_ly, 73.4, Winter 1996: 804-820.
2“ David, "Deviant," 807.
289 An example of a story that runs counter to expectations is the cliched example journalism educators use
to teach news values, "Man Bites Dog."
2"" Elsewhere in this article (805), the author alluded to the possibility that imagery may explain why
"concrete, event-driven articles have better appeal than abstract, issue-driven articles.

97

deviance, something born out by this study?91 Yet it is also interesting to note
that this research did not compare visual imagery to text, thus it did not test the
additive nature of pictures for recall. Subjects read only lead paragraphs of news
stories, which suggests the possibility of an additive recall effect for highly
imagistic language.
Effects of content and illustration

Paivio stated picture superiority over verbal systems under dual coding
theory is so sound that it is an ”empirical generalization."292 There is also
evidence that audiences gain more information from positive visual material than
from negative visual material. Extremely negative visuals, such as those that
depict excessive violence, are believed to disturb viewers so greatly that they can
be distracted from verbal messages?93 In a 1992 study on the effects of negative
TV news images on memory, researchers found that intense, vivid images
inhibited encoding of visual and verbal lnfonnation presented prior to it?94 For
this reason then, video used in this experiment was carefully chosen to assure it
was neither too vivid nor intense that its presence would affect normal recall
rates. Narrative that preceded the image presentation was also not remembered
as well as voice-over was.295 To avoid this pitfall, film used in the experiment has
voice-overs that were complementary to the information presented in the video

and not redundant to it.

 

Personalization. . .is another technique to make the story more concrete."
29' David “Deviant,” 818.
292 Paivio "Retrospect," 266.

293 Robinson, John P. and Levy, Mark R. The Main Source: Learning From Television News, Beverly
Hills: Sage. 1986.

29" Scott, Randall K. and Golf, David R., “How Excitation from Prior Programming Affects Television
News Recall,” Journalism Merly, 69.4, 1992, 615-620.

98

Yet there are still some bothersome questions about certain media,
especially in terms of what constitutes verbal and what constitutes images in
television. There is quite a bit of variability in the research literature. How can
researchers categorize spoken words on television as verbal and film footage as
visual? Since the audience is processing all information from television news
within a visual channel, isn't the kind of processing that is going on here more
likely associative processing of imagens? Yet a case can be made that film-
footage stories may be categorized as non-verbal, while news narratives can be
counted as verbal because the TV anchor presented this information in narrative
form, unaccompanied by graphic or visual illustration. This kind of story is
presented in verbal fashion. Secondly, there is empirical evidence of a picture
superiority effect between these two kinds of stories. In a study reported in 1991,
audience news comprehension was shown to be at its lowest when the entire
newscast consisted of verbal news narratives delivered by TV anchors. It
improved when newscasts consisted entirely of stories accompanied by film and
was best when film-footage stories were alternated with news narratives.296
From a dual coding perspective, then, this study suggests that in the broadcast
media, at least, broadcast information that allows for referential processing will
aid recall greater than associative processing alone. Still, for the purposes of this
study, a traditional dual coding perspective of what is imagery and what is verbal

information will be adopted.

 

29’ Price and Czilli “Modeling,” 55-78.
2% Brosius, Hans-Bemd. "Format Effects on Comprehension of Television News," Journalism Quarterly,
68.1, Fall 1991:396-401.

99

There is also evidence to suggest care should be taken to select visual
information that complements text information rather than visual information that
independently adds information not covered in the accompanying text?97 The
study showed that while graphics enhanced reader recall of information about the
geological makeup of both the San Francisco Bay and the Los Angeles areas,
they detracted from reader recall of lnfonnation available only from the text. This
study also suggests that unless the text prompts some measure of associative
processing and non-verbal lnfonnation is unable to create referential processing,
then the only processing possible is representational. Therefore, this study
reinforces the notion of the additive effect of picture superiority in dual coding
theory by showing what kind of processing occurs in the absence of additive
visual material.

It is well known in the industry that certain kinds of stories create more
interest than others. Researchers have looked at this and have confirmed the
findings: news based on domestic issues is far more likely to be recalled better
than lntemational news.298 Vivid visual components and localized stories are also
better recalled than others?99 In addition, others credit various media certain

advantages over the rest. Some mass media researchers believe television has

 

297 Rarnparasad, J. "Information Graphics in Newspapers," Newmr Research Journg, 12.4, 1991:92-
103.

298 Pew ReseaLh Center for the People & the Press. “Muted and Mixed Public Response to Peace in
Kosovo,” 1999. <http://www.people-press.org/>.

299 Graber, Doris A. “Seeing is Remembering: How Visuals Contribute to Learning in Television News,”
Journg of Communications, 40.3, 1990: 134-155; Gunter, Barrie. Poor Reception:
Misunderstanding and Forgetting Broadcast News, Hillsdale: Erlbaum. 1987; op. cit: Price and
Czilli, p. 58.

100

a greater hold over audiences who have lower cognitive ability or lack strong
interest in the news.3O°

Visual processing

In January 2000, a pair of communication researchers from the Netherlands used
a dual coding perspective to examine why it seemed children derived more
information from television news than from print news.301 They tested 192, 4‘"-
through 6"‘-graders abilities to recall news from television news stories, a text-
only version of the same stories, a print form supplemented with a photo and an
audio version. The results of this experiment were consistent with dual coding
theory—that recall was greatest for television news. TV’s superior recall, the
researchers noted, occurred when verbal information was supplemented with
redundant television images. TV stories were also recalled better than audio and
text-only versions and printed versions accompanied by photos.302 These results,
when considered within a dual coding theory framework, are consistent with the
theory. However, this study attempts to make the point that the television
superiority recall effect is due to redundant images. This distinction is made
based on early DCT research that showed subjects paired pictures and letter
combinations. But it fails to account for more recent DCT research, which notes
that image information is nested?03 When we invoke a visual lnfonnation

memory, it recalls visual lnfonnation sequentially to build a whole. For this

 

30° Neuman, W. Russell, Just Marion, R., and Crigler, Ann. N. Common Knowledge: News and the
Construction of Political Meaning, Chicago: University of Chicago Press, 1992.

30' Walma van der Molen, Juliette H. and van der Voort, Torn H.A., “The Impact of Television, Print and
Audio on Children’s Recall of the News: A Study of Three Alternative Explanations for the Dual
Coding Hypothesis,” Human Commpnication Research, 26.1, January 2000: 3-26.

302 Walma van der Molen “Impact,” 21.

”3 Paivio, “Mental,” 38-39.

 

 

101

reason, then, complementary visual and verbal information, when presented
together as part of a whole, are, in effect, a nested object, which is also
consistent with DCT theory.

Dual coding theory was first postulated in the early 1970s, a little more
than a decade before computers had diffused to the consumer market. Once
personal computers entered the market, the educational community quite
naturally embraced them as a useful Ieaming tool. Thus, educational researchers
began to apply dual coding theory to this new channel. Indeed, a parallel
development was occurring in journalism. In the personal computers early days,
it was used to create the news, but by the late 1980s, media pundits began to
foresee it as a delivery channel.

As early as 1992, research was being published that pointed to the
potential for the Web as an innovative news channel. One such study compared
news recall based on presentations by newspapers, television, radio and
computer screens; it featured a large diverse sample of university students drawn
from three US. higher education institutions.304 Subjects were able to recall news
from presentations read in newspapers and computers at much higher levels
than from broadcasts in two electronic media.

One point educators seem to know about intuitively is that younger
audiences process information far better in the visual modality than they do in
text or verbal ones. Fortunately, as we grow older, we seem to develop

complementary skills in assessing verbal symbolism.305 The educational

 

30‘ DeFleur, 1010-1022.
305 Simpson, Timothy J., "Message into the Medium: An Extension of Dual Coding Hypothesis," Proc.

102

literature speaks to the value of combining verbal and visual stimuli to assure the
most effective transfer of information.306 Thus, we see the educational
community looks to referential processing as the place to go for the best Ieaming
expenences.

Indeed, Paivio points out in his literature describing referential processing
that when we are confronted with a visual image alone (imagens), our minds call
up a word to describe it.307 Similarly, when we are confronted with verbal or
textual lnfonnation alone (logogens) we conjure a mental image from our
memory.308 Thus we see a case being made in both the theoretical and
educational literature for the necessity of using complementary imagens and
logogens together to deliver messages to the audience. Educators suggest using
imagens to convey ”striking" information, while reserving logogens for
presentations that describe impact, that describe logical progressions or even
persuasive messages?09 These are principles that precisely track the goals of
photojournalism, electronic news photography and taped interviews in radio
journalism, which strengthens the case for use of DCT in this proposed research.

Educational researchers are also using DCT to study hypermedia.31o One
reason is because of the multiple modalities DCT explains that people use to

create mental images in memory. Because people use multiple modalities, the

 

of lntemational Visual Literacy Association, 12-16 October 1994:255-263.

30" Op. cit., Simpson, "Message,": Mayer, RE, and Anderson, R.B. "Animations Need Narration: An
Experimental Test of Dual Coding Hypothesis," Journal of Educational Psychology, 83, 1991: 484-
490; Mayer, RE, and Anderson, R.B. "The Instructive Animation: Helping Students Build
Connections Between Words and Pictures in Multimedia Learning," Journal of Educational
Psychology, 84, 1992: 444-452.

3‘” Paivio “Mental,” 62-63.

3°“ Paivio “Mental,” 62-63.

309 Simpson "Medium," 259.

103

 

literature says, the multiple presentations hypermedia uses to present
information will only enhance Ieaming. Additionally, since people personalize
information with imagens and logogens drawn from their own memories, this, too,
will enhance Ieaming, according to DCT.311 Additionally, presentation of
information within multiple modalities will certainly facilitate associative
processing.

Research Questions

The experimental portion of this dissertation will suggest a means to test the
referential processing functions of memory that, according to DCT, occur as
people process imagens and logogens. The larger question deals with whether
or not the superior effect of pictures will enhance recall of information of both
verbal and non-verbal lnfonnation in news stories. More specific questions will
deal with superiority of recall of among various text and text and visual
presentations.

As should be the case, the literature will drive the formation of the
questions to be answered in the experimental portion of this work. Thus, there
are four questions posed for discovery in the experiment. The first of these will
test the picture superiority effect DCT research has reported within its
experimental literature, that the addition of images to text in any presentation

aids in memory creating more concrete images of those presentations:

 

3” Najjar “Multimedia,” 134.

3 ” Borsook, Terry K., and Higgenbotham-Wheat, Nancy. "A Psychology of Hypermedia: A Conceptual
Framework for R&D," Association for Educational Communications and Technology annual
meeting, 5-9 Feb. 1992:13-14.

104

H1: Subject recall scores for news accompanied by
visual presentations will be higher than news without
accompanying visual presentations, or:

H1: Mv> M:

The null hypothesis for this is the difference between the mean scores
collected in the recall tests for articles accompanied by visual presentation will be
less than or equal to those for articles without visual presentations, or:

H1: M., =Mt

The second hypothesis speaks to the differences between each of the
proposed visual presentation mode. These include text with still photos, text
accompanied by complementary (as opposed to redundant) multimedia video or

text accompanied by complementary multimedia audio and video content:

H2: Recall of online news stories consisting of text
accompanied by still photographs will be higher than
those consisting of text alone, or:

H2: Mt<Mp

The alternative hypothesis for H2 is:

H2: Mt>Mp

105

H3: Recall of online stories consisting of text
accompanied by streamed (multimedia) video will be
higher than both stories accompanied by still
photography or stories consisting of text alone 01".

H3: M,<MV and Mp<Mv

The alternative hypotheses for H3 are:

H4: Recall of online stories consisting of text
accompanied by streamed multimedia video will be
higher than stories accompanied by still photography,
streamed video or stories consisting of text alone, or:
H4: Mma>va and Mwna>Mp and Mma>Mt

The alternative hypotheses for H4 are:

H4: vaaSva and vaagMp and vaafiM,

This last hypothesis suggests a linear increase in recall scores with the
lowest recall being expected from text alone, and the highest recall being
expected with the multimedia-intensive mode of text, video and embedded audio.
Thus, the text alone score is expected to be lower than the recall scores of all
other presentation modes. The text with photo recall score is expected to be

higher than the recall scores of text alone, but lower than those achieved by the

106

F-"
M 2

stories accompanied by multimedia video and multimedia video and embedded
audio. Similarly, the text and multimedia video mode recall scores are expected
to be lower than the text and multimedia video and embedded audio mode recall
scores. Lastly, the text and multimedia video and embedded audio mode recall
scores are expected to be the highest recall scores collected in this experiment.
Conclusions

All the elements are in place to support multimedia news sites. There are
a plethora of new access channels available; there are enough lntemet-ready,
low-cost computers available that the time is right for a new mass audience to
emerge, equipped with the necessary tools to access it; streaming has come to
the fore as an interim multimedia technology that will put audio and video into the
hands of even those whose only access is a dialup connection; and lntemet 2
promises sufficient innovation to diminish the negative effects of multimedia
access that lacks the bandwidth to handle it?12

With so many online news media sites now on the Web, clearly, the time
has passed to ask whether news should be offered on the Web. Instead, the
more pertinent question is, how should news he presented on the Web. The
experiment detailed in the following pages looks closely at that question with an
eye toward recall. What is the best way to present news on the Web, if recall of
the news is the goal?

No research has looked at news media in a native environment. That is,

what little research has been done on this question has mimicked online news

 

312 CNN Sci-Tech. Kellan, Ann,.“Intemet 2 Holds Promise of Technological Leap,” 24 Feb. 1999.
<http://www.cnn.com/TECI-I/computing/9902/24/internet2/>.

107

 

venues. This is not a limitation of the researchers and their research; rather, it is
a limitation imposed by the absence of a large number of multimedia news sites
and the desire by their managers to present news that is not produced solely by
the parent firm. Thus, radio stations create so-called Web-radio sites that
produce audio output alone?” TV stations create so-called Web-TV sites that
produce multimedia video output alone;314 and print media firms produce so-
called online newspapers that produce print, photo and graphical illustrations
alone.315

This research project will look at how well multimedia lnfonnation serves
audiences when presented on the same site together. However, Instead of
looking at how the synergy of all four presentation modes work together, this
project takes the logical first step of evaluating the component presentation
modes’ individual abilities to engender news recall by measuring it and applying
a traditional quantitative means (supported by the literature) of evaluating that in
the form of analysis of variance.

There have been a great deal of studies looking at computer-based
classroom teaching tools that compare classroom performance of those taught
using conventional methods versus those taught using computer-based tools. As
presented above, computer-based teaching methodologies provide a clear

advantage to those taught with computer-based tools. Based on these results, it

 

3’3 e.g., Netradio.com. 2001. <http://www.netradio.com/>.

3 1" e.g., ABCNews.com. 2001.
<http://www.abcnews.go.com/sections/us/video_index/video_index.html>.
3'5 e.g., LansingStateJoumalcom. 2001. <http://www.1ansingstatejournal.com/>.

108

might easily be surmised that news recall might be improved when viewed as
computer media compared to more traditional, offline news products.

Reading and navigation research points to the less memor— intensive
nature of the news reading task. In addition, research has shown reading
comprehension on computers is similar, if not better, than reading text-based
documents. More recent research shows a clear recall advantage for visual
presentation of the news, as long as the material is redundant. But what if the
material presented in the news site is not redundant, but is complementary? That
is another unique question this research will address. Will news recall be higher
for visual information, as current research with redundant material shows, and as
dual coding theory also suggests? Or will it be lower, as some attention theory-
based research suggests?

Lastly, it is important to consider the theoretical perspective this project
proposes to create. If we believe that theories are indeed the roadmaps by which
we plot the results of our research, then it is important to identify theories that will

serve both traditional and evolving media formats.

109

Chapter 3

METHODS

This study will assess the values of various ways to deliver online news
content. It will compare average recall rates of news items that are text-only, text
accompanied by a single still photograph, text accompanied by a single digitized
video segment and text accompanied by embedded audio and a single, digitized
video segment. Using dual coding theory as a guide, it was hypothesized the
highest recall rates would be generated by those multimedia presentation modes
that used a combination of text and images to deliver the message.

The study portion of this dissertation was constructed along traditional
methodological lines already used by previous studies in the dual coding theory
experimental literature. Subject pool size, experimental structure and data
analyses tracked experiments conducted over the nearly 20-year history of this
theory's existence.

Sample

Subjects who participated in this study were undergraduate students,
aged 18-25, recruited from a mass media and society lecture class at Michigan
State University. They were each offered extra credit toward their class grades
for participating in the study. An estimate of required sample size was computed,

using the following formula:

110

FIGURE 2: Estimate of Required Sample Size Formula

Ta Sd

projected difference

 

n = + K
In this computation, which is derived from hypothesis testing, n is the
required sample size for the Analysis of Variance (ANOVA), the statistical
procedure used to analyze the data collected in this study. Researchers use
significance to test hypotheses to see if the same results would occur if a
different approach were used to seek the same results. Ta (or T—alpha) is the
significance level of the test, which in this case is 0.05 (5%). “In the research
literature, results of statistical tests are usually reported using the p-value. The p-
value provides an objective measure of the strength of evidence, which the data
supplies in favor of the null hypothesis. It is the probability of getting a result as
extreme or more extreme than the one observed if the proposed null hypothesis
is correct.”316 Smaller p-values suggest the null hypothesis is incorrect. Sd is the
standard deviation, which In this case, was set to .25. Sd was set to .25 as a
conservative standard deviation unit. One standard deviation unit away from the
mean within a normal curve represents 68 percent of the population. The
literature predicted mean recall scores to occur within a relatively narrow range of
40 to 55 percent. It was for this reason this conservative standard deviation was
used. The projected difference was derived from the literature, which ranged

from five to eight percent. K is the number of confounders identified. Similar to

 

3‘6 Hyp_erstat Online. “Logic of Hypothesis Testing,” Lane, David M. 2002-

111

fl!

the projected difference, potential confounders were derived from the literature.
In the end, 105 subjects were recruited and 95 actually completed the study.
Research instrument

Subjects accessed Web pages designed to mimic the look and feel of an
online news site. Originally designed to appear as if it were a local online
newspaper, it was named the Lansing Eagle (See Appendix C for a sample page
template). The design of the instrument was based on a survey of Web designs
of 160 randomly selected U.S. online newspaper sites, which at the time317
represented approximately 10 percent of all online newspapers on the Web.
First, the survey determined if sites had an index (a table of contents) and if they
did, how and in what position on the page they displayed it. Nearly all those
analyzed—98 percent—displayed their site index in a banner device somewhere
on the page. The remaining two percent simply listed their contents within a
column.

The Lansing Eagle featured a banner identifying the publication across the
top. It used a three-column format, featuring non-functional editorial and
advertising links on the outer columns and news copy in the middle column. This
design was consistent with approximately 68 percent of the existing online news
sites of the time, as determined by the design survey (See Appendix B for the
tabulated results of this survey).

Stories appeared in each of the four proposed formats or levels--text alone

(T), text and a still photograph (T + PH), text accompanied by a multimedia video

 

<http://davidm1ane.com/hyperstat/).
317The survey of online news site designs was conducted during the summer of 1998.

112

clip embedded within the story (T + V) and text with both embedded multimedia
video and embedded audio (T + V + A). Again, this range of presentation modes
represented the full range of available content delivery modes being used on the
Web at the time.

News copy and still photography used in the study were taken from both
the New York Times and Washington Post online news sites, while television
news stories were collected from CNN News daily broadcasts. Three TV news
segments were first taped, using a VCR, from a cable television broadcast, then
converted to computer files using Apple Macintosh computers. Two of the
segments were then converted into QuickTime movie formats, using QuickTime
Pro. The third film segment was used to create the embedded audio soundtrack.
The segment used for this piece was a short clip featuring a taped interview
between a reporter and a seated man. The sound track from this clip was saved
into a .wav file, using SoundEdit then compressed into an audio digital file, using
Adobe Premiere.

Delivering all this multimedia material through a single campus network,
via the World Wide Web, to even 18 subjects at a time (the limitation imposed by
seating in each of the computer labs used for the experiment), would have been
more than any network of the day could sustain. The bandwidth required would
have taxed the small lab, which, at the time, used the Appleshare network
protocol. Thus it was necessary to store all the Web documents, the Java scripts,
QuickTime Movies, the Web pages, the still photography, the text and the .wav

files locally, within a single file, on each individual computer desktop used in the

113

study. Stored locally, file access did not move through networks and create
slowdowns. Instead, subjects could view news material at speeds they might
normally experience when viewing online news as individuals.

Computers subjects used to view the news material were PowerPCs, each
equipped with multimedia capabilities and hypertext browsers, QuickTime Movie
Players and headphones. Subjects used headphones to listen to sound from the
experiment's news material because subject matter presentation was
randomized with a Java script.318 Thus, subjects sitting next to each other but
listening to multimedia content from different articles would be less likely to be
disturbed by messages from other, different articles being listened to at
neighboring stations. In addition, subjects used headphones because
researchers hoped to create an optimal online news reading experience by
eliminating as much competing noise from the lab environment as possible.

As mentioned earlier, the online news articles were presented to subjects
in four different viewing orders, randomized at its particular station by a Java
script embedded in the Web sites and stored on each computer’s desktop along
with the HTML code for the Web page. This script caused the computer to
present the four news story pages of the Lansing Eagle in random order to each
viewer. This practice is called counterbalancing and was done to avoid
confounding the variables. Confounding occurs when variables vary together in

such a way that it becomes impossible to determine which variable is responsible

 

3 '8 My former student, Andrew Kurtz, then a designer employed by Michigan State University’s
Communication Technology Laboratory, East Lansing, Michigan, graciously provided coding for
this Java script.

114

IM,’

for an observed effect.319 In this case, if every subject viewed the multimedia
news stories in the same order, the presentation mode and the viewing order
would be confounded. There would be no way to know if the various presentation
modes cause recall to be higher or lower, or if it was simply the most recently
read/viewed mode that was remembered best. Counterbalancing—randomly
varying the order in which subjects view the pages of the Web site—was used to
reduce the risk that the order of presentation of the material might influence how
well or poorly subjects recalled information they processed from articles.
Viewing the news

Subjects were asked to refrain from reading, listening or viewing any news for 48
hours prior to their participation in this experiment. This action attempted to
minimize their familiarity with the day's news events and the possibility that this
familiarity might skew the experiment's results. Subjects were then scheduled
into seven groups of approximately 16 subjects each to view all four levels of
news presentation modes.

When subjects sat down at each individual computer station, they were
confronted by a white computer screen with a single line of text, asking them to
wait for instructions before mouse-clicking on the link that provided them access
to the online news site. Subjects were not told, prior to the experiment, that the
questions they would be asked would be about the stories. Rather, they were
instructed to read the news material as they normally would any material on the
Web. The goal here was for subjects to pay the normal amount of attention they

paid when viewing material on the Web. The strategy worked for most subjects.

 

3 '9 Lane, "Counterbalancing."

115

A number of subjects noted in the verbal comment section at the end of the
questionnaire that their recall scores might have been higher had they been
informed of the questions they would be asked.

Once subjects clicked on the link to the news site, they were presented
with what appeared to them to be the first page of the site. In fact, they were
each confronted with one of the four pages presented in random order. Stories
appeared within a form that featured the site's banner across the top of each
page, with news story and advertising banners along both sides of each page.
According to instructions they received prior to accessing the news on the site,
subjects read through each of the four Web pages then stopped when confronted
with the white page with the single line of text. Then they filled out a
questionnaire that asked a number of questions dealing with the facts of the story
as well as a number that attempted to identify potential interaction effects.
Questions were carefully worded so that their complexity was similar from one
question to another. In general, their length was limited to one of below 30 words.
Each was cast in an interrogative form and addressed itself directly to facts
contained within single sentences or single graphical or audio items within each
story.

Once subjects finished filling out the questionnaire, they were excused.
Each new group of subjects entered the labs and chose lab stations only after
the previous group had left the room. This reduced the possibility of subjects
sharing lnfonnation about the experiment with those about to enter the lab to

participate.

116

‘I

It should also be pointed out at this time that a conscious decision was
made to assure the various multimedia portions of the article were, in no way,
repetitive or redundant with the text or other multimedia components of the same
story. With stories constructed so that redundancy did not exist, it freed the
researcher to be able to test the efficacy of each presentation mode with little
fear that the recall response might be skewed by the presence of redundant
material.

Recall questionnaire

Questionnaires’ first sheet consisted of a consent form required by
Michigan State University’s Human Subjects Committee. In it, subjects were
informed they were being asked to “participate in a study that deals with how
people process news information from Web sites.” They were also informed that
their participation in this study would not them some extra credit in their TC 100,
Introduction to Mass Media class and were reminded they had agreed not to read
newspapers nor listen to radio or TV newscasts for 48 hours prior to the study.
The remainder of the consent form was filled with boilerplate language provided
by the university assuring them their participation in the study was voluntary and
all individual responses would be kept private. Lastly, the consent form asked
that subjects confirm that they were aged 18 years or older.

The questionnaire was subdivided into two sections (See Appendix A for a
sample questionnaire copy). The first section consisted of 20 short-answer,
objective, close-ended questions dealing with the facts of each story; four to six

questions each dealt with content on a story-by-story basis. It was reasoned that

117

questions should be grouped according to the stories, rather than be randomized
because dual coding theory specifies that text information is stored sequentially
and image information is stored in a nested fashion. Whether remembered data
from the stories was either imaginal or textual or both, the goal of the study was
to determine if subjects retained information. Thus it was essential to test
subjects in a fashion that would parallel their memory structure.

Care was taken to assure that each question dealt specifically with
information that was wholly contained within a single story element. For example,
there was a particular question that dealt with a filmed segment about how some
stolen money was used. There was also another question about the same story,
dealing specifically with the substance of an audio interview.

The second section of questions consisted of 20 questions aimed at
eliciting demographic and attitudinal information that might create interaction
effects within the data analysis. Interaction occurs when the effect of one
independent variable differs because of the level of the other variable or if the
combination of the variables leads to unanticipated results.320

Question 1 asked subjects to specify if they were male or female. At the
time this study was performed in late 1997, use of online venues was a male-
dominated activity. It wasn’t until several years later that those numbers began to
approximate real population norms. Question 2 asked subjects to stipulate their
race. The idea here was to cross match responses against ethnic backgrounds to

see if any behaviors were peculiar to particular ethnic backgrounds. Questions 3

and 4 asked if subjects’ primary language was English or if they had any trouble

118

reading the English language. Since all articles, film and audio were done in
English, it would be important to determine which subjects might have had
difficulty with reading and understanding the material. Questions 5-7 probed
subjects about any other physical disability (hearing, sight, Ieaming disability)
that might interfere with their reading, listening, watching or understanding the
material. Question 8 asked if there was any “particular problem” with completing
the experiment. This question was asked in an attempt to detect any unforeseen
problems with either the research instrument or the questionnaire.

Questions 9-15 attempted to discern negative or positive attitudes subject
may have held about the use of computers in general. At the time the research
was conducted, there were a number of studies that showed the presence of
negative attitudes about computers that might skew the results of studies that
involved computer use?21 Presented as Lickert scale questions, the queries were
selected, based on their reliability scores, from a number of articles that listed
various attitudinal checks on computer use among various demographic
groups.322

The last five questions on the questionnaire were designed by the
researcher to determine if subjects held any biases for or against various media.

Four questions presented subjects with a Likert scale asking them if they

 

32° Lane, "Interaction."
32' e.g., Loyd, Brenda H. and Loyd, Douglas E. "The Reliability and Validity of an Instrument for the
Assessment of Computer Attitudes," Educational and Psychological Measurement, 45, 1985: 903-
908.
322Loyd, "Reliability, 903-908;Loyd, Brenda H. and Gressard, Clarice. "The Effects of Sex, Age and
Computer Experience on Computer Attitudes," AEDS Journal, 17, 1984: 67-77; Loyd and Gressard.
"Reliability and Factorial Validity of Computer Attitude Scales," Educational and Psychological
Measurement, 44, 1984: 501-505; Koslowsky, Meni, Lazar, Aryeh and Hoffman, Michael.
”Validating an Attitude Toward Computer Scale," Educationpl and Psychological Measuremen_t, 48,

119

considered each kind of media trustworthy or untnrstworthy. The fifth question
asked from which medium did they get most of their news: print media, radio, TV
or the lntemet.

Consideration was also given as to whether to administer the
questionnaire online or not. It was decided to eschew an online questionnaire
since most students of the day were accustomed to taking exams and quizzes
that were paper-based. The goal of this particular portion of the experiment was
to have subjects focus on recalling the information in the Web site and to not
allow things like interactivity and novelty interfere or improve that memory
process. Granted, interactivity and novelty certainly played a role in the subjects”
reading of the new content; but in choosing a traditional means to test recall, the
possibility of the results being tainted by the novelty or interactivity of the testing
instrument were eliminated so that results of that effect might be linked directly
back to the Web media and not the testing instrument.

Data analysis

The model for this study is as follows:

Recall = f(modes, demographics, confounders) - error

Recall was measured by an objective test that determined the amount of
information remembered from the various articles. It is a function of the text and
images presented in each presentation mode and the subjects’ minds’ abilities or
inabilities to process what they read into short-term memory. The modes in this

model include a text-only article, an article accompanied by a single photo, and

 

1988: 517-520.

120

article accompanied by a single news video and an article accompanied by a
single news video and an embedded audio track.

Data collected in this study was analyzed using Analysis of Variance
(ANOVA), according to a within-subjects design. Simply put, each subject was
exposed to each of the four presentation modes. The reason for this design is
that subjects are inevitably different from each other. Some might have been
better readers than others are or some might have been better acclimated to the
use of computers than others. In statistical analysis, all differences among
subjects’ scores that cannot be explained by the experimental treatment are said
to be "error," though some scholars say it would be more accurately defined as
"unexplained variation."323 It is not error, in the traditional sense that it is due to
some kind of mistake. Rather, within this kind of experimental design, error is
used to explain whatever differences exist among the subjects can be measured
and separately identified.

The within-subjects design was used to reduce error. In a within-subjects
design, the same subjects are tested in each condition, which allows differences
among subjects to be measured and separated from error. This design also
increases the power of the results from the study. Power is defined as ”the
probability of rejecting a false null hypothesis."324 Experiments with little or no
power are generally inconclusive. Since each subject viewed each of the four
presentation modes, the statistical measures are more capable of explaining

individual differences among subjects in the experiment.

 

32’ HyperStat Online. Lane, David M., "Partitioning the Sums 0f Squares,"

<http://www.ruf.rice.edu/~lane/hyperstat>.

121

Historically, DCT experimental literature shows typical predicted
differences among treatments to be from 5 to 8 percent. Expected mean scores
generally range from 40 to 55 percent and with the range seeming to vary from

25 percent to 60 percent.

 

32’ Lane, "Power."

122

Chapter 4

RESULTS

The experimental portion of this dissertation was designed to use the
theoretical framework proposed earlier in this work to test the representational,
referential and associative processing functions of memory that, according to
DCT, occur as people process imagens and logogens. The larger question deals
with whether or not the superior effect of pictures and sound enhances recall of
information of both verbal and non-verbal information in news stories and to what
degree. More specific questions deal with superiority of recall among various
text, text and visual and text and visual and audio presentation modes.

A total of 107 subjects viewed the news presented in the Lansing Eagle
Web news site. Eight of the subjects questionnaires had to be discarded owing to
them either not following directions to complete the study correctly or for failing to
fill out the questionnaire correctly. The total number of subjects who correctly
completed the study and filled out questionnaires was 95. Overall, 51 were males
(53.7%), and 44 were females (46.3%), which roughly approximates the sexual
characteristics of U.S. population data. Data collected were analyzed using a

one-way analysis of variance.

123

Table 1: Recall Score Means and Standard Deviations

 

 

 

 

 

 

 

 

 

Mean Std. Deviation
MM Video and
Embedded Audio '589 289
MM Video, Text .312 .267
Text and Photo .314 .259
Text alone .434 .310

 

The table above shows the mean recall scores and standard deviations.
The multimedia video and embedded audio treatment had the highest recall
score; the text-only treatment had the next highest average recall scores; and
both the text with still photos and text with multimedia video treatments had
nearly identical average recall scores and ranked third in the experiment. The
mean recall score for the multimedia video and embedded audio presentation
mode was 58.9% with a standard deviation of .289. The next highest average
recall score was for the text-only presentation mode, with a mean score of 43.4%
and a standard deviation of .310. Both single-item treatments netted nearly
identical recall rates, which was consistent with dual coding theory. Text with a
still photo had an average recall score of 31.2% and a standard deviation of .267
and text with multimedia video had an average recall score of 31.4% with a
standard deviation of .259.

Further analysis, using contrasts of means, showed that all the means
were different from each other, save multimedia video and still photography, both
of which were considered the same. Though this runs counter to the proposed
hypothesis for this study, it is theoretically consistent with dual coding theory.

This result will be discussed more thoroughly in the discussion chapter of this

124

dissertation. All the recall score means were significant at the .95 level. In
addition, it should be noted that each of the modes had p-values of .000.

Table 2: Comparison of Means

 

 

 

 

 

 

 

 

 

Source F Significance

MM Video, Embedded Audio, Text

Vs. Text Only 25'043 '000
MM Video, Embedded Audio, Text Vs. Still

Photo and Text 16270 “000
MM Video, Embedded Audio, Text vs. MM

Video and Text 15'455 '000
Text Only Vs. Still Photo, Text 66.16 .000
Still Photo and Text Vs. MM Video and Text .006 .941

 

The first hypothesis for this study tested the picture superiority effect DCT

research has reported within its experimental literature:

H1: Subject recall scores for news accompanied by
visual presentations will be higher than news without
accompanying visual presentations, or:

H1: Mv>M(

The alternative hypothesis for this is the difference between the mean
scores collected in the recall tests for articles accompanied by visual
presentation predicted the scores would be no different to those for articles

without visual presentations, or:

H1: MV=Mt

125

 

 

This research failed to unequivocally disprove the H1 alternative
hypothesis. Mean recall for the multimedia video and embedded audio plus text
page was the highest of all tested presentation modes at 58.9%. Yet recall of
news presented in multimedia video plus text (31.2%) and photography plus text
(31.4%) modes were both below the text-only recall score (43.4%).

The second hypothesis spoke to the differences between each of the

proposed visual presentation modes:

H2: Recall of online news stories consisting of text
accompanied by still photographs will be higher than
those consisting of text alone, or:

H2: Mf<Mp

The H2 alternative hypothesis, M,>Mp, was not disproved. Average recall
scores for the story accompanied by a still photo (31.4%) were lower than the
average recall score for the story presented in text-alone mode. This result runs
counter to DCT literature, which suggests higher recall scores should have
resulted from the text accompanied by a photo.

As part of the analysis of variance, 20 questions discussed in the
previous chapter were included in the questionnaire in an attempt to identify
potential interaction effects. None of these interaction affects tested significant

at the .95 level, and thus, they were excluded from any further consideration.

126

All of this suggests some kind of experimental design error on the part of
the researcher. One possible explanation for this could be that the text-alone
story consisted of breaking news, while the other three stories consisted of
follow-up news stories. This will be expanded upon in the Discussion section of

this dissertation.

H3: Recall of online stories consisting of text
accompanied by streamed video will be higher than
both stories accompanied by still photography or
stories consisting of text alone or".
H3: M,<M., and Mp<Mv
This research also failed to disprove the H3 alternative hypotheses, Mt>Mv
and Mp>Mv. Average recall of material in the text-alone presentation mode was
higher than average recall of stories in both the still-photo-with-text and
multimedia-with-text presentation modes. In addition, both the still-photo-with-text
mode and the multimedia-with-text mode average recall scores were nearly
equal. Again, this particular finding is likely a result of the same experimental
design error described in the section above, and will be discussed in more detail
in the Discussion section of this dissertation.
Average recall scores for the stilI-photography-with-text presentation mode
(31.4%) was virtually the same to average recall for the multimedia-video-with-
text (31.2%) presentation mode. Though these results are inconsistent with the

hypothesis, they are consistent with the theory. DCT research suggests single

127

visual elements accompanying stories would not these kinds of similar scores.
Thus, this research points to a conclusion supported by the theory that single
visual elements, will return the same recall scores simply, because they are
single visual elements accompanying text. Whether these image components are

moving or still does not appear to affect recall.

H4: Recall of online stories consisting of text
accompanied by streamed multimedia video and

embedded audio will be higher than stories
accompanied by still photography, streamed video or

stories consisting of text alone, or:

H4: Mvma>vaand Mvma>Mp and Mvma>M,

The alternative hypothesis for H4 would be mathematically expressed as:
H4: vaangv and vaagMp and vaasM,

As expressed, this research did not unequivocally disprove the H4 null
hypothesis. The hypothesis that news presented with text and embedded audio
and multimedia video would have greater recall scores than all other treatments
was proven. However, the linear progression that suggested multimedia video
with text, still photography with text and text-alone presentation modes would
have recall scores in descending order was not borne out.

In the end, the presentation mode with the greatest number of multimedia
modes accompanying the text news stories had the highest recall score. This

result is consistent with the theory. Additionally, the fact that the stories

128

accompanied with single visual elements had equal scores is also consistent with
the theory, though not with the hypotheses. Finally, the text-alone presentation’s
second highest recall score is not consistent with the theory nor is it consistent
with previous experiments. The text-alone story was the only story of the four
presented in the experiment that was breaking news. For this reason, it is
believed this resulted from a design error. Had the research attempted to control
for interest shown in all four stories, it is quite possible this explanation might
have been born out with facts. In the absence of this question, the reason for why

recall occurred in the all-text mode can only be surmised to be content novelty.

129

Chapter 5

DISCUSSION

Introduction

The experimental data collected in this trial show some remarkable lnfonnation
that speaks to an almost preordained success for multimedia news. The theory
used to guide this research, dual coding theory, states that the most concrete
memories created when people process imaged information along with text
information.

In this particular study, 18- to 25-year-old college students read an online
news publication in a laboratory environment that had four Web news
presentation modes: text-only, text with photos, text with video and text with
video and embedded audio. Immediately following their reading of this material,
the students were tested on their recall of information presented in the articles.

Data collected was analyzed using analysis of variance. The total number
of subjects who correctly completed the study and filled out questionnaires was
95. Overall, 51 were males (53.7%), and 44 were females (46.3%), which roughly
approximates the sexual characteristics of the U.S. population. The results of the
study showed that the highest recall rate was nearly 60 percent for the
multimedia video and embedded audio, plus text treatment. The next highest
recall rate was generated by the text-only treatment at a little more than 43
percent, followed by nearly identical scores fractionally higher than 31 percent,

for both the still photo plus text mode and the multimedia video plus text mode.

130

Mixed results

These results represent a mixed bag of success when considered in
context of dual coding theory. Since the early 1970s, when dual coding theory
was first drafted, a so-called “picture superiority effect on recall” has been
documented. Based on the combined experimental experience of nearly three
decades of research, it was not surprising to see the multimedia video and
embedded audio mode to have the highest recall rate.

The inclusion of two separate imagery modes in this treatment virtually
guaranteed this treatment would score the highest, according to DCT. What was
perhaps the most telling part of this was that neither the text, the video nor the
embedded audio contained any redundant material; none of the information in
the article was repeated to subjects in any way in this mode, nor in any other
mode. This was done deliberately to assure no recall score results were skewed
by reiteration that would not normally occur in real-life news reading.

With an average recall score nearly 20 percent higher than the nearest
scoring mode, clearly, the embedded audio and multimedia video treatment
scored the highest recall. This is not surprising when the additive effects
described by DCT research are considered. Called the additivity hypothesis?25 it
was first proposed to explain the superior ability of concrete words to be recalled
over subjects’ ability to recall abstract words and later extended to explain a
similar picture superiority effect. For example, DCT predicts symbolic
comparisons of concrete objects will be faster with pictures than with words

because the imagens are accessed more quickly when pictures serve as

131

stimuli?26 In episodic memory, “the mnemonic contribution of the non-verbal
(visual-image) trace component consistently turned out to be greater than that of
the verbal component.”327

Therefore, it should not be a surprise that the highest scoring treatment in
this study had the greatest amount of recall among the subjects. The fact that
these two image components—the embedded audio and the multimedia video—
were complementary??? and were not repetitive, enabled the experiment to avoid
the repetition lag paradigm of free recall. Repetition lag refers to the
phenomenon in which recall of repeated items in a list increase with the number
of items that intervene between repeated events?29 Images separated by blocks
of text might call into question the recall subjects exhibited for the information
contained in the multimedia video and embedded audio mode story. By making
the information contained in the two imagery components of the story non-
repetitive, neither item was repeated and could therefore not be the cause of the
repetition lag.

There were two surprises in the experimental data. First was the fact that
the text—only treatment scored second highest in the recall test. This runs counter

to theoretical expectations. The text-only treatment should have scored the

lowest, since it was not accompanied by any image information whatsoever.

 

’25 Paivio “Retrospect,” 265.

326 Paivio, Allan. “Manipulation and Use of Representational Information,” Mental RepreseMnfl
D_u_al Coding ApproaLh, New York: Oxford University Press, 1990: 180.

327 Paivio Mental, 142.

328The material presented in video and audio forms in no way repeated information presented in the
article’s text.

”9 Paivio “Retrospect,” 265.

132

lire-“rm! -

There is one possible explanation for this occurrence. This article was
derived from information originally published in a New York Times article about
negotiations between then-U.N. Secretary General Anan Kofi and Libyan
President Muammar Qadafi to release prisoners in that country suspected to be
terrorists responsible for blowing up a commercial passenger plane in the skies
of Ireland the previous year. The error committed here was that unlike the articles
presented in the other three modes, this one was the only breaking news story.
The other three were all follow-ups to stories already broken and part of
continuing coverage. For this reason, it is believed, the story garnered the
second highest recall score.

News, especially breaking news is quite popular. In fact, one of the more
interesting outgrovvths of online news is how often stories may now be updated
within a single day. For the first time since it’s inception, broadcast news is
experiencing significant competition in an area it heretofore had no competition
in—its ability to update stories. Unlike print media, electronic media could break
into any part of its program day to bring new lnfonnation to its audience. Now
Web media are capable of the same thing. Since the audience for this media is
asynchronous—they “tune in” and consume news on their own schedules, and
are not limited to the editions or programs times—online journalists have felt free
to update the news whenever events dictate.

The novelty of news is what attracts us to it. Each day, it is involves a

unique panoply of events and people, as unpredictable as the next day’s

133

direction of the wind. As readers and daily consumers of this new lnfonnation, we
pay it our highest homage by reading it first and giving it our greatest attention.

In this experiment, newsreaders did precisely that—to a point. Recall of
two stories that, according to DCT, should have featured higher scores than the
text-only treatment, did not outscore the text-only story. And yet, the one story
that had the most multimedia components, and was still merely a follow-up news
story, outscored the news of the day.

While it may be argued that this story was an lntemational news story,
and, as such, would generally score lower in reader interest than anything with a
domestic slant, the margin by which this story outscored all others clearly points
to an effect that goes way beyond mere news interest. The multimedia story
outscored the text-only story by nearly a 20 percent difference and the nearest
multimedia competitors—text and still photos and text and video—by almost 30
percent.

Another point should be raised here, too. Why did both the text with
photos and text with multimedia video score nearly identical recall scores? The
answer to this question is one that is consistent with the theory. DCT
experimental literature says text, when accompanied by images, will be
remembered better. Consistent with the high score of the multimedia video and
embedded audio mode, the theory bears out that the more imagery data that
accompanies text, the higher the recall. Consequently, the two treatments that
were accompanied by only one image each both scored nearly identical scores.

This, too, is consistent with the theory, since the theory does not accord greater

134

power to either still photography or video, in terms of making memory of text
more concrete.

Though it can be legitimately argued, from a statistical analysis standpoint,
that reader interest in various stories may certainly impact on recall, it remains to
be seen if that could actually be controlled for. How could a question be framed
to get at that particular interaction that did not strike at the heart of what is reader
recall? After all, interest is what draws us into a story, what keeps reading and
what sparks our memories into recalling. Therefore, in an attempt to account for
it, we might actually find ourselves in the unenviable position of trying to control
for our dependent variable.

In sum then, the general ranking of most multimedia as the highest recall
and the two single-element image and text modes gaining near identical scores
were consistent with what DCT predicted. The overall ranking, in order from
highest amount of accompanying imagery to lowest amount was not borne out,
but can be directly attributed to researcher error.

Conclusions

There are so many Web media sites today that it is easy, from a news consumer
perspective to view the Web as a separate channel. Despite the relative ease
with which news audiences are embracing the Web, it is clear that news content
providers are not approaching this channel with the same reckless abandon
audiences are. There are several media databases claiming to index more than

15,000 online publications.330 More than 5,300 of these are newspapers?31 Yet

 

”’0 One such database is NewsStation.com <http://www.newsstation.com>.
33 1 This was the number of online newspapers featured on E&P Online’s MediaOnline directory in

135

‘ 9 Luci“.-.

relatively few of these sites are using or even experimenting with multimedia
news presentations.

Two of the more advanced sites using this form are NewYorkTimes. com
and CNN.com. Both these sites began this multimedia experimentation little more
than two years ago. Since then, MSN.com is the only major news outlet to adopt
a similar delivery strategy.

One reason for this is that while the Web has emerged as a major new
commercial channel, the news industry has failed to generate as much money
from it as it has from its offline products. For this reason, they seem reluctant to
invest money in online news products until they see greater returns from them in
the form of advertising revenues or access fees. Another is that traditional media
rarely, if ever, combine forces to present news jointly. They are more
accustomed to competing with each other to deliver news faster, more
comprehensively to the audiences they serve.

But Web media represent a confluence of news product delivery. As this
study has shown, online news producers must reorient their thinking, in terms of
news delivery, if maximizing recall is a desired goal. Web sites will compete for
audience share with multimedia product offerings, not single medium offerings,
as it has been done in the past. Additionally, if multimedia increases recall of
editorial material, it should similarly increase recall of advertising material in
consumers’ minds. Differences between print and broadcast products will

disappear as the weaknesses and strengths of each channel converge to support

 

December 2001.

136

‘nmmt _

each other and create a more complete and efficient news product on a wholly
separate channel.

What appears to baffle most new media venture management teams,
especially those employed by firms whose experience lies chiefly in print, is how
to fashion the delivery vehicle. With the arrival of the World Wide Web on the
populist scene in 1993 or so, media firms began looking to the Web as a vehicle
for content delivery. Now with worldwide Web audiences being estimated at a
little more than 200 million people, with a little more than 112 million logging on
from Canada and the United States?32 the way is clear for news organizations to
launch online multimedia projects.

One of the goals of this work was to show that the Web is indeed a
singular and new media channel, just as TV, radio and print media are each
individual channels used to transmit information from senders to receivers. The
various media products people consume via this channel are all Web media and
not individual online newspapers, Web radio or Web TV; it is the Web's
multimedia capability that melds all these disparate media into a single new
medium as well as melding each of their individual abilities to deliver audiences
news content via a new and singular channel. The greatest benefits and returns
from the Web are in store for those who conceptualize Web media as such and
forego less relevant categories, such as online newspapers, online TV or online

radio.

 

”2 N_u_a Internet Surveys. "How Many Online?", Dec. 2001.
<http://www.nua.ie/surveys/how_many_online/index.html>.

137

:6 3TH" I

‘rll’

Categories, such as “online newspapers,” serve only to limit the
opportunities available to Web media entrepreneurs. Sites whose content is
conceptualized using traditional media models are limited by the parameters of
those models. This research has shown that the presence of visual elements on
an online news site enhances the ability of that site to deliver memorable news
content. The news recall scores created by the multimedia video and embedded
audio treatment (approximately 60 percent) outperformed recall scores for
articles accompanied by only one visual element (30 percent) by nearly 100
percent margin. And yet, the model for using visual elements to accompany news
text in traditional newspapers is to seldom have any accompanying visual
material in the news hole, and if any is used at all there, it is seldom more than
one visual element. Yet clearly, the theoretical imperative demonstrated in this
experiment is that the more visual elements that accompany a story, the more
recall their presence will engender.

To extend this logic further, perhaps the magazine model of using multiple
images to illustrate news might be a better model to use to conceptualize online
news sites, but for one thing: magazines use static images to illustrate content
and in this experiment, the news treatment that scored highest featured non-
static visual components embedded audio and multimedia video). Granted, the
single-element, static and moving visual component treatments scored nearly
identical in the recall trial, but the treatment that featured multiple visual elements
stood head and shoulders above all other treatments in terms of recall scores.

This seems to again point to the relative uselessness of using traditional media

138

models for online content sites because there is no traditional news media model
that uses multiple channels—text, video and audio—to deliver the same
message at the same time to the same audience with such high recall scores.

Another point worth raising is the apparent equal value the younger
audience in this experiment accorded both the static-image photo and the
multimedia video. These results point to a visually jaded audience segment.
Considering their exposure since early childhood to television and radio,
videocassette recorders, televisions that receive not 10-15 channels, but
hundreds and now, as they pass from adolescence to early adulthood, they are
presented with the Web. All this accompanies the traditional media experience
available to their parents’ generation.

While it is true that dual coding theory predicted single visual elements
scoring identical recall scores, it must also be remembered that dual coding
theory is primarily a theory derived from classroom experimentation. The content
it was created to evaluate, word and picture pairs, initially, then classroom
material, is hardly as compelling as news content might be. In this experiment,
the researcher purposefully chose follow-up news items in an attempt to reduce
the compelling nature of the news and thus remove it as a possible factor in the
recall scores.

The experimental part of this dissertation was, by and large, a success. It
was not without its flaws. But those flaws did not detract from the overall goal of
determining which multimedia presentation modes are most effective in

delivering news to online newsreaders.

139

Ultimately, this work shows that for those online news producers who seek
the greatest amounts of news recall, dusting off the age-old adage of “hit-'em-
with-it-and-hit-em-with-it-again” will no longer carry the day. Instead, the new
media adage should read: “hit-‘em-with-it-and-show-it-to-them-again”. The old,
one-tvvo punch of summarizing the quote in the sentence immediately prior to
actually delivering the quote is a tried-and-true method of underscoring the
importance of information in feature writing. Online news media content providers
must take that one step further. They must use that time-tested method print
media method, certainly, but additionally, bring the news home to online
audiences with audio and video embedded in the same interactive article. DCT
shows us the value of accompanying text with imagery. But a newsperson’s
instinct tells them it was the synergy and plethora of components that sold the
information to readers far more effectively than mere partnering of words and
images.

In addition, the value of novelty and interactivity cannot be ignored either.
At the time this experiment was performed, online news, especially online video
news was almost unheard of. The possibility was acknowledged, but issues of
bandwidth precluded their inclusion in normal media streams. For this reason, it
is possible that a measure of the substantially higher recall the multimedia video-
and embedded audio-laden piece of news garnered in this experiment might
legitimately be laid at the feet of novelty.

Yet, interactivity, also, clearly plays a role in this. The sum total of

“interactivity" demanded by newspaper—tuming pages and reading—or even in

140

television—“channeI-surfing” with a remote are no more attention getting nor
engaging than simply reading or staring at the screen.

However, the point and click of a mouse, the required keystrokes
necessary to navigate the Web by means of URLs acts as a primer to readers’
mental pumps prior to digesting news content from Web sites. Certainly, this
attention they must focus to merely turn the page must act, at some level, to
catalyze greater attention to be focused on the news itself.

Lastly, though the novelty of the Web may now be dimming, it is still, in the
words of words of Michael Lewis, “the new, new thing.”333 As such, the casual
users of this newer channel are just now gaining an expertise in using it and, for
a while, at any rate, will continue to approach their news with a greater freshness
to it, to perceive it as a more novel thing than they will other, less attractive (to

them), more traditional channels.

 

33’ Lewis, Michael. The New, New Thing: A Silicon Valley Stogr. New York: W.W. Norton Publishing,
1999.

141

Chapter 6

CONCLUSIONS

Introduction
This chapter contains a discussion of what has come before.

First, it will evaluate Allan Paivio’s Dual Coding Theory (DCT) for fitness
as an evaluative tool for the messages and content in traditional mass media.
Then it will consider its value as an evaluative tool of Web media. As mentioned
earlier in this work, DCT has gone through a similar progression as an evaluative
tool for classroom tools.

Next, the chapter will contain a discussion of the evolution of the World
Wide Web and Web media. It will also consider how the Web, access to it and
the changing face of the mass media industry segment has spurred change
within this media channel. It will also discuss various economic and technological
obstacles that might impede rapid advancement of technology in this segment.

Another area this discussion will delve into is the issue of convergence
and how that will affect the success of this new media channel. The Web exists
as a channel, but how well will media companies be able to use it to
communicate with disaffected audiences, such as young people, aged 18 to 34.

Who is this mass audience, where is it, what are its tastes and what is its
technology, its interests, will traditional mass media products survive? Or is this

audience segmented, even more so than cable TV? If it is, why is it? And what of

142

the digital divide? Are the disenfranchised an audience, too? Or are they truly
disenfranchised? How wide is that gap? Or is there a gap at all?

What is the future of news? What is the future of mass media? Will it still
be advertising driven? Or will there be a new economy...one in which information
is the commodity? Can we legitimately label computer media and, more
specifically, “mass media?”

There is another parallel area worth a bit of discussion, as well. Since
DCT enjoys some popularity among educational researchers, a discussion of
media convergence in the classroom would also, certainly be in order, especially
in the context of journalism education and journalism education research.

At the base of any discussion of this material, the first question that needs
to be answered is, “Is recall a desired objective in journalism?” Put simply, “Do
we want people to remember what they read, hear or see in the news media?”

The answer is yes, we do. The U.S. Constitution calls for a public that can
discuss the issues of the day within a free forum and its First Amendment calls
for a free press to provide information to stimulate this discussion within a free
forum of ideas. Thomas Jefferson, the Constitution’s author, in crafting this
philosophy, borrowed from Libertarian principle, which calls for a marketplace of
ideas in which citizens gather information so they can discover the tnlth and
create good from it. Today, this country runs on a democratic system that
absolutely relies on an informed public. Press freedom has guaranteed people
have enough lnfonnation to operate this free society the Constitution, along with

its citizens, has created.

143

Thus, those engaged in the craft of writing and reporting news should
naturally be concerned that they present information to their audiences in a way
in which it can be easily remembered. If it is easily remembered, then citizens’
abilities to participate freely and equally in the marketplace of ideas will be
enhanced. Recall, then, is a highly desirable goal.

From an lntemational perspective, news recall is also a highly desirable
goal. Over and over again, we have been shown that North America is not an
island. Improved communication has done what it has done throughout history: it
has served to reduce the physical and cultural distances between people.
Knowing where disasters strike enables us to know where to direct aid. Knowing
where wealth accumulates enables us to determine where economic opportunity
exists. Knowing where war is being fought, enables us to know where we might
try to bring peace. If these ends and means are indeed valuable, and can be
aided by creating more memorable news accounts, then again, news recall is a
highly desirable goal.

DCT: a mass media research tool

Steve Lacy once told an introduction to mass media class at Michigan State
University that a good theory is one explains the past, relates to the present and
predicts the future.334 Another MSU professor, Robert LaRose once said that a
theory is a roadmap for research, charting the process of exploration and that a

good research progress adds to the map.335

 

33" Lacy, Steven. Lecture, “Introduction to Mass Media,” Michigan State University, College of
Communication Arts and Sciences, Fall 1995.
3” LaRose, Robert. Evergreen Grill, East Lansing, MI, Fall 1997.

144

By any measure dual coding theory survives the first measure. It was
derived in the early 1970s and has been used since that time to address how
people place lnfonnation into long and short-term memory. Research has run a
full gamut of making picture-word associations for pre-school children on up to
comparisons students Ieaming subjects via traditional classroom teaching versus
multimedia computer programs. The second measure can will be determined at
the end of this work. Chapter 2 and this one will discuss DCT’s validity as a mass
media research tool.

DCT is a theory that fits well with mass media research. When we read,
we process verbal information and place it in long- and short-term memory. DCT
explains how we process information for memory. Briefly, the theory suggests
that the human mind Ieams lnfonnation far better when it processes both verbal
and visual information on a subject than when it is confronted with text
information alone. Verbal information is defined as text and numbers, while visual
information encompasses a continuum of external stimuli ranging from images to
sound, smells, tastes, even touch. When we recall things, the more that we can
draw from memory about the information we wish to recall, according to DCT, the
easier it is to recall it. Research, over time, has demonstrated an additive,
superior recall effect associated with pictures, defined as images by the theory.

From this brief description of a theory about which several books have
been written, it is easy to see how useful it might be to study mass media. The
print media produces information that is either text only or text with images in

both editorial and advertising material. Sometimes the text presents lnfonnation

145

alone or it also features photographic material. Other times, photographic
material presents its lnfonnation alone.

Television is a visual medium that also incorporates text. The amount of
visual information TV presents far outweighs the text it presents, but it does
combine both kinds of information. Radio presents lnfonnation that, as defined by
DCT, is solely considered imagery. Web media presents information that in both
visual and text forms.

Thus, there are three broad, mass media genres that present news
information in a way that, according to DCT, maximizes its ability to be stored in
long-tenn memory. The fourth, radio, presents information to audiences in a way
that shows a greater propensity for Ieaming than just text alone.

DCT assumes the existence of two cognitive processes: verbal and
imagery systems. Though each of these systems are structurally and functionally
distinct, they are not independent. They can operate independently; people can
Ieam and recall material using either system or both at the same time. Little mass
media content appears as text or imagery alone. Most content is a mixture of
verbal and text information, which lends itself to studies that use a theory
suggesting verbal and imagery systems stimulate memory together. Yet DCT
also addresses imagery or text that does stand alone, stating that it creates less
concrete and more abstract memories without the presence of both.

For online media, the theory has an even greater relevance. Since DCT
assumes verbal and non-verbal systems act in concert to stimulate memory

function, this theory, then, shows a great deal of promise in evaluating how

146

people process multimedia infonnation—media products that feature text,
photos, video and audio together.

Indeed the experiment in this dissertation points to greater recall with
complementary material and lower recall with fewer audiovisual story
components. This seems to build on the picture superiority recall effect Paivio
notes in earlier studies. This study shows the addition of more imagery along with
text, that is (contrary to previous studies) not redundant with text, provides the
highest recall.

DCT’s hierarchal structure parallels the stnlcture of the proposed items of
study, online news articles. Consider at the top of this structure is a sensory
stimulus of some kind that triggers memory function. Online news consumers got
to news sites because they associate specific sites from visual, even auditory
cues—in the case of Real.com’s RealPlayer—that tell them this place they are
visiting contains news. Site logos, sounds stimulate memory through
representational or symbolic memory functions that what they are about to read,
watch or listen to will be news. The structure of the articles they read are very
similar to the next level of the theory: articles can be subdivided by components,
into verbal or imagery system-related pieces. Some pieces with consist of only
text, while others will consist of text and photos and still others might consist of
compressed broadcasts alone. At the lowest level of DCT’s stnlctural hierarchy
are the imagens and the logogens, representational units that generate word-
and image-related memories. Similarly, words within articles or visual objects

within pictures and film trigger memories and enable recall to occur.

147

Mass media researchers has devoted a substantial amount of time and
energy to determine which media, broadcast or print, generates the highest news
recall rates. DCT focuses this question on the individual, asking if a person is
inherently more comfortable processing verbal or imagery system-based
material. Theorists suggest this determination can only be based on an
individual’s experiential dominance—that which they have been exposed to most
over time. Media preference studies have shown that younger audiences, as a
group, tend to favor broadcast media over print media. Thus, from a theoretical
standpoint, younger audiences tend more favorably toward image-based memory
systems. However, research using school children as subjects continues to
underscore the underlying assumption of DCT that verbal plus imagery system-
based material creates the most concrete memories. If anything, then, the theory
tells us what we already know—that text alone will not hold readers’ interest and
thus, will likely affect their ability to recall what news they’ve read in that form.

Another interesting parallel created by DCT is the way in which it theorizes
verbal and image system data is committed to memory. The theory suggests that
people commit verbal information sequentially and that non-verbal information is
placed in memory within a nested structure. Sentences are words placed within a
certain sequence to provide meaning. It stands to reason that material presented
sequentially, read sequentially will likely be recalled sequentially. This
underscores the importance of the news writing style (the inverted pyramid
writing style), which calls for paragraphs to refer backward to the previous

paragraph and forward to set up what will be written in the following paragraph.

148

In essence then, adherence to this writing style will foster the sequential nature of
the information that will enhance recall. Images and sound, too, must use
composition forms to create nested environments. Engaging photos and
broadcasts should be set against backdrops that created nested environments to
enhance recall. Backdrops or sound in the background will aid viewers and
listeners in creating associations that will later aid in recall. Broadcasts that
feature talking heads, photos that feature landscapes alone or people alone
(“grip-and-grins” and “talking heads,” for example) are less likely to engender
recall, according to DCT, for their lack of nested environments.

Structural information, according to DCT, is recalled in a simultaneous
way. When people visualize parts that make up a whole, they can also visualize
all those parts assembled and created as a whole. Similarly, when people
attempt to recall verbal and image systems presented together, such as an
online news story, they remember all the pieces together as they look for the
relevant image they wish to recall specifically. Here again, to aid in that
simultaneous recall, it makes sense to construct news stories with components
that fit easily together, whose organization is intuitive, so that when simultaneous
recall is required, it can be accomplished easily with as little mental hesitation as
possible.

One of the means by which recall is catalyzed, according to DCT, is by
means of referential interconnection. This is a between-system link that
connects imagens and logogens so that objects can be named and names can

evoke images. It is important to note that this interconnection does not run

149

counter to the assumption that the verbal and imagery systems are independent.
DCT’s notion of independence simply states that one system need not be in
operation for the other to be in operation. However, it is possible that logogens
might stimulate image-based memories, just as imagens may stimulate verbal-
based memories. From a journalistic perspective, then, most U.S. citizens can
easily put the correct name to a picture of President George W. Bush, when seen
on a Web news site. Similarly, when confronted with President Bush’s name in
print, these same people can easily conjure an image of George W. Bush.

A second interconnection is called an associative interconnection. This is
a within-system link that allows logogens to be associated with other logogens
and imagens to be mentally associated with other imagens. Both verbal and
imagery systems are bi-directionally interconnected; in other words, each can be
used to recall the other. DCT also assumes interconnection to be one-to-many;
that various logogens and imagens will stimulate recall of not one single memory,
but several, or in some cases, many memories. Again, using George W. Bush
example, some might recall the questionable manner in which the Florida vote
was tallied in his election when confronted by his photo; or they might recall that
his daughters were arrested for underage drinking; or they might recall how he
looked while inspecting the demolished World Trade Center in September 2001.
Similarly, his name in print might evoke pictures of all these images in our minds
as well. Mass media seems to already recognize that bi-directional
interconnection exists. Text and visual components of news are seldom

presented alone to audiences. Film has voiceovers, text is usually presented

150

together with illustration or photography and that photography is usually
accompanied by cutlines to further explain what readers/viewers see.

DCT assumes associative processing in the verbal system is hierarchical.
For example, “vehicle,” is a category label for “truck” and “car". The theory says
hierarchies are based in the cultural organization of our language structures.
Media depend heavily upon cultural organization of language structure to deliver
messages to their audiences. This is especially true for those media that attempt
to appeal to some kind of localized audience. “Speaking the same language” is
very important to local media, such as newspapers, non-network-affiliated radio
and broadcast television stations, in terms of reaching their audiences. This point
does not serve to make study of Web media products difficult with this theory.
While it is true Web media attempt to appeal to mass audiences, the Web’s
audience is, in general, younger, hipper, has higher income levels and is more
educated. Therefore, cultural organization of language structure for that audience
is just as critical as it might be for more traditional, localized media, despite the
global geographic diffusion of the audience.

Thus far, discussion has covered how memory is created and how it is
created so that it might be easily recalled. DCT also addresses how memory is
activated. As might be expected, DCT states that memory is activated by the
joint functions of stimuli and individual difference, simply put, the need to recall
memory and affective, behavioral and sensorimotor qualities of individuals. DCT
specifies two kinds of stimuli that come into play in memory activation: target and

contextual stimuli. Target stimuli are those things that trigger memory—an article

151

or photo at which a reader might be looking. Contextual stimuli are those objects
within the photo or text that act as an instruction set to activate or recall other
information already in memory. For example, we are reading an article in the
Wall Street Journal that live hog prices have fallen lower than they have in nearly
50 years. The article is the target stimulus. The words, “live hog prices,” are the
contextual stimuli that cause us to begin to recall many thing we might associate
with hog prices—hog farmers we know, the lowest prices we can remember, the
things that low prices will cause.

The uses of imagery, more concrete words and descriptive terms have all
been shown, through experimentation, to create more concrete memory than
simple verbal stimuli does. Journalists write using simple terms, short sentences
and use active voice. By minimizing the use of descriptive terms, adjectives and
eliminating passive voice, reporters write clearly, concise, and, in doing so, write
more concretely. According to DCT, then, by accompanying this more concretely
written text with photographs, or, in the case of Web pages, with embedded
audio or multimedia video, journalists increase the ease with which memory of
the material in their stories can be activated.

Memory creation: news recall

When memories are activated, three different processing tasks are activated,
depending upon the form of stimuli activated. The most direct means is by direct
activation of logogens by words or imagens by images. This is referred to as
representational processing. Thus seeing a familiar face in a photo or reading a

familiar turn of phrase in a story would both be examples of representational

152

processing. Each of the contextual triggers directly stimulate memories of words
and images read or seen in other stories read previously. Referential processing
is cross-system stimulation of words that evoke images or images that evoke
words. Theoretically, this form of processing is considered more indirect since it
does involve a mental crossover of systems to activate memories. Associative
processing involves activation of one logogen by another or one imagen by
another.

In direct activation, information read in an article from a Web news site
would trigger a memory of information in a reader’s memory in an act of
representational processing. Information in the stories, delivered by means of still
or video images, sound files, even text act as a contextual triggers to activate
data stored in our memories. In an example of cross-system stimulation, words
might activate old images; new images might activate memories derived from
previous readings in text, thus triggering the act of referential processing.
Additionally, associative processing also occurs when online news readers see
news in text and their memories are stimulated, recalling information they read in
the original or first story they read on a subject either the previous day or even,
as is the practice in online news, on a story written the previous hour.

Each of these kinds of processing functions is not unique to the online
venue. which once again, leaves this researcher in a state of surprise over the
lack of previous research in traditional media that might have used this theory as
a guide. DCT's utility as a means to describe how people read and process

information would have been the same for older media studies since the process

153

(readers reading, listening or watching news stories), aside from the tools used to
access this information, is virtually the same.

Representational systems serve another function in creating memory.
Called the mnemonic function, it explains how imagery and verbal systems aid in
creating memory. One assumption that governs this part of DCT is that the
pathway used to create a particular memory is modality specific.336 In other
words, if the memory was encoded verbally, it will be retrieved verbally and
similarly, if it was encoded by means of the image system, it will be recalled as
an image. Representational coding—words evoking word memories, images
evoking imaged memories—is usually sufficient for some level of accuracy in
recognition and recall tests.

Referential encoding is also possible, as long as the contexts in which the
input items are presented are not confusing. Viewing a multimedia news video on
an online news Web site would be a contextually appropriate way to view online
news material. Online news readers can referentially encode lnfonnation
gathered by reading/viewing news material on a Web site because it is presented
to them in an expected fashion. The history of news presentation is such that we
have come to expect news to be presented visually, because as an audience we
already have experience with TV news and photojournalism, in an audio format,
since we have already experienced radio news and news in text because we
have already experienced newspapers. By contrast, we may experience difficulty
in recalling information from animated graphics, because as an audience, we

have very little if any experience with information presented in this venue.

154

DCT also notes that the likely success of referential encoding will be much
higher if subjects are given instructions as to the nature of the tasks prior to
processing the information.337 Objects, such as sliders, radio buttons and links
fulfill this function on Web pages. Since experienced Web users Ieam how to use
these devices through experimentation or on-site instruction. Their function
serves as instruction on how to access material and also provides clues as to the
nature of the material they will access in using these features prior to processing
this information. For example, a slider button underneath a framed visual image
cues viewers they are confronted with a video frame on a Web page. It tells them
the lnfonnation they are about to process is moving and not a still photo and will
likely, but not always be accompanied by a sound track. Thus, the user is
“instructed” to receive information in a particular way by the cues created by the
frame surrounding the text or images they are about to process. Additionally, the
presence of these cues serves another purpose: they serve as a cue to the
presence of interactive material. Since this material is interactive, it requires
some action on the part of the reader beyond simply reading or watching
information displayed on a screen. Readers must engage in an activity to activate
this information, thereby heightening their awareness of what is in front of their
eyes, which readies them for a flow of lnfonnation from the Web site. In effect,
this interactive material serves to stimulate readers from the mental lethargy that
may have resulted from continued exposure to text material or still images, which

excite less involvement on a reader’s part since it requires only visual and mental

 

”6 Paivio Mental, 55-58.
”7 Paivio Mental, 250.

 

155

processing and not some measure of physical activity, such as button pushing ort
slider activation.

Associative encoding is the activation of associated representations in the
same symbolic modality as the input item?38 Thus, a subject who recalled the
subject matter in a video as a mental representation of a film clip in which news
was imparted would have engaged in associative encoding to put that
information into memory. What is interesting about this theory is that video has
no greater value than does still photography. The theory makes no qualitative
distinctions between these two visual modes. Instead, it suggests a quantitative
variable that exists; that the only modifier on the concreteness or quality of
memory is the amount of visual information that accompanies text. In fact, DCT
suggests is that multimedia is a superior recall creator because it stipulates a
superior picture recall effect: the presence of images accompanying text, creates
far more concrete images in memory than either text alone or images alone can
create.

Statements made by Paivio in which he said research showed the
“mnemonic superiority of images” was pegged at having a 2:1 advantage over
verbal codes further support this point339 It clearly lays out a theoretical
imperative that a multimedia channel dedicated to information delivery should
use images whenever possible. It keeps coming back to the goal of news
organizations: to inform entertain and educate audiences. Each of these goals

requires the audience to Ieam some of the material they read, to commit

 

33“ Paivio Mental, 144.
”9 Paivio Mental, 76.

 

156

something to memory, if for no other reason than to add it to their memories'
collective. Indeed the results this dissertation’s experiment underscores this
same mnemonic superiority. The mode that netted the highest recall score
(58.9%) featured two image codes—multimedia video and embedded audio. This
score was nearly twice the recall scores of the still photo mode (31.2%) and that
of the multimedia video (31.4%).

These experimental results illustrate the additive nature of images when
presented in concert with text. Here again, we see the additive nature of the
presence of image codes (sound and video). When twice as many image codes
are present in different modes, the results show a linear improvement in recall.
Based on this, it is surprising there are so many news Web sites that do not
feature as much image coded material as possible to sell their messages to their
audiences.

Theoretically, though seemingly not borne out by the experiment done in
this study, text alone is far more difficult to use in memory creation. Memory
created from verbal information must be stored sequentially, and is recalled
sequentially. Because verbal stimulated memory creation is sequential,
according to DCT, any words used to create memory must be extremely well
written and very well organized. This requirement may speak more eloquently
than any other reason as to why it seems so difficult to create memory from text
alone. So much of what we read is not written as well as it might be, or, more
accurately has not been edited as well as it might be. Poorly written material

doesn’t “hang together" very well, frequently because it is poorly organized.

157

T'TT-‘ml

Organizational problems in writing would make a sequentially ordered memory
creation process a difficult task.

Similarly, it can also be noted that single-coded information, stand-alone
text, stand-alone photos, audio or film, according to the theory, will not generate
as much recall as multimedia information will because information presented to
users that contains both text and image lnfonnation creates more concrete
memory. Additionally, it will be stored as both imagery and verbal information
because it will have been received as text and imagery lnfonnation.

When readers consume online news, they must engage two kinds of
memory to make sense of what they are reading: episodic and semantic
memory. Semantic memory is the part of memory in which information is
manipulated, transformed and compared to create meaning?‘1o Episodic memory
is memory about specific events at a particular time and place.341 News is
inherently event-driven and thus, on some level, those who read, listen to it or
watch it are processing it in terms of what has occurred or what they have
experienced previously as it relates to the subject matter. Thus, at least a portion
of the kind of memory engaged to process news information is episodic. This is
especially true for news presented on the Web since it is updated far more
frequently than news in other media. Another thing inherent to news is that
events must be related to those who they involve; in other words, news
consumers evaluate and relate what they read against how the events contained

in it might affect them or others they know. DCT recognizes that semantic

 

3‘0 Paivio Mental, 120-121.
3‘" Tulving, Endel,. “Episodic and Semantic Memory,” Eds: Endel Tulving and Wayne Donaldson,

 

158

memory objects are made of connected, sequenced or nested units. Similarly,
episodic memory tends to be sequential. But beyond this at the end of the
lnfonnation presentation the process, what is remembered or recalled is not the
representation of the parts, but rather the representation of the whole. It is not
letters we remember, nor phrases nor even sentence fragments. Instead, we
remember concepts imparted by the text and synthesized against what we know.

Another important point is that DCT assumes no structural or functional
differences between the episodic or semantic memory. Thus when studying a
multimedia channel, the theory driving the research suggests a methodology that
lends itself to certain kinds of research and not others. Research aimed at the
function of mass media, that which asks questions of the end result—recall or
some comparative measure of recall—will find DCT more useful than those who
would delve more deeply into the cognitive process of deriving meaning from
what is read, listened to or watched.

Meaning creation, or memory trace, as it is referred to in DCT literature, is
derived from both external and internal stimuli. We see a story on a Web site
about a successful attack on the Afghan city of Kabul. The external stimulus is
the Web site on which a user finds the news item as well as the news item itself.
The internal stimuli might be the knowledge a user might have that Kabul is the
capital city of Afghanistan, which allows this user to realize that this attack is
aimed at a significant target and that it is of strategic importance. When this

information is stored in memory, it would then be stored within the context of

 

Organization of Memog, (New York: Academic Press, 1972: 58-74.

159

lnfonnation about the capital of Afghanistan, the war being fought in Afghanistan
and other related memory items.

Just as there are three kinds of encoding, so, too, are there three kinds of
processing, or decoding. These three means of decoding memories are identified
using the same distinctive labels as types of encoding: representational,
referential and associative. Representational decoding occurs from text as a
result of direct activation by language stimuli and it occurs from images as a
result of direct activation by non-language stimuli. Referential decoding occurs as
visual information stored in memory helps add meaning to linguistic stimuli and
text information stored in memory will also aid in meaning constnlction when
viewing visual lnfonnation. Since both kinds of stimuli would be present, it
theoretically stands to reason that whatever meaning is synthesized from
material that contains both kinds of information—linguistic and imagery—the
factual information derived in this fashion will be more concrete than memory
derived from either linguistic material alone or imagery alone. Associative
decoding relies on episodic memory, which, in essence is memory called up that
is essentially “jogged” or stimulated by what is being read. This associative
encoding and decoding are essentially the grist of follow-ups. News, especially
big news, is seldom handled in a single story. Were it not for associative memory
in any form, follow-ups would be a tiresome exercise at best, with too great a
necessity to rewrite stories from their beginnings. Our associative memories

allow reporters to write but a few paragraphs, run the same sound or video clip to

160

“mart _

refresh readers/listeners minds about the news of yesterday that now has a new
twist.
The experiment

The experimental part of this dissertation was, by and large, a success. It
was not without its flaws. But those flaws did not detract from the overall goal of
determining which multimedia presentation modes are most effective in
delivering news to online news readers.

The experimental portion of this dissertation was designed to use the
theoretical framework proposed earlier in this work to test the representational,
referential and associative processing functions of memory that, according to
DCT, occur as people process imagens and logogens. The larger question deals
with whether or not the superior effect of pictures and sound will enhance recall
of information of both verbal and non-verbal information in news stories and to
what degree. More specific questions deal with superiority of recall among
various text, text and visual and text and visual and audio presentation modes.

Clearly, the embedded audio and multimedia video treatment scored the
highest recall. This is not surprising when the additive effects described by DCT
research are considered. Called the additivity hypothesis, it was first proposed to
explain the superior ability of concrete words to be recalled over subjects’ ability
to recall abstract words and later extended to explain a similar picture superiority
effect. For example, DCT predicts symbolic comparisons of concrete objects will

be faster with pictures than with words because the imagens are accessed more

161

“I

quickly when pictures serve as stimuli?” In episodic memory, “the mnemonic
contribution of the non-verbal (visual-image) trace component consistently turned
out to be greater than that of the verbal component?“

Therefore, it should not be a surprise that the highest scoring treatment in
this study featured two image components (audio and multimedia video) in
addition to the verbal (text) component. The fact that these two image
components—the embedded audio and the multimedia video—were
complementary?“ and was not repetitive, enabled the experiment to avoid the
repetition lag paradigm of free recall. Repetition lag refers to the phenomenon in
which recall of repeated items in a list increase with the number of items that
intervene between repeated events?45

Ultimately, what this means is that that for those online news writers who
seek the greatest impact, in terms of news recall, will merely dust off the age-old
adage of “hit-‘em-with-it-and-hit-em-with-it-again.” The old, one-two punch of
summarizing the quote in the sentence immediately prior to actually delivering
the quote is a tried-tried-and-true method of underscoring the importance of
lnfonnation in feature writing. This same strategy is clearly useful for delivering
information to news audiences. DCT shows us the value of accompanying text
with imagery. But a newsperson’s instinct tells them that it was the synergy and
plethora of components that sold the information to readers far more effectively

than mere partnering of words and images.

 

342Paivio Mental, 180.

3’3 Paivio, Mental, 142.

3"Ihe material presented in video and audio forms in no way repeated information presented in the
article’s text.

 

162

nil-lam. 1’! '—

In addition, the value of novelty and interactivity cannot be ignored either.
At the time this experiment was performed, online news, especially online video
news was almost unheard of. The possibility was acknowledged, but issues of
bandwidth precluded their inclusion in normal media streams. For this reason, it
is possible that a measure of the substantially higher recall the multimedia video-
and embedded audio-laden piece of news garnered in this experiment might
legitimately be laid at the feet of novelty.

Yet, interactivity, also, clearly plays a role in this. The sum total of
“interactivity” demanded by newspaper—tuming pages and reading—or even in
television—“channel-surfing” with a remote are no more attention getting nor
engaging than simply reading or staring at the screen.

However, the point and click of a mouse, the required keystrokes
necessary to navigate the Web by means of URLs act as a primer to readers’
mental pumps prior to digesting news content from Web sites. Certainly, this
attention they must focus to merely tum the page must act, at some level, to
catalyze greater attention to be focused on the news itself.

Lastly, though the novelty of the Web may now be dimming, it is still, in the
words of words of Michael Lewis, “the new, new thing.”346 As such, those who
are denizens of this newer channel are just now gaining an expertise in using it
and, for a while, at any rate, will continue to approach their news with a greater

freshness to it than they will other, less attractive, more traditional channels.

 

3’5 Paivio, “Retrospect,” 265.
3’6 Lewis, New.

163

Another portion of the experiment that deserves at least some discussion
are its flaws: the fact that the text-only portion of the experiment was the second
highest scorer and that there was not an attempt to discern if interest in the
stories read might have skewed the results.

News is interesting or there wouldn't be newspapers, TV and radio news
programs and online news. In fact, one of the more interesting outgrowths of
online news is how often stories may now be updated within a single day. For the
first time since it’s inception, broadcast news is experiencing significant
competition in an area it heretofore had no competition in—its ability to update
stories. Unlike print media, electronic media could break into any part of its
program day to bring new information to its audience. Now Web media are
capable of the same thing. Since the audience for this media is asynchronous—
they “tune in” and consume news on their own schedules, and are not limited to
the editions or programs times—online journalists have felt free to update the
news whenever events dictate.

The novelty of news is what attracts us to it. Each day, it is involves a
unique panoply of events and people, as unpredictable as the next day’s
direction of the wind. As readers and daily consumers of this new lnfonnation, we
pay it our highest homage by reading it first and giving it out greatest attention.

In this experiment, newsreaders did precisely that—to a point. Recall of
two stories that, according to DCT, should have featured higher scores than the

text-only treatment, did not outscore the text-only story. And yet, the one story

164

‘1‘“?«352 III, I—

that had the most multimedia components, and was still merely a follow-up news
story, outscored the news of the day.

While it may be argued that this story was an lntemational news story,
and, as such, would generally score lower in reader interest than anything with a
domestic slant, the margin by which this story outscored all others clearly points
to an effect that goes way beyond mere news interest. The multimedia story
outscored the text-only story by nearly a 20 percent difference and the nearest
multimedia competitors—text and still photos and text and video—by almost 30
percent.

Though it can be legitimately argued, from a statistical analysis standpoint,
that reader interest in various stories may certainly impact on recall, it remains to
be seen if that could actually be controlled for. How could a question be framed
to get at that particular interaction that did not strike at the heart of what is reader
recall? After all, interest is what draws us into a story; it’s what keeps us reading
and what sparks our memories into recalling. Therefore, in an attempt to account
for it, we might actually find ourselves in the unenviable position of trying to
control for our dependent variable. Perhaps the only way to do this would be to
ask a set of questions at the end that attempted to identify high- and low-reader
interest by subject area such as international news, or by issue, such as “the war
in Kosovo.” The questions would probe reader interest with test subject matter
prior to their participation in the experiment. However, these answers might be
colored by the fact that test subjects had just read articles on this particular

subject.

165

tar-“““il' _

What of the Web?

At this point, it is important to note that while many of the larger operations
in online journalism are already constructing news sites that feature as much
multimedia as the Web can sustain, the majority of all online media are not.
Many, many operations that might do this, still don’t, now more than 10 years
after the Web’s popularization as a mass medium. And that is a point worth
making. Clearly, nearly every newspaper in the country is convinced of the
wisdom of having its own Web site. Yet the vast majority still merely repackages
print news onto Web sites with neither thought nor care about the fact that the
Web is inherently a different channel with an inherently different audience than
traditional print media. While many media consultants pay lip service to
convergence, only the major news organizations are actually following through
on assuring convergence—the use of text, still photography, videos and
embedded audio—is a part of their Web offering.

It can be argued that many of the operations that have not incorporated
multimedia into their content do not do so because they either do not have
access to that kind of material nor can they afford to purchase it, or, if they could
purchase it, could not purchase it in a timely enough fashion to be useful to a
news operation. Then, too, the same kind of economy that exists in print and
electronic media today, is proving disturbingly elusive on the Web.

Just this past quarter, Amazoncom posted its first quarterly profit after
nearly five years of existence. These same woes seem to plague lntemet

advertisements. The sheer breadth and depth of the Web has served to diffuse

166

the relatively small amount of lntemet advertising available to relatively few sites,
making profits in this channel somewhat scarce. With advertising profits low and
subscription fees non-existent, or nearly so, the question becomes how can any
publication not making money from its lntemet operation justify purchasing
freelance editorial material?

Simply put, it cannot. And since many cannot, and do not have affiliations
with broadcast operations, multimedia remains far out of the grasp of the majority
of online news sites, the revelations of this work, notwithstanding. Until lntemet
news operations show profits, it is unlikely they will embrace multimedia as they
should to maximize reader recall.

This begs the question, then, what are steps these operations might take
to incorporate multimedia into their sites that do not quite reach video and
embedded audio? The answer is to mimic the actions of advertisers. lntemet
advertisers have long been using animation to attract attention to their
messages. Why could Web news operations resort to the same tactics? Indeed,
using animations to liven up information graphics or to create enticing photo-
illustrations might well aid in improving reader recall of online news lnfonnation.

Lastly, it should be noted that bandwidth, the capacity of the Web to
handle large amounts of data, be they graphics, audio or video lnfonnation is still
largely limited—too limited for large scale use of video and audio on Web news
sites. lntemet 2, the consortium of universities and corporations devoting
substantial amounts of time resources and personnel to solving Web bandwidth

issues with hardware and software innovation has yet to make significant gains in

167

 

large-scale bandwidth improvement. ThUs, web media must wait for newer, faster
communication protocols that will allow high-speed transfer of high-capacity data.

In the absence of real improvement in Web-wide bandwidth, the best
solution for web companies is to minimize the size of multimedia files and lengths
of presentations. That way, Web media can still present multimedia content
delivery to audiences and not overload the network. In recent years, the more
ubiquitous diffusion of broadband access, such as digital subscriber lines (DSL),
cable modems and integrated service data networks (ISDN), have all contributed
to greater speed available through last-mile connections. These advances have
occurred within the last two years and their adoption bodes well for higher-
capacity services, such as online media, that wish to incorporate more high-
capacity content into their offerings. Indeed, the same is true for wireless
communication. Once this protocol gains an advantage in bandwidth capacity, it,
too, will aid in online news sites’ quest to include more multimedia into news
delivery.

One other concern weighs heavily on any discussion of the Web and that
is of the so-called Digital Divide. This concept holds that there is a population that
cannot avail itself of the value of the lntemet. The reasons vary: for some, they
lack the economic means to purchase the necessary equipment to access the
content; others lack the necessary English language skills that the majority of
lntemet users have; and for still others, the cultural landscape is so alien, that

access is, by no means relevant.

168

For those in the former condition, relief seems to be on the way.
Computers are cheaper now than they have ever been.

Conclusion

No research project may be called complete before an agenda for further
discovery might be laid out. This work is no exception.

First, it should be replicated to see if the results from this work can, indeed
be replicated. If the original results can be replicated, then this work will be
proven to be all the more valid. Additionally, now that high-speed lntemet access
is far more ubiquitous than it was when this work was done, there is also merit in
replicating the work to see if further interest in this content might be spurred by
higher speed access.

This work looked at how 18-25 year old college students read online news
and remembered what they read. But there are many of groups worthy of that
same look, especially more middle-aged mainstream newspaper reading groups.
It should also be duplicated across various educationally representative groups
within the news reading public.

There are lessons to be gained by Web advertisers from this work as well.
Many printed publications seek the same answers about their advertisements as
we might about editorial. How well is it read? How much do people remember?
How might recall be enchanced? Does the presence of certain kinds of content
enhance advertising recall? And what of advertorial? Do the same rules for recall

apply for this kind of content as in editorial material?

169

All these questions bear investigation, using DCT as a guide to further our

understanding of how readers read and process online information.

170

APPENDICES

171

ONLINE NEWS RECALL QUESTIONNAIRE
Part I
1. What did Gail Lloyd say Admiral Scudi did with the money?

2. What did Scudi to provoke the anger of his superior officers in the U.S. Navy?

3. What did Robert Maginnis have to say about the U.S. Navy's handling of the
Scudi case?

4. What penalty did the U.S. Navy penalty impose on John Scudi?

5. How did women’s advocacy groups have to say about the Navy’s handling of
John Scudi?

6. To which ethnic background do the rebels near Kosovo belong?

7. What do western diplomats believe the Serbian leader will respond to rebel
activties in the region?

8. What do the Albanian rebels call themselves?

172

,.____....,, ill—ll
1 . '_

 

9. Why is it so difficult to control the Albanian rebels?

10. What evidence is there that even the region’s children are at risk from the
Albanian rebels?

11. What did Republican New York Congressman Peter King say about the
chances Congress would vote to proceed on impeachment hearings?

12. What were the two conditions the GOP would require in a proposal to
censure President Clinton?

13. What issue about President Bill Clinton did the House Judiciary Committee

decide to drop in favor of considering the facts surrounding the president’s
relationship with Monica Lewinsky?

14. What kind of plan did Rep. Peter King create for dealing with President
Clinton?

15. Who is Annan Kofi?

173

‘z? 4.. urinal #

 

16. What did he want to talk to Libyan Leader Muamar Ghadhafi about?

17. Where are the suspects being held?

18. If Annan’s discussion is successful, where might the suspects be tried?

19. Where might the suspects be jailed?

20. Why did Libya’s news agency say Gadhafi was not qualified to sign any
agreements on Libya’s behalf?

Part II
Please circle the appropriate answer.
1) You are:

Male Female

2) Please indicate your race:

Black American Indian
Hispanic Asian
Caucasian Other

3) Is English your primary language?

Yes No

174

mum-ll

4) Do you have trouble reading the English language?

Yes No
5) Do you have a Ieaming disability (dyslexia, attention deficit disorder, etc.)?

Yes No

6) Do you have a severe hearing problem that would affect your ability to listen
to the news?

Yes No

7) Do you have a severe vision problem that would affect your ability to watch
the news?

Yes No
8) Did you have any particular problems completing the experiment?
Yes No
If yes, please explain:

Indicate how much you agree of disagree with each of the following statements,
on a scale of one to five, where 5: Strongly Agree and 1= Strongly Disagree.

9) I don't feel in control when I use a computer.

Strongly Agree 1 2 3 4 5 Strongly Disagree
10) I have problems using computers.

Strongly Agree 1 2 3 4 5 Strongly Disagree
11) Computers are a fast and efficient means of getting information.

Strongly Agree 1 2 3 4 5 Strongly Disagree
12) Computers are bringing us into a bright new era.

Strongly Agree 1 2 3 4 5 Strongly Disagree
13) I would hate to go without using a computer for more than a few days.

Strongly Agree 1 2 3 4 5 Strongly Disagree

175

'L ’Jud— "

 

14) I often use computers for recreational activities.
Strongly Agree 1 2 3 4 5 Strongly Disagree

15) I often use computers use computers for news and information gathering.
Strongly Agree 1 2 3 4 5 Strongly Disagree

16) Which of the following best describes your feelings about news in print
media?

Trustworthy 1 2 3 4 5 Untrustworthy

17) Which of the following best describes your feelings about television news?
Trustworthy 1 2 3 4 5 Untrustworthy

18) Which of the following best describes your feelings about radio news?
Trustworthy 1 2 3 4 5 Untrustworthy

19) Which of the following best describes your feelings about news on the
lntemet?

Trustworthy 1 2 3 4 5 Untrustworthy
20) Where do you get most of your news? (Circle only one.)

Print media Radio TV lntemet

176

APPENDIX 8

ONLINE NEWS SITE DESIGN ANALYSIS SURVEY DATA — SUMMER 1997

Note: The “NAME” column is the name of the online news site. The “INDEX”
column notes if the index was depicted in a banner along the side of the page or
set off in frames. “T EMPLATE” refers to a particular color scheme used on the
page, either two color (2/C), three-color (3/C) or four-color WC) and whether the
page was justified-left (JL) or centered (CTR.) “MM” referred to whether or not
multimedia was in evidence anywhere on the site; “AUDIO”, “VIDEO”, “ANIM”
and “STILLS” referred to whether or not audio, video, animation or still
photography was in evidence on the site. Starred responses (*) in these four
categories indicate the individual multimedia mode was in evidence in the site’s
advertising content only

Name Index Template MM Audio Video Anlm Stills
Alabama Live Frames 2/C - JL No No No No Yes
Birm. (Ala.) Post-Herald Banner 3/C - JL No No No Yes Yes
Decatur Daily Frames Mixed No No No Yes Yes
Anchorage Daily News Columns 2/C - Center No No No No No
Juneau Empire Banner 2/C - JL No No No No Yes
Arizona Republic Banner 2/C - JL No No No Yes No
Arizona Daily Star Banner 2/C - JL No No No No Yes
Prescott Courier Online Banner 1/C - JL No No No Yes Yes
Yuma Daily Sun Banner 1/C - JL No No No Yes Yes
Arkansas Online Banner 2/C - JL No No No No Yes
NW Ark. News Online Banner 1/C - JL No No No Yes* Yes
Cabot Star-Herald Banner 1/C - JL No No No No Yes

Log Cabin Democrat Banner 1/C-JL No No No Yes Yes

177

Name Index Template MM Audio Video Anlm Stills
Alameda Times-Star Frames 1/C - JL No No No Yes Yes
Bakersfield Californian Banner 2/C - JL No No No Yes No
Contra Costa Times Frames 2/C - JL No No No Yes No
Fresno Bee Banner 2/C - JL No No No Yes* Yes
Daily Independent Banner 2/C - JL No No No Yes Yes
Desert Dispatch Frames 1/C - JL No No No No Yes

SF Examiner/Chronicle Banner 1/C-JL No No No Yes Yes

LA Times Banner 3/C - JL No No No Yes* Yes
Malibu Times Frames 1/C - JL No No No No Yes
Modesto Bee Banner 2/C - JL No No No No Yes

Orange County Register Banner 3/C - JL No No No Yes Yes
Pasadena Herald Tribune Frame 2/C - JL No No No Yes Yes
Riverside Press-Enterprise Frames 1/C - JL No No No No Yes
SignOn San Diego Banner 3IC - JL No No No Yes* Yes
Telegram Tribune Online Banner 2/C - JL No No No Yes* Yes
Santa Rosa Press Dem. Banner 3/C - JL No No No No Yes
Stockton RecordNet Banner 2/C - JL No No No Yes Yes
Press Telegram (LA) Frames 3/C - JL No No No No Yes
Sacramento Bee None 3/C - JL No No No Yes* Yes
San Jose Mercury News Banner 3/C - JL No No No No Yes
Torrance Daily Breeze Banner 1/C - JL No No No Yes No

Boulder Daily Camera Banner 2/C - JL No No No Yes Yes

178

Name Index Template
Denver Post Online Banner 3/C - JL
Colorado Springs Gazette Banner 1/C - JL
Glenwood Post Banner 1/C -JL
Greeley Trib Online Banner 2/C - JL
Hartford Courant Banner 2/C -JL
Washington Post Banner 1/C - JL
Bradenton Herald Banner 1/C — Ctr.
Florida Times Union Banner 4/C - JL
The Miami Herald Banner Mixed - JL
Naples Daily News Banner 1/C - JL
Ocala Star Banner Banner (r) 1/C - JL
Palm Beach Post Banner 2/C - JL
Fort Myers News-Press Banner 1/C - JL
St. Augustine Record Frame Mixed - JL
Ft. Laud. Sun-Sentinel Banner (r) 2/C - JL
Gainesville Sun Banner 2/C — JL
Tallahassee Democrat Banner 1/C - JL
Tampa Tribune Banner 1/C - JL
TC/Palm Beach Banner Mixed - JL
Atlanta J-C Banner 2/C - JL
Augusta Chronicle Banner 2/C - JL

Columbus Ledger-Enquirer Banner 2/C - JL

179

MM AudIondeo Anlm Stills

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

Yes*
No
No
Yes
Yes
No
No
No
Yes*
Yes
Yes*
Yes*
Yes*
Yes
Yes
No
No
Yes
No
Yes*
Yes

Yes

Yes
Yes
Yes
Yes
No

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No

Yes
Yes
Yes
Yes
Yes

Yes

Name Index Template
Macon Telegraph Banner 2/C - JL
Athens Daily News Banner 1/C - JL
Savannah Morning News Banner 2/C -JL
Chicago Tribune Banner 2/C - JL
Chicago Sun-Times Banner (top) 1/C - JL
Arlington Hts. Dly. Herald Banner 2/C - JL
Edwardsville Intelligencer Banner 2/C - JL
Evansville Courier Banner 2IC - JL
Gary Post Tribune Frame 1/C - JL
Fort Wayne News Sent. Frame 2/C - JL
Indianapolis Star Banner 1/C - JL
Cedar Rapids Gazette Banner 1/C - JL
Topeka Capital Journal Banner 2/C - JL
Newton Kansan None 2/C - JL
Witchita Online Column 4/C - JL
Courier Journal Online Banner 1/C - JL
Kentucky Connect Frame/Banner 3/C - JL
NOLA Live Banner 1/C - JL
The Capital Online Banner 1/C - JL
Sunspot Banner 2/C - JL
Boston Globe Online Banner 2/C - JL

GazetteNET Banner 2/C - JL

180

MM AudioVideo Anim Stills

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes*
No
No
Yes*
No

Yes

Yes
No

No

Yes
Yes
Yes
No

Yes
No

No

Yes
Yes
Yes
Yes
Yes
Yes
Yes
No

Yes
Yes
Yes

Yes

Name Index Template MM Audio Video Anlm Stills
Detroit News Online Banner 1/C - JL No No No Yes Yes
HDT Info Connect None 1/C - JR No No No No No
Holland Sentinel Online Banner 2/C - JL No No No No Yes
Brainerd Dispatch Online Banner 2/C - JL No No No No Yes
TribUniverse | Duluth Banner 1/C - JL No No No No Yes
PioneerPlanet Banner Mixed - JL No No No Yes* Yes
startribune.com Banner Mixed - JL No No No No Yes
Sun Herald Online Banner 2/C - JL No No No No Yes
The Examiner Banner 1/C - JL No No No No Yes
Hannibal Courier-Post Banner 1/C - JL No No No Yes No
The Star.Kansas City Banner 1/C - JL No No No Yes* Yes
POSTnet Banner 2/C - JL No No No No Yes
thelndependent.com Frame 2/C - JL No No No No Yes
RJ Online Banner 2/C - JL No No No No Yes
SUN Online Banner Mixed No No No Yes Yes
Billings Gazzette Frames 21C - JL No No No Yes Yes
Bozeman Daily Chronicle Column 2/C - Jl. No No No No Yes
Havre Daily News Column 2/C - JL No No No Yes Yes
Helena Independent Banner 3/C — JL No No No ' No Yes
Daily Inter Lake Banner 3/C - JL No No No No Yes
Livingston Enterprise Frames 2/C - JL No No No No Yes

Missoulian Online Banner 2/C—Ctr. No No No No Yes

181

Name Index
Alliance Times-Herald Frames

Grand Island Daily Banner
Independent

Kearney Cyber Hub Banner
North Platte Telegraph Banner

starherald.com Banner

Columbus Telegram.com Banner 2/C — JL

Hastings Tribune Banner
Lincoln Journal Star Banner
Norfolk dail News Column
Omaha.com Column
RJ Online Banner
Las Vegas SUN Column
Nevada Appeal Banner
Elko Daily Free Press Column
tahoe.com Column
NevadaNet Banner

Concord Monitor Online Banner
Foster’s Daily Democrat Banner
N.H. Sentinel Source Column
The UnionLeader.com Frame

The Telegraph Banner

Template

2/C — JL
3/C - JL

3/C - JL
2/C — JL
2/C — JL

21c — JL
nc—JL
uc—JL
mc—JL
nc—JL
3/c — JL
3/c- JL
3/c - JL
3/c — JL
3/c - JL
3/c — JL

3/C — Ctr.

2/C - JL
2/C — JL
2/C - JL

182

MM AudioVideo Anim Stills

No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

No

No

No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No

No

Yes*

No
No
No
No
Yes
No
No
No
No
Yes*
No
Yes
Yes*
Yes
No
Yes*
Yes
Yes

No

Yes

Yes

Yes
Yes
Yes
Yes
Yes
Yes
No

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Yes

APPENDIX C

ONLINE NEWS SITE TEMPLATE

IMAGES IN THIS DISSERTATION ARE PRESENTED IN COLOR.

"[
Bounding-92

LL‘LS’B‘ ‘35

D SE15

D Fnierrainmeni
LEusness __. 7 I
P LIIEBYYIG
P9891?

D Heart or

Web mm

Yuslimrzuiur

/COIS.COM

m

--.Tansing Eagle

the only place you need to look
Headline News:

Apollo 11 Mission Celebrated

‘ Text with mere text with more leitt Err
hi you to put in Ten With more lr-xt

' With more text for you to put in Text

With more text: With more text for you

to put in. Tex: with more next with

Cir-l. her- tru 9.4 It's [high

 

 

 

ripen. u use”, triers- text for you to put in
Text with more text with more test for
you to put in Text With more text
With more text for you to put in
Something Else

Text with more lest with more last for you to put in Text With
more text With more text for you to put ui Text: With more text
With more text for you to put in Test With more text With more
test for you to put in Test with more test With more test for you
to put 11)..

Something Else

Text With more text With more text for you in pill m Ttld With
more text With more test for you to put in Teh'l: With mere text
With more text for you to put 111 Tex: With more tax: With more
test for you to put in Tera: with more tel-rt With more text for you
In [Jul m.

Something Else

Text With more text With more text for you to put in Text With
more text With more text fer you to put in Text: With mere text
With more test for you to put in Text. With more taxi: wilti more
test for you to put in Text with more test With more text for you
to put m.

611!“ Lmnt‘h, NIL-#1: mend

183

 

 

irrriinJ-wblinr

"tili't‘o‘lt
Illllliilifi
’fl]

gap orl he sure

FI‘ '

If!

 

BIBLIOGRAPHY

184

 

Bibliography

ABCNews.com. American Broadcasting Company. 2001.
<http:llwww.abcnews.go.com/sections/uslvideo_index/video_index.htm|/>.

“A Little History of the World Wide Web.” W3C. W3 Consortium. 2000.
<http://www.w3.orngistory.htmI/>.

Baddeley, Alan D. Short Term Memory. New York: Oxford University Press,
1 986.

Baggett, P. and Ehrenfeucht, Andrezj. “Encoding and Retaining Information in
the Visuals and Verbals of an Educational Movie,” Educational
gimmnicajjon am Technology Journal, 31 (1983): 23-32.

Baran, Paul. ”ARPANET—The U.S. DOD-Sponsored Network," The 'Lectric
Learning Wetlgenter. 1995. <http://www.chaos.com/Iearn/History.htmll>.

Barans, Stanley J. ”Media, Media Industries and Media Audiences: Television.”

Introduction to Mass Communication: Media Literacy and Culture.
Mountain View: Mayfield, 2001.

Bielby, Denise B. “Whose Stories Are They? Fans Engagement with Soap Opera

Narratives in Three Sites of Fan Activity,” Journal of Broadcasting and
Electronic Media, 43 (1999): 35-52.

Biocca, Frank and Levy, Mark R., Communication in the Age of Virtual Reality
Hillsdale: Lawrence Erlbaum Associates, 1995.

Borsook, Terry K., and Higgenbotham-Wheat, Nancy. "A Psychology of
Hypermedia: A Conceptual Framework for R&D." Association for
Educational Communications and Technology annual meeting, 5-9 Feb.
1 992: 1 3-14.

Bosco, J. "An Analysis of Interactive Video,” Educational Technology, 25 (1986):
7-16.

185

Bower, G.H., and Glass, A.L. “Structural Units and Redintegrative Power of
Picture Fragments.” Journal of Experimental Psychology: Human Learning
and Memogy, 2 (1986): 456-466.

Bransford, John D. and Stein, Barry S. The Ideal Problem Solver: A Guide to
Improving ThinkingI Learning and Creativiy. New York: W.H. Freeman,
1984.

Brosius, Hans-Bemd. ”Format Effects on Comprehension of Television News,"

Journalism Quarterly, 68.1 (1991): 396-401.

Budd, D., Whitney, P. and Turley, K.J. "Individual Differences in Working Memory
Strategies for Reading Expository Text.” Memom and Cognition, 23
(1995): 735 - 748.

Cable Television Laboratories, Inc. “Cablemodem.org,” Qa_blemogem.org.
CabIeLabs. 2001. <http://www.cablemodem.orgl>.

Cerf, Vinton. CYHIST Community Memory: Discussion List of the History of
Cyberspace, Posted Message. 12 Dec. 1998.

Challis, Bradford, Velichkovsky, Boris and Craik, Fergus. “Levels of
Processingon a Variety of Memory Tasks: New Findings and Theoretical
Implications,” Consciousness and Cognition, 5 (1996): 142-165.

Chun, Dorothy M., and Plass, Jan L. “Research on Text Comprehension in
Multimedia Environments,” Language, Learning and Technology. July
(1997): 60-81.

CNNI, Turner Broadcasting. 1998: <http://www.cnn.com/>. .

Coyle, James, and Brill, Ann. “Examining Metaphor Use in On-Line Advertising.”
Association of Educators in Journalism and Mass Communication Annual
Meeting. Baltimore, MD. 1998.

Cronbach, L. and Snow, R., “Aptitude-Treatment Interaction,” The Theom in
Practice Database. The Encyclopedia of Psychology Society. 1999.
<http2/ltip.psychology.org/>.

186

David, Prabu. “Role of Imagery in Recall of Deviant News," JoumaIism & Mass

Communication Quarterly, 73.4 (1996): 804-820.

DeFIeur, Melvin, Davenport, Lucinda, “Audience Recall of News Stories
Presented by Newspaper, Computer, Television and Radio,” JoumaIism
Quarterly, 69.4, (1992): 1010-1022.

DeFIeur, Melvin L. and Ball-Rokeach, Sandra, “Theoretical Strategies for
Persuasion,” Theories of Mass Communication, Longman: New York,
1989: 276-277.

Enzer, Matisse. “Veronica,” Glossau of lntemet Terms. Matisse Enzer. 2001.
<http://www.matisse.net/files/glossary.html/>.

Ernst & Young, Global Online Retailing: Key U.S. Findings.15 January 2001.

Fidler, Roger. Mediamogphosis: Understanding New Media. Thousand Oaks:
Pine Forge Press, 1997.

Fletcher, D. "The Effectiveness and Cost of Interactive Videodisc Instruction,"
Machine-Mediated Learning, 3 (1989): 361-385.

“Free Data,” FIND/SVP. 1998.
<http://www.cyberdialogue.com/free_data/free_data.htmI/>.

“General Demographics Summary—Primary Place of WW Access,” GVU's
Ninth WWW User §_L_II'VBV. Georgia Tech University. 1998.
<http://www.gvu.gatech.edu/user_surveys/survey-1 998-04/reports/1 998-
04-General.htmI/>.

Gibson, James Jerome, The Senses Crcmsigered as Perceptual Systems.
Boston: Houghton Mifflin, 1966.

Grimes, Tom, “Encoding TV News Messages into Memory," JoumaIism
Quarterly, 67.4, Winter 1990: 757-766.

187

”I

 

Glaser, Candyce Williams and Prestidge, Linda K., “Using Technology to Support
Special Education Teachers” Implementation of the Theory of Anchored
Instruction,” American Educational Research Association Annual Meeting,
Washington, DC. 21 April 1995.

Graber, Doris A. “Seeing is Remembering: How Visuals Contribute to Learning in
Television News.” Journal of Communications, 40.3 (1990): 134-155.

Gunter, Barrie. Poor Reception: Misunderstanding and Forgetting Broadcast
News. Erlbaum & Associates: HiIlsdaIe. 1987.

"GVU's Tenth WWW User Survey Graphs" Graphics Visualization Unit. 2001.
Georgia Tech University.
http://www.gvu.gatech.edu/user_surveys/survey1 9981 Olgraphs/graphs. ht
ml#technologyl>.

Harper, Christopher, AM That’s The Way It Will E, New York: New York
University Press, 1998.

Harrison, Susan M.. "A Comparison of Still, Animated or Non-Illustrated Online
Help with Written or Spoken Instructions in a Graphical User Interface,“
Proc. of Computer-Human Interfaces, 1995.

Hedgedus, Peter. “Stroud’s Reviews: Gopher/ArchielFinger,” intemet.com. INT
Media Group Incorporated. 2001.
<http://www.intemet.com/cws/gopher.htmI/>.

“How Many Online?” Nua lntemet Surveys 18 March 2001. Scope
Communication Group. <http://www.nua.com.ie/>.

“IAB lntemet Advertising Revenue Report Pegs Total at $1.792 Billion For 30

2001.” lntemet Advertising Bureau. lntemet Advertising Bureau. 2001
<http://www.iab.net/>.

“Interactive Advertising Bureau (IAB) Reports $8.2 Billion Ad Revenue in the
United States for Year 2000; Fourth Quarter Totals $2.2 Billion.”

Interactive Advertising Bureau. Interactive Advertising Bureau. 2001.
<http://www.iab.net/>.

188

Infogate. 2001. Infogate Inc. <http:l/www.infogate.com/>.

“lntemet Advertising Revenues More Than Double in 1998.” lntemet Advertising
Bureau. lntemet Advertising Bureau. 1998 <http:l/www.iab.net/>.

Jerram, Peter, “Who is Using Multimedia?” New Media. October 1994: 48-58.

Kearsley, Greg, “Elaboration Theory,” M The_ory in Practice Database, The
Encyclopedia of Psychology Society. 2001. <http:l/tip.psychology.orgl>.

Kearsley, Greg, “Gestalt Theory,” The Theory in Practice Database, The
Encyclopedia pf Psychology Society. 2001. <http:l/tip.psychology.org/>.

Kearsley, Greg, “Symbol Systems,” 313 Theory in Practice Database, The
Encyclopedia of Psychology Society. 2001. <http:lltip.psychology.orgl>.

Kellan, Ann. “lntemet 2 Holds Promise of Technological Leap.” CNN Sci-Tech.
24 Feb. 1999. <http:l/www.cnn.com/TECH/computing/9902/24/intemet2/>.

Kenney, Keith, “Memory and Comprehension of TV News Visuals.” News
Photographer, August 1992: 53-54.

Kobayashi, Susumu, "Theoretical Issues Concerning Superiority of Pictures Over

Words and Sentences in Memory,” Perception and Motor Skills, 3 (1986):
783-792.

Koslowsky, Meni, Lazar, Aryeh and Hoffman, Michael. “Validating an Attitude

Toward Computer Scale." Educational and Psychological Measurement,
48 (1988): 517-520.

Kosslyn, Stephen M. and Pomerantz, James R. “Imagery, Propositions and the

Form of lntemal Representations.” Cognitive Psychology. 9 (1977): 52-
76.

Kristula, David. “The History of the Internet.” Dave’s Site. davesite.com. 1997.
<http:l/www.davesite.comMebstation/net-history.shtmI/>.

189

 

Lee, Seungwhan. “The Convergence of the Web and Television: Current
Technological Situation and Its Future.” Association for Education in
JoumaIism and Mass Communication. Baltimore, MD. 1998.

Lane, David M. “ANOVA.” HyperStat Online. 1999.
<http:l/davidmlane.com/hyperstat/intro_ANOVA.html/>.

Lane, David M. “Logic of Hypothesis Testing.” Hyperstat Online. 2002.
<http:l/davidmlane.com/hyperstatl).

Lane, David M. ”Partitioning the Sums of Squares.” HyperStat Online.1999.
<http:l/www.ruf.rice.edu/~Iane/hyperstatl>.

Lane, David M. “Type I and Type II Errors.” HyperStat Online. 1999.
<http:l/davidmlane.com/hyperstat/A1 8652.htm|/>.

LansingStateJoumaI.com. Gannett News Corporation. 2001.
<http:l/www.Iansingstatejoumal.com/>.

Leiner, Barry M., Cerf, Vinton, 6., Clark, David D., Kahn, Robert E., KIeinrock,
Leonard, Lynch, Daniel C., Postel, Jon, Roberts, Larry G. and Wolff,
Stephen. “A Brief History of the lntemet.” Intemet§ociem. lntemet
Society. 2001. <http:l/www.isoc.org/intemet/history/brief.htmI#Origins/>.

Lewis, Michael. The New, New Thing A §ilimn Valley Stow. New York: W.W.
Norton Publishing, 1999.

Lindlof, TR. and and Schatzer, M.J. “Media Ethnography in Virtual Space:

Strategies, Limits and Possibilities.” Joumal of Broadcasting and
Electronic Me_d§, 42 (1998): 170-189.

Lomicka, Lara L. “To Gloss or Not to Gloss: An Investigation of Reading

Comprehension Online.” Language Learning and Technology. January
(1998): 41-50.

Loyd, Brenda H. and Gressard, Clarice. “Reliability and Factorial Validity of

Computer Attitude Scales." Educational and Psychological Measurement,
44, (1984): 501-505.

190

Loyd, Brenda H. and Gressard, Clarice. ”The Effects of Sex, Age andComputer
Experience on Computer Attitudes.“ AEDS Journal, 17 (1984): 67-77.

Loyd, Brenda H. and Loyd, Douglas E. “The Reliability and Validity of an
Instrument for the Assessment of Computer Attitudes." Educational and
Psychological Measurement, 45 (1985): 903-908.

Lowe, Phillip J. “The Effect of Flashback on Children’s Understanding of
Television Crime Content.” Journal of Broadcasting and Electronic Media,
43 (1999): 83-98.

Lundsten, Apryl and Doiel, Robert. "Internet 2: What is I2?", Builder.com. CNET
Networks Inc. 27 April 1999.
<http:l/www.builder.com/Servers/Intemet2/index.htmI/>.

Maney, Kevin. “High Tech Tablets: Next Step for Newspapers.” _U_§A Todayy Tech
Remrt, Gannett News Corporation. 28 Feb. 1999.
<http:l/www.usatoday.coml>.

Marschack, ML. and Hunt, R.R., “A Re-Examination of the Role of Imagery in

Learning and Memory,” Journal of Experimental Psychology: Learning,
Memogy and Cognition, 15 (1989): 710-720.

Mayer, RE, and Anderson, R.B. "Animations Need Narration: An Experimental
Test of Dual Coding Hypothesis." Journal of Educational Psychology, 83
(1991 ): 484-490.

Mayer, RE, and Anderson, RE. ”The Instructive Animation: Helping Students
Build Connections Between Words and Pictures in Multimedia Learning."
Journal of Educational Psychology, 84 (1992): 444-452.

“Mediainfo.com,” Editor & Publisher Medialnfo Links Online Media Directogy, The
Editor and Publisher Company, Inc. 23 Oct. 2001.

<http:l/www.editorandpublisher.comleditorandpublisher/business_resourc
es/mediastats.jsp>.

Middlestadt, Susan E. and Bamhurst, Kevin G. “The Influence of Layout on the
Perceived Tone of Newspaper Articles,” JoumaIism & Mass

Communication Quarterly, 76 ( 1999): 264-276.

191

 

“Muted and Mixed Public Response to Peace in Kosovo,” Pew Research Center
for the People & the Press. 1999. <http:l/www.people-press.org/>.

Najjar, Lawrence J. “A Review of the Fundamental Effects of Multimedia
Information Presentation on Learning,” Technical Remrt GIT-GVU-95-20.
June 1995.

Najjar, Lawrence J. “Dual Coding as a Possible Explanation for the Effects of

Multimedia on Learning,” Geoggia Tech Graphics Visualization & Usabilig
Qenter Technical Remrt GIT-GVU-95-29. 1995.

Najjar, Lawrence J. “Multimedia Information and Learning,” Journal of
Educational Multimedia and Hypermedia 5 (1996):129-150.

Neuman, W. Russell, Just, Marion R. and Crigler, Ann. N. Qommdn K_now|edgg
News and the Construction of Political Mepning. Chicago: University of
Chicago Press, 1992.

Netradio.com. 2001. <http:l/www.netradio.com/>.

New York Times on the Web, New York Times Company. 1998:
<http:l/www.nytimes.com/>.

“Newspapers Worldwide: North America,” NewsDirectoLycom. Tucows Inc. Dec.
2001, <http:l/www.ecola.com/newslpress/na/usl>.

“Total Revenue From Advertising —1998." Nua lntemet Surveys. Scope
Communication Group. 1999.
<http:l/www.nua.ie/surveys/analysis/graphs_charts/1 9989raphs/ad_reven
ue.html/>.

O'Hara, Kenton and Sellen, Abigail.“A Comparison of Reading Paper and Online
Documents,” CHI 97, Atlanta, Georgia, 22-27 March 1997.

"Online Newcomers More Middle-Brow, Less Work-Oriented: The lntemet News
Audience Goes Ordinary," Pew Research Center for 113 People and the
Press. 1999. <http:l/www.peoplepress.orgltech98sum.htm/>.

192

 

Osborne, D.J., and Holton, D. ”Reading from Screen versus Paper: There is No
Difference,” lntemational Journal of Mdeachine Studies, 28 (1988) 1-9.

Outing, Steve. “Why Web TV Is Not Viewtron Revisited” E&P Online. The Editor
and Publisher Company, Inc. 9-10 Dec. 1996.
<http:l/www.mediainfo.com/>.

Outing, Steven. “Fidler Moves to Kent State.” E&P Online. The Editor and
Publisher Company, Inc. 9 Sept. 1996. <http:l/www.mediainfo.com>.

“Packet Switching.” Free giline Dictionagy of Computing. Imperial College.
2001. <http:l/www.foldoc.org>.

Paivio, Allan, “Coding Distinctions and Repetition Effects in Memory,” ed: Bower,

G.H. The Psychologyof Learning and Motivation, New York: Academic
Press, 1975.

Paivio, Allan. “Dual Coding Theory: Retrospect and Current Status.” Canadian
Journal of Psychology, 45, (1991): 255-287.

Paivio, Allan. Imagegy and Verbal Processes. New York: Rhinehart and Winston,
1971 .

Paivio, Allan, Mental Representations: A Dual Coding Approach, New York:
Oxford University Press,1990.

Paivio, Allan. “Paired-Associate of Learning and Free Recall of Nouns as a
Function of Concreteness, Specificity, Imagery and Meaningfulness,”
Psychological Reports, 20, (1967): 239-245.

Paivio, Allan and Csapo, K. ”Picture Superiority in Free Recall: Imagery or Dual
Coding?” Journal of Cognitive Psychology, 1973:176-206.

Paivio, AIIan, “Spacing of Repetitions in the Incidental and Intentional Free Recall
of Pictures and Words,” Journal of Verbal Learning and Verbal Behavior,
13 (1974): 497-511.

193

Peny, Stephen D. “Inhibiting Speech Through Exemplar Distribution: Can We
Predict a Spiral of Silence?” Journal of Broadcasting and Electronic
Media, 44 (2000): 268-282.

Pluck, G.C., and Brown, R.G. ”Curiosity is Stimulated by Incomplete

Information.” Proceedings of the British Psychological Socieg. 7 (1999):
5.

Price-Waterhouse. ”IAB lntemet Advertising Revenue Report: Third-Quarter

Highlights, 1999.” nteractive Advertising Bureau. March 2000.
<http:l/www.iab.net/>.

"Profile of a Net Buyer.” Cyberstats Online. 1997.
<http:l/www.zdnet.com/zdimaglcyberstats/1 997/06/users. htm|/>.

Proffitt, Brian. “BrowserWatch Stats Station.” intemet.com. INT Media Group
Incoporated. 2001. <http:llbrowsenrvatch.intemet.com/stats/stats.htmI>.

Ramparasad, J. ”Information Graphics in Newspapers,” Newspaper Research
Journal, 12.4 (1991): 92-103.

Reigeluth, Charles M., Instructional Theories in Acflon: Lessons Illustrating
_S_<flected Theories and Models. Hillsdale: L. Erlbaum Associates, 1987.

Rieber, L.P., Tzeng, S., Tribble, K., and Chu, G. “Feedback and Elaboration
within a Computer-Based Simulation: A Dual Coding Perspective,”
American Educational Research Association Annual Meeting. New York,
New York. April 1996.

Rigny, J.W. and Lutz, K.A. “Effect of Graphic Analogies of Concepts in Chemistry

on Learning and Attitude,” Journal of Educational Psychology, 68 (1976):
305-31 1. .

Robinson, John P. and Levy, Mark R. The Main Source: Learning From
Television News, Beverly Hills: Sage. 1986.

 

Rogers, Everett M. Diffusion of Innovations. New York: Free Press. 1995.

194

 

 

Sadoski, Mark. “An Exploratory Study of the Relationship Between Reported
Imagery and the Comprehension and Recall of a Story,” Reading
Research Quarterly, 19 (1983): 110-123.

Sadoski, Mark, Paivio, Allan, and Goetz, Emest T. "A Critique of Schema Theory

in Reading and a Dual Coding AItemative,” Reading Research Quarterly,
26 (1991 ): 463-468.

Salomon, GavrieI. Interaction of Media, Cognition, and Learning. Hillsdale: L
Eribaum Associates, 1994.

Scott, Randall K. and Goff, David H., “How Excitation from Prior Programming

Affects Television News Recall,” JoumaIism Quarterly, 69.4 (1992): 615-
620.

Sentius Corporation. “Guide to Online Publishing for Reading Comprehension -
Section II,” Sentius Corporation. 1997. <http:l/www.sentius.com/>.

Senzig, Don, CYHISTCommdnity Memory: Discussion List on the History of
cmerspace, Posted Message. 7 Nov. 1998.

Simpson, Timothy J., ”Message into the Medium: An Extension of Dual Coding
Hypothesis,” Proc. of lntemational Visual Literacy Association, 12-16
October 1994 :255-263.

Sims, Roderick C. “Engagement, Control and Learner. A Theoretical Appraisal of
Interactivity,” Australian Society for Computers in Learning in Tertiary
Education Conference, Adelaide, 1996.

Smith, Douglas, and Alexander, Robert. Fumbling the Future: How Xerox

InventedJ hen Ignored the First Personal Computer. New York: Morrow,
1988: 89.

Snow, R., “Aptitude-Treatment Interaction as a Framework for Research on
Individual Differences in Learning,” eds: Ackerman, P., Stemberg, R.J.,

Glaser, R. Learning and Individual Differences New York: W.H. Freeman,
1989.

195

Stark, Pegie and Garcia, Mario. Eyes on the News, St. Petersburg: Poynter
Institute of Media Studies, (1991).

Straubhaar, Joseph, and LaRose, Robert. Qommdnicamis Media in the
Information Sociey, Updated Edition. Belmont: Wadsworth, 1997.

Tappendorf, Sean. ”ARPANET and Beyond,” The World-Wide Web Virtual
Library: The lntemet Studies Page. 1995. Asia Pacific Research Online.
<http:llwww.ciolek.comNVWWVLPages/QItyPages/NetStudies.htmI/>.

Theys, Mitchell. “My Home Page.” Purdue University. 1998.
<http:l/garage.ecn.purdue.edu/~theyslstufflmisclcomp_hist/>.

Tolley, B. Stuart. “Newspaper Effectiveness: The Data,” Journal of Advertising
Research, 31, Feb/March 1991: 12.

Treichler, D.G. ”Are You Missing the Boat in Training Aid?” Film and A-V
Communication, (1967): 14-16.

Tulving, Endel and Donaldson, Wayne. “Episodic and Semantic Memory.”
Organization of Memory. New York: Academic Press, 1972.

"TV News Viewership Declines: The lntemet News Audience Goes Ordinary.”
Pew Research Center for the People and the Press. 2001.
<http:llwww.people-press.org/mediarpt.html>.

“TV Viewership Falls: Fall Off Greater for Younger Adults and Computer Users,”
Pew Research Qenter for the People and the Press. 18 March 2001. The
Pew Foundation. <http:llwww.people-press.org/mediarpt.htmI/>.

“U.S. Has 33% Share of lntemet Users Worldwide Year-End 2000 According to
the Computer Industry Almanac.” Computer Industry Almanac. Computer
Industry Almanac Inc. 2001 <http:llwww.c-i-a.com/>.

Vonder Haar, Steven. “AOL Slows Growth as Membership Chums,” nter@ctive
W_e_e_k, 20 May 1996: 15.

196

Walma van der Molen, Juliette H. and van der Voort, Tom HA. “The Impact of
Television, Print and Audio on Children’s Recall of the News: A Study of
Three AItemative Explanations for the Dual Coding Hypothesis.” Human
Qommdnica_tion Resea_rc_h. 26.1 (2000): 3-26.

Wearden, Stanley T., Fidler, Roger, Schierhom, Ann B., and Shierhom, Carl.
”Portrait vs. Landscape: Potential Users' Preferences for Screen
Orientation,” Association for Education in JoumaIism and Mass
Communication Annual Convention, New Orleans, 1998.

Weber, D. and Hamish, J.D. “Visual Imagery for Words: The Hebb Test.” Journal
of Experimental Psychology, 54, (1974): 409-414.

Webster’s New World Dictionagy of American English, 3rd College Edition, ed.:
Neufeldt, Victoria, New York: Prentice Hall, 1994, 271.

Wertheimer, Max, Productive Thinking, New York: Harper 8. Row, 1959.

Wilson, Brent and Cole, Peggy. “A Critical Review of Elaboration Theory.”
Educational Technology Reseerch and Development. 40 (1992): 63-69.

WirLdowsmedi_a.com, Microsoft Corporation. 2001.
<http:llwww.windowsmedia.com/>.

Wolfe, Mark R. “The Wired Lot: Lifestyle Innovation Diffusion and Industrial
Networking in the Rise of San Francisco’s Multimedia Gulch.” Urban
Affairs Review, May (1999): 707-728.

Worden, Aileen, Walker, Neff, Bharat, Krishna and Hudson, Scott. ”Making
Computers Easier for Older Adults to Use: Area Cursors and Sticky
Icons.” CHI '97, Atlanta, Georgia, 22-27 March 1997.

197

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IIIIIIJIIIIIIIHIIIIIIIIIZIIIIIIIZIIII

302