Download - E-Commerce

eatablesurveyorInternet και Εφαρμογές Web

14 Δεκ 2013 (πριν από 3 χρόνια και 7 μήνες)

116 εμφανίσεις

THE EFFECT MULTIMEDIA WEBPAGE DESIGN HAS ON CONTENT
TRANSFER OVER A VERY FAST NETWORK


by



Jon Storslee














A Dissertation Presented in Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy














ARIZONA STATE UNIVERS
ITY



May 2001














THE EFFECT MULTIMEDIA WEBPAGE DESIGN HAS ON CONTENT
TRANSFER OVER A VERY FAST NETWORK

by


Jon Storslee




has been approved


April 2001




APPROVED:



, Chair














Supervisory Committee





ACCEPTED:








Division Director








Dean, Graduate College






iii







ABSTRACT

The goal of this research was to create a set of baseline data that
identified appropriate media for cognitive processing and retention of the
content delivered on very fast networks. The World Wide Web is a relatively
new m
edium for delivery of instructional material and very little research
exists relating to the inclusion of media into Webpages. This report
describes a Webpage media model created using hypermedia, visual literacy,
interface design, cognition, and media re
search. The common theme in the
literature was that exposure to content in multiple media formats enhances
the cognitive processing of content resulting in improved retention.
Logically, there has to be a point at which adding media to Webpages would
beco
me counter productive to the transfer of the content but the literature did
not identify a point of diminishing return. The Webpage model uses
network bandwidth as a factor in determining the influence media has on the
cognitive processing of Web
-
based co
ntent. The model assumes the
cognitive gain that media provides will be negated as cognitive interference
increases with Webpage download time.

This study tested the Webpage model’s assumptions for very fast
networks. The model predicts that Webpages wit
h text, graphics, animation






iv







and sound provides the best content retention and Webpages with text,
graphics and animation are an alternative to Webpages with sound.

This study used a complementary method of quantitative and
qualitative data analysis. The qu
antitative analysis found that students
performed statistically better when Webpages included media instead of just
text. Webpages with text, graphics, animation and sound performed
statistically better than Webpages with text and graphics or Webpages with

text
-
only. Webpages with text, graphics and animation are an alternative to
Webpages that incorporate sound because no statistical difference was found
between the two designs. The qualitative portion of the study supported the
quantitative findings. Ho
wever, thirty
-
seven percent of the respondents
using Webpages with sound felt that the sound was distracting. The findings
imply that Web developers need to be aware of the speed of the network the
students will be using to access the content and how long

the media they are
using takes to download.








v







DEDICATION

This dissertation is dedicated to the teachers and students who worked
or attended Carminati Elementary School from 1993
-
1996. I would have
never appreciated Dr. Bitter’s sage advice about the Inte
rnet and its impact
on education without them. Carminati’s administration allowed me the
freedom to explore using their classroom and their network.

I would also like to dedicate this to Deedee, Adam, Aaron, and
Andrea for getting me started on my disser
tation and for showing me how
important technology is to education, to my mom for helping me fund my
education and to Alicia for putting up with me and helping me edit this
work.






vi







ACKNOWLEDGEMENTS

Dr. Gary Bitter chaired my dissertation committee and his sa
ge advice
about the power of the Internet has proven invaluable. Dr. Bitter gave me
the freedom and resources to pursue research on the Internet. I also
appreciate all the hockey, football, and baseball tickets he gave me over the
years.

Dr. James Midd
leton was the guiding force in my research. The
number of hours he spent with me was greatly appreciated. I know that I
would not have completed this work without his tireless effort and patience.

Dr. Wilhelmina Savenye, Dr. Gloria Wilson, and Dr. Michae
l
Flemister offered numerous suggestions and managed to understand my
topic despite my poor writing skills. Dr. Alex Yu was a very big help with
the statistics and data analysis. And last but not least, a special thank you to
all my friends who kept askin
g me when I was finishing and for all of their
words of encouragement.







vii








TABLE OF CONTENTS







Page

LIST OF TABLES

................................
................................
................................
........

xii

LIST OF FIGURES

................................
................................
................................
......

xiii

CHAPTER


1

Introduction

................................
................................
.............................

1




Statement of the Problem

................................
................................
.....

1




Purpose of the Study

................................
................................
.............

6




Limitations of the Stu
dy
................................
................................
........

7




Definition of Terms
................................
................................
..............

8




Research Question

................................
................................
...............

13




Null Hypothesis

................................
................................
...................

13




Significance of the Study

................................
................................
.....

13


2

Literature Review
................................
................................
..........................

16




Introduction

................................
................................
..........................

16





Cognition and Media
................................
................................
.............

16





Schema Theory

................................
................................
...........

17





Cone of Experience

................................
................................
.....

19





Cognitive

Apprenticeship

................................
...........................

20





Dual Coding Theory

................................
................................
...

21





Multiple Coding Theory

................................
..............................

21





Cognitive Flexibility Theory
................................
........................

22





Multimedia Generative Theory

................................
....................

23





Cognitive Loading

................................
................................
.....

23





Cognitive Overload

................................
................................
......

25







viii







CHAPTER




Page





Cognitive Overhead

................................
................................
.....

26





Cognitive Economy

................................
................................
.....

26





Cognitive Interference

................................
................................
.

27
Hypermedia

................................
................................
..........................

28





History and Evolution

................................
................................
.

28





Website Usability

................................
................................
........

30




Visual Literacy

................................
................................
......................

31





Text, Graphics, and Animation in Electronic Media

...................

32





Graphical User Interface

................................
..............................

32




Summary

................................
................................
.............................

34


3

Webpage Media Model

................................
................................
..................

37




Theoretical Assumptions

................................
................................
......

37




Pilot Study

................................
................................
............................

46




Assessment of Networks

................................
................................
......

47





Model for Slow Network

................................
............................

47





Model for Moderate Speed to Fast Network
...............................

48





Model for Very Fast Network


................................
.....................

50




Very Fast Network
Model
Tested

................................
.........................

52


4

Method


................................
................................
.............................

55




Research Purpose

................................
................................
..................

55




Research Question

................................
................................
................

55




The Null Hypothesis

................................
................................
.............

55




Subjects

................................
................................
.............................

56




Design of Experiment

................................
................................
..........

57







ix







CHAPTER




Page




Instruments

................................
................................
...........................

57




Treatment

................................
................................
.............................

58





Webpage Design

................................
................................
.........

59





Content

................................
................................
.........................

60






The Ma
in Webpage


................................
................................
.....

61





The Linked Webpages


................................
................................

61




Quantitative Data Collection

................................
................................

63





The Design


................................
................................
..................

63





Pretest


................................
................................
.........................

63





Posttest


................................
................................
.......................

64




Quantitative Data Analysis


................................
................................
..

65




Qualitative Data Collection

................................
................................
..

66





The Design


................................
................................
..................

66





The Interview Subjects


................................
...............................

67





The Questions


................................
................................
.............

67




Qualitative Data Analysis


................................
................................
....

68





Field Notes

................................
................................
..................

68





Narrative and Interview Data


................................
......................

69




Summary


................................
................................
............................

69


5

Results

................................
................................
................................
......

71




Overview

................................
................................
.............................

71




Research Question

................................
................................
................

71







x







CHAPTER




Page




The Null Hypothesis

................................
................................
............

72




Quantita
tive Data Analysis

................................
................................
..

72





All Questions

................................
................................
..............

73





Text
-
Only Questions Removed
................................
...................

75





Descriptive Questions Considered

................................
..............

79




Qualitative Data Findings

................................
................................
....

82





Summary of Field Notes

................................
.............................

82





Posttest Descriptive Responses

................................
...................

82






All types of Webpag
es

................................
.......................

83






Text
-
Only Webpages

................................
.........................

84






Text and graphics Webpages

................................
.............

85






Text, graphics, and animation Webpages


..........................

86






Text, graphics, animation, and sound Webpages

...............

88





Interview Responses


................................
................................
...

90




Summary


................................
................................
............................

95


6


Discussion

................................
................................
.............................

96




Implica
tions of the Study

................................
................................
.....

99





Designer of Web
-
based Content

................................
..................

99





Educators
................................
................................
......................

101




Recommendations for Future Research

................................
...............

102

REFERENCES

................................
................................
................................
.............

105







xi







APPENDIX



A

TEXT
-
ONLY PAGES WEBPAGES

................................
.............................

111


B

TEXT AND GRAPHICS PAGES WEBPAGES

................................
...........

125


C

TEXT, GRAPHICS, AND ANIMAT
ION WEBPAGES

..............................

143


D

TEXT, GRAPHICS, ANIMATION, AND SOUND WEBPAGES

..............

162


E

PRETEST WEBPAGE

................................
................................
..................

181


F

POSTTEST WEBPAGE

................................
................................
................

188


G

TEST RUBRIC

................................
................................
.............................

194








xii







LIST OF TABLES

Table





Page



1.

Sample Size, Mean and Standard Deviation of Variables

.............................

73


2.

Generalized Lin
ear Model Analysis

................................
..............................

74


3.

General Linear Models Procedure: Ryan Multiple Range Test

....................

75


4.

Sample Size, Mean and Standard Deviation of Variables

.............................

76


5.

Generalized Linear Model Analysis

................................
..............................

77


6.

General Linear Models Procedure: Ryan Multiple Range Test

....................

78


7.

Sample Size, Mean and Standard Deviation of Variables

.............................

80


8.

Generalized Linear Model Analysis

................................
..............................

80


9.

General Linear Models Procedure: Ryan Multiple Range Test

....................

81









xiii







LIST OF FIGURES

Figure





Page

1.

Cognitive Loading Factors

................................
................................
............

40

2.

Text
-
only Cogn
itive Loading

................................
................................
........

41

3.

Text & Graphics Cognitive Loading
................................
.............................

43

4.

Text, Graphics & Animation Cognitive Loading

................................
.........

44

5.

Text, Graphics, Animation & Sound Cognitive Loading

.............................

45

6.

Slow Network Connection

................................
................................
.............

49

7.

Medium to Fast Network Connection

................................
...........................

51

8.

Very Fast Network Connec
tion

................................
................................
....

54







CHAPTER 1


INTRODUCTION


Statement of the Problem

The introduction of the World Wide Web (Web) as an instructional
delivery tool makes media selection a critical issue for Webpage designs.
Little research is available regarding appropriate medi
a selection for
Webpage design and the resulting effect on the cognitive transfer of content.
Bandwidth (the amount of data the Internet is capable of transferring at any
time) limitations may restrict the type of media used on Webpages because
download ti
me negatively influences the cognitive processing of instruction
by distracting students. Glover (1997a), in his “speed versus design”
discussion, eloquently stated that media selection problem:

“There is a delicate balance between high quality design and

quick
loading pages. This balance is what can separate your page from the
rest. Unfortunately everyone's tolerance for quick loading vs. good
-
looking pages is different. Some people do not mind waiting for a
page to load if they are in for a treat in desi
gn and/or functionality, yet












others simply want to get to the "meat" of the page as soon as possible
without all the bells and whistles.” (p. 1)

The Internet has grown exponentially since the introduction of the
Web in 1991 (Rutkowski, 1997). According to
predictions, this growth
should continue because businesses only recently discovered the power of
the Internet. Unfortunately, bandwidth has grown only 50% a year (Nielsen,
1997a) during the 1990s. If the Internet usage and bandwidth development
trends con
tinue as predicted, the growth of the Web eventually would
outpace the ability of the Internet infrastructure to handle the data transfer
requirements. The Web may already be experiencing a slowdown.
According to a survey conducted by Internet Archive, ap
proximately three
quarters of the most popular servers on the Internet delivered content in a
“timely fashion” (Smith, 1997, p. 1). Former President Clinton’s desire to
put every child on the Internet (United States Department of Education,
1996a) would be

fruitless if the Internet experienced a bandwidth traffic jam
that rendered classroom connections useless.

Compounding the bandwidth problem was the development of new
software tools that made development of high bandwidth media such as
sound, video, and
animation both easier to produce and to include on












Webpages. Web developers were quick to adopt the new multimedia
software because they believed that increased interactivity attracted new
users without alienating current users (Nielsen, 1997b, 2000). This

assumption was questionable because of the longer download time and the
additional software the users needed to incorporate into their Web browsers
before viewing the interactive Websites. Even with new video and audio
compression along with streaming tec
hnology, the file sizes of the media
have been clogging Internet
-
based Local Area Networks (Radosevich &
Fitzloff, 1998).


In 1997, the average user visited 20 Webpages each day and each
page contained approximately 70 KB of data (Smith, 1997, p. 1). Acco
rding
to Smith, the typical Website was approximately 20% HTML text added to
80% images, sounds, and executables. Moreover, with the development of
new multimedia tools, these figures also would increase. The Internet

Archive claimed that the 1,000 most po
pular sites accounted for
approximately half of all the Web traffic (Smith, 1997, p. 1). Imagine how
the Internet would slowdown if all these popular sites adopted the use of
high bandwidth media.













Several steps have been taken toward solving the problems
facing the
Internet. In 1994 when Berners
-
Lee (1997), the inventor of Hypertext
Transfer Protocol (http), founded the World Wide Web Consortium (W3C,
1997) specifically to address problems associated with the growth and
evolution of the Web. Further assis
tance came in 1995 when the National
Science Foundation (National Science Foundation, 1995) allocated funds to
improve Internet infrastructure. Then, in 1996, the United States Congress
passed the Telecommunications Act (United States Department of
Educat
ion, 1996b) designed to promote commercial investment in the
Internet infrastructure.

Although these steps were taken to ease access problems, keeping up
with the Internet’s exponential traffic growth will be difficult for the
Internet’s Backbone provider
s (Nielsen, 1997a). John Sidgmore, Chief
Operating Officer for WorldCom/UUNET Technologies, a major Internet
Backbone provider, stated at the NetWorld+Interop Conference: “Bandwidth
demand is doubling every three to four months. That's a growth rate of
1,0
00% per year" (Gittlen, 1998, p. 1). Finally, Keynote, an Internet
consulting agency, demonstrated that the data transfer rate from the Internet
was approximately 4.5% slower than was portrayed in the data collected












from the previous year (O'Donnell, 1998)
. Bob O'Donnell, a writer for
InfoWorld

magazine, summarized the situation:

“Despite well
-
publicized efforts by many telecom companies to
increase the size of the pipes used across large sections of the
Internet's backbone, the overall throughput on the ne
t for typical Web
content is still abysmal”. (O'Donnell, 1998, p.1)

In 1998, Sidgmore warned, “We haven't seen the worst of bandwidth
consumption yet!” Furthermore, he stated, “When new applications start to
come on the Internet, we will see (data transfe
r) levels we have never seen
before!" (cited in Gittlen, 1998, p. 1). When the increase in Internet users
was combined with the larger file sizes required by the new multimedia
Webpages, a slowdown of the Internet appeared inevitable. Therefore,
Sidgmore c
oncluded, "If you're not scared, then you don't understand!"
(cited in Gittlen, 1998, p. 1).

One solution to the problem of the Internet slowing down was to
minimize the file sizes transferred over the Internet by identifying the most
effective media for
Web delivery. Unfortunately, educational technology
research has not focused on media delivery using the Internet. Other than a
number of anecdotal records about Internet usage in the classroom, little












research is available that identifies the effectivene
ss of media delivered over
the Internet. Because the Web was a new method of delivering media,
research in this area is limited. Historically, the media was stored on CD
-
ROM, hard drive, or a local area network only. Unfortunately, Internet data
transfer
limitations posed new problems that have not been addressed in
existing hypermedia research, such as slow access to the media.

Purpose of the Study


The purpose of this study was to create a set of baseline data that
identified appropriate media for cogni
tive processing of the content on
educational Websites in order to ensure the greatest degree of student
success in content transfer.

A Webpage media model was created because several Webpages on
the Internet offered practical advice on Webpage design, bu
t most of these
pages showed no evidence that any of the advice was grounded in research.
For example, Webmonkey: design:
The Foundations of Web Design

(Veen,
1998), located at Website
http://www.hotwired.com/webmonkey/design/tutorials/tutorial3.html, and
GLOVER.COM: The Web site & Pages of Jeffrey M. Glover

(Glover,
1997a) located at Website http://www.jeffglover.com/ offered practical












advice on Webpage design. However, only
Yale C/AIM Web Style Guide

(
Lynch & Horton, 1997)
located at http://info.med.yale.
edu/caim/manual/
showed evidence of grounding in prior research. Research into Web
usability (Spool, Scanlon, Schroeder, Snyder, & DeAngelo, 1997) illustrated
that Webpages “don’t act like software” (p. 11). The statement implied that
research into softwar
e design might not apply to Webpages. Contrary to that
seemingly popular wisdom, Web usability studies found that text
-
only pages
provided the best access to information and were the most usable (Nielsen,
1997b, 2000; Spool et al., 1997). Web usability was

based on the need to
find specific information from a Website. The results of usability studies,
however, might not apply to the transfer of content. Moreover, the results
might not substantiate the usability as a teaching tool.

Limitations of the Study


Application of the results of the study should be limited to content
delivered over a network defined as very fast. The results from the pilot
study suggested that the addition of sound added interference to cognitive
processing of the content when using
a slower network. The attrition rate of
the subjects who used sound during the pilot study (as described in Chapter
3) was approximately 50% compared to the zero attrition rate during the












study. The study was conducted on computers with all the necessary
software previously installed. If subjects had been required to load additional
software on the computer, they would have experienced cognitive
interference that could have hindered their cognitive processing of the
content. Consequently, the results from
this study cannot be applied to
students who access these Webpages either from home or from computers
that were not properly configured.

The subjects in this study were from Computer Information Systems
classes. A reasonable assumption would be that stude
nts from this field
would have more exposure to computers than the general population.
Consequently, the results may not apply to the general college population.

Definition of Terms

For this study, the following terms were used:

ACT

(Adaptive Control of T
hought): A statistical model that predicts
the transfer of information from short
-
term to long
-
term memory.

Backbone:

The main communications line for a network. The
backbone for a network is the data line that supports the majority of data
transfer over t
he network.













Bandwidth:
The amount of data measured in bits per second that can
be transferred over a data line.

Bit:
The smallest unit of binary (1 or 0) data.

Browser:
A software program designed to access the resources of the
Internet.

CAI

(Computer
-
assi
sted instruction): The use of a computer to help in
the delivery of instruction. Computer usage is designed to supplement
instruction, not to replace the instructor.

CBI

(Computer
-
based instruction): The use of a computer to deliver
instruction.

C.G.I
. (C
ommon Gateway Interface): A program designed to work
with HTML and the Web server’s resources to add interactivity to
Webpages.

Cognitive Loading
: The amount of cognitive resources needed to
process a new piece of information.

DOS

(Disk Operating System):

A text based operating system
typically used on IBM PC compatible computers.













Fast network connection
: A T
-
1 or a T
-
3 data line to the school
combined with 10 megabits per second (Mbps) shared Ethernet connection
to the desktop of the user.

Flash
: A softwa
re tool created by Macromedia that creates animations
that are delivered easily over the Internet.


GUI

(Graphical User Interface): An icon based operating system
designed to replace a text
-
based operating system. Windows and Mac OS
are examples of GUIs.

http

(hypertext transfer protocol): An Internet protocol used to transfer
Webpages over the Internet.

Hypermedia
: Hypermedia is the linking together of documents that
utilize more that one media source.

Hypertext
: Hypertext is the linking (or connecting)
of documents
together using text
-
based connections.

Hypertext Markup Language

(HTML): A simple text based language
that is used to create a viewable document on the World Wide Web.

Imagen
: A non
-
verbal cognitive channel or pathway used to process
new in
formation. An imagen is one of two ways information is processed.













Internet
: The Internet is a series of data lines and computers
connected around the world forming the largest computer network in the
world.

Internet Service Provider
(ISP): An organization

that, for a user fee,
provides access to the Internet.

logogen
: A verbal cognitive channel or pathway used to process new
information. A logogen is one of two ways information is processed.

Moderate speed network connection
: A connection speed of 56,600
b
its per second.

Mbps
: An abbreviation for mega bits per second or million bits per
second.

Pipe:

A pipe is jargon for a network’s bandwidth or data transfer
capability.

Points of Presence

(POP): A term used to describe a network that is
so fast that remo
te locations are virtually in the same room.

Point
-
to
-
Point Protocol

(PPP): An Internet protocol (rule) that allows
computers from remote locations to connect via a phone line.

T
-
1

(Trunk line level 1): T
-
1 lines are data lines capable of
transferring
data at 1.54 Mbps.













T
-
3

(Trunk line level 3): T
-
3 lines are data lines capable of
transferring data at 45 Mbps.

Typography
: The style and arrangement of fonts used.

UNIX
: A text
-
based operating system typically used on servers and
mainframe computers
.

Usabi
lity
: The ease in which a Webpage (or hypertext) document is
used. A Webpage is considered very usable when information is easily
accessible.

Very fast network connection
: A T3 fiber optic connection with at
least a 10 Mbps on demand switched Ethernet conn
ection to the desktop of
the user.

WBI

(Web
-
based instruction)
:

The use of the World Wide Web to
deliver instruction.

Webpage
: A hypertext document stored on the Internet that combines
text and media
that can be read by a browser.

Website
: Webpages tha
t are linked together that is on the same server.

World Wide Web Consortium
(W3C): A non
-
profit organization
founded in 1994 to establish common standards for use on the Internet
.













World Wide Web
(Web or WWW): Hypertext environment that allows
access to res
ources on the Internet.

Research Question


The research was designed to address the following question: Does
the Webpage design model (see Chapter 3) for a very fast network correctly
predict which Webpage design provides the best content transfer? The
W
ebpage model predicts that Webpages with text, graphics, animation, and
sound would provide the best cognitive support resulting in the highest
degree of content transfer.

The Null Hypothesis

The null hypothesis was stated as follows: An analysis compari
ng
pretest and posttest scores of groups of subjects will demonstrate that there
is no significant difference between the Webpage design and the degree of
content transfer.

Significance of the Study

Misanchuk (1995) identified the need for new research in
multimedia
because the changes in technology could make previous research obsolete.
The Internet is a new delivery mechanism for content material; and although












previous multimedia research might be applied, additional research was
necessary to verify that
premise (McManus, 1995.)

Throughout the literature a common theme exists that exposure to
content in multiple formats enhances the cognitive processing of content.
Logically, there has to be a point at which adding media to Webpages would
become counter
productive to the transfer of content. In keeping with the
purpose of this study, to ensure the greatest degree of student success in
content transfer, four Webpage designs were tested to determine which
design provided the best cognitive support for tran
sfer of content. The
designs tested were:

1.

Text
-
only Webpages (Appendix A);

2.

Webpages with text and graphics (Appendix B);

3.

Webpages with text, graphics, and animation (Appendix C); and

4.

Webpages with text, graphics, animation, and sound (Appendix D).

Chapt
er 2, “Literature Review,” is a literature review. Chapter 3,
“Webpage Media Model,” is devoted to creating a research model based
upon the conclusions drawn from the literature review and the pilot study.
Chapter 4, “Method,” is a description of the Webp
age designs and the
research method. Chapter 5, “Results,” is a presentation of the qualitative












and quantitative data gathered during this study and analyzed using this
design method. Chapter 6, “Discussion,” is a discussion of the results from
this stu
dy and a recommendation for further research.













CHAPTER 2

LITERATURE REVIEW

Introduction


Throughout the literature, there exists a common theme that exposure
to content in multiple formats enhances cognitive processing of content
material. Hypermedia or m
ultimedia presentations, as described in current
belief statements, provides multiple representations and, thereby, enhances a
student’s ability to create a mental model that serves to build connections
from prior knowledge to new knowledge. Logically, acc
ording to Tergan
(1997) there has to be a point at which adding collateral media
representations interferes with or becomes counter productive to the transfer
of the content. To provide insight into the type of Webpage media designs
that were hypothesized
to be the most effective for student cognitive transfer
of content, the literature review includes research from cognition and media,
hypermedia, and visual literacy.

Cognition and Media

Wittrock (1992) described the cognitive approach to education as the
process by which an instructor induces a mental model that relates to
previous knowledge. The mental model acts as a bridge or scaffold between





the student’s prior knowledge and the new knowledge. The amount of
cognitive resources needed to process a new

piece of information or bridge
the gap between prior knowledge and new knowledge is a function of the
student’s prior knowledge and the cognitive loading associated with the
presentation of the new material (Mayer, 1997; Sweller, 1990; Sweller &
Chandler,

1994). The following sections will document the cognitive
process of linking prior knowledge to new knowledge and the factors that
influence the cognitive processing of new information.

Schema Theory

Thorndike’s Law of Connection Forming states that a
ll things being
equal, the more exposure to the new knowledge, the stronger the cognitive
link (connection) that is built (Thorndike, 1971). The building of links using
prior knowledge as a foundation evolved into what is commonly referred to
as Schema The
ory.

The Schema Theory of cognition states that the transfer of knowledge
is based upon construction of mental representations that scaffold, or bridge,
the gap between prior knowledge and new knowledge (Mayer, 1987;
Tergan, 1997). Prior knowledge is segme
nted into categories called
schema

that serve as a foundation for the scaffolds. Jonassen (1989) reported that












hypermedia is capable of mapping or building a knowledge structure that
crosses from the expert to the novice. Jonassen also emphasized the
neces
sity of finding scaffolds capable of spanning that identified knowledge
gap. The scaffold bridges the two cognitive structures by overlapping or
connecting the learner’s schema to experts. Once the overlapping starts, the
knowledge mapping can commence.

The Law of Effect stated that the closer the stimulus to usage of the
new knowledge, the stronger the link to prior knowledge. Thorndike’s Law
of Connection Forming and the accompanying Law of Effect implied that
adding supplemental scaffolds provided more

opportunities to bridge the
knowledge gap and successfully map new knowledge (Thorndike, 1971).

The ACT (Adaptive Control of Thought) Theory, developed by
Anderson in conjunction with several colleagues, is a statistical analysis that
predicts the proba
bility that knowledge would be stored in long
-
term
memory (Anderson, 1993, 1996; Anderson, Corbett, Koedinger, & Pelletier,
1995; Anderson, Matessa, & Lebiere, 1997). The ACT model is based upon
the number of times the subject was exposed to the new knowle
dge and the
strength of activation or exposure. The method of exposure and the existence
of any interfering knowledge is a factor in determining the strength of












activation or how well the student retains the new knowledge. The more
exposure and the strong
er the activation translates into a higher probability
that the knowledge is retained. Multimedia offers several tools to aid in the
exposure of new information and increasing activation to the information
without distracting students with repetitive and
boring content (Mayer,
1997). Dale created a model that ranks the different teaching environments
for strength in activating new knowledge.

Cone of Experience

Dale’s “cone of experience” model (Dale, 1969) of instructional
technology, developed in the 19
50s, implied that the closer the teaching
environment was to the “real world” conditions surrounding the knowledge,
the stronger the transfer of information (Dale, 1969; Saettler, 1968, 1990).
Real world experiences provide a broader cognitive foundation t
hat is
needed before moving on to more abstract concepts within that framework.
By using real world conditions, transferring the knowledge from an artificial
environment to the work place is not required. The theoretical construct of
the cone of experienc
e provided by Dale was the prelude to the construct of
cognitive apprenticeship.













Cognitive Apprenticeship

Cognitive apprenticeship (Brown, Collins, & Duguid, 1989) promoted
on
-
the
-
job training (OJT) or training within the same environment in which
the know
ledge was to be used. Real world experiences provided a schema
that was broader and easier to adapt to the work environment. The need to
simulate the real world further supported the findings of research into
sensory experiences, such as hearing, sight, an
d touch conducted by Dale in
the 1950s (Saettler, 1968, 1990; Dale, 1969; Tergan, 1997). Dale found that
by adding sensory input, both learning and retention were improved because
more links from the student’s schema are built to the new knowledge. By
util
izing the same framework that existed in the work environment, the
cognitive loading associated with transferring the content to an unfamiliar
environment was either minimized or eliminated. Creating a virtual
environment that represented the work environ
ment of the apprenticeship
required using several media mixed together. The use of multiple
representations provided a mixture of stimuli. As a result, the student
experienced the same look and feel as found within the work environment.













Dual Coding Theory

Dual Coding Theory was developed in the early 1970s, and was
employed to explain why using graphics or movies was helpful to the
delivery of instruction. Dual Coding Theory stated that students process
information using two channels (Paivio, 1971, 1991; S
aettler, 1990).
According to the theory, one channel of information processes verbal
information called
logogens
, and a second channel called
imagens

processes
nonverbal information (i.e. pictures). Processing information from different
channels creates m
ore scaffolds from which to link to the new knowledge.
Paivio reported that pictures were more likely to be remembered than text
because visuals were processed in both the verbal and nonverbal channel
portions of the brain although text was processed on
ly in the verbal channel.
His research also revealed that concrete words were more likely to be
remembered than were abstract words.

Multiple Coding Theory

The “contiguity principle” was contained within the Multiple Coding
Theory (Mayer, 1997). The princ
iple was used to explain why multimedia
was successful in transferring content material. The contiguity principle, an
extension of Dual Coding Theory, assumed that human beings processed












information using both verbal and visual representations (Mayer, 1997
;
Mayer & Anderson, 1991, 1992; Simpson, 1995). New information,
presented both verbally and visually, established three possible connections
to prior knowledge. First, the information could be linked to verbal
representations. Second, information could be

linked to visual
representations. Finally, information linkages could be made using
combinations of the two references. The contiguity principle also suggested
that information transfer was enhanced when more representations of
knowledge in the form of li
nks and symbols were made (Simpson, 1995).

Cognitive Flexibility Theory

Multiple Coding Theory held that several cognitive channels could be
accessed at once. The Cognitive Flexibility Theory promoted the use of
multiple methods and media systems to repre
sent knowledge and to build
stronger links by accessing several channels. According to cognitive
flexibility constructs, the combination of multiple methods and media
systems provided additional opportunities for the learner to access the
multiple scaffold
s needed to process the information (Jacobson & Spiro,
1995). Consequently, the additional scaffolds improve the ability of the
learner to create a mental representation of the information. Multimedia












offers several tools that can be used to create menta
l representations of new
content that enhance the student’s ability to create mental representations.

Multimedia Generative Theory

Multimedia Generative Theory is an updated version of Whittrock’s
Generative Theory (Mayer, 1997). Generative Theory held tha
t learning was
more meaningful and that stronger cognitive links were created through
active participation of the learner (Wittrock, 1992). Sensory input from the
environment helped the learner build stronger cognitive links and, when
combined with elabora
tion by the student, improved the links to prior
knowledge. The basic theme of Multimedia Generative Theory was “the
design of multimedia instruction affects the degree to which learners engage
in cognitive processing required for learning” (Mayer, 1997, p
. 7).
Multimedia enhances the cognitive processing of new knowledge by
stimulating both verbal and visual processing systems, resulting in more
links being built to prior knowledge. Existing research suggested that the
appropriate use of media enhanced the

cognitive processing of the content.

Cognitive Loading

Cognitive Loading Theory, an extension of Dual Coding Theory,
holds that humans have limited resources for processing new knowledge.












Each individual has a cognitive threshold point that when crossed
, hinders
the ability to process additional information (Kulhavy, Stock, & Kealy,
1990; Sweller & Chandler, 1994). This theory explains why a student’s
content retention will drop when exposed to distracting environment. The
environmental distraction tak
es up cognitive resources the student would
have used to process the content material.

Sweller (1990) and Sweller with Chandler (1994) concluded that
“instructional materials can be restructured to eliminate cognitive activities
that act as impediments to

learning” (Sweller, 1990, p. 129). Therefore,
according to this theory, instructional designers have to take steps to ensure
that the selected media and verbiage will not overwhelm the student’s
cognitive resources. Selecting text and media that create
mental images of
the content is an effective way to use the student’s cognitive resources and
enhance the cognitive processing of new information (Mayer, 1997; Sweller,
1990; Sweller & Chandler, 1994). The addition of graphics to the Webpages
used in this

study (see Appendix B), or animation (see Appendix C), were
designed as examples of media that were included with the intention of
easing the cognitive load placed upon student cognitive resources. The
media in the cited examples depicted complex concept
s that would have












required long textual explanations without the use of graphics.
Unfortunately, poor use of media can overwhelm students and add to the
cognitive loading of the instructional material. However, Preece (1993)
recommended that perception an
d attention be considered when designing
instructional material because cognitive loading will be affected.

Cognitive Overload

Cognitive overload occurs when the cognitive resources of the learner
are overwhelmed, and the processing of new knowledge is h
indered.
Jonassen and Wang (1993) identified cognitive overload as a common
problem with hypertext designs. According to this research, cognitive
overload was caused either by too many choices or by a lack of the cognitive
foundation necessary to process i
nformation about the new environment. An
example of cognitive overload in hypermedia was the presence of too many
(or several poorly organized links) on one Webpage. In those instances, the
user was either confused or overwhelmed by the number of options (
Horton,
2000; Ebersole, 1997). The cognitive reaction resulted in temporary
cognitive paralysis. Excessive simultaneous media interaction, slow content
material delivery, and/or the inability to access Webpages would be












additional examples of situations t
hat could cause cognitive overload.
Unfortunately, students vary in their ability to handle cognitive loading.

Cognitive Overhead


The concept of cognitive overhead was introduced by Conklin (1987)
to describe why some users were experiencing cognitive ov
erload when
using hypertext while other users were successful. The amount of overhead
required for a user, according to Conklin, was based on the cognitive
foundation the user had prior to attempting a task. The stronger the
foundation, the lower the chanc
e for overloading the user’s cognitive ability.
The weaker the foundation, the more likely the user’s cognitive ability will
be overwhelmed. Another potential problem when designing instructional
material was overwhelming users with too many stimuli.

Cog
nitive Economy

Cognitive economy emerged from the work of Tergan (1997). The
report of that study cautioned against overwhelming the user with too much
information and/or too many stimuli. The theory assumed that a limited
number of cognitive resources wer
e available to solve problems. The amount
available for a specific task was dependent upon the amount of cognitive
overhead consumed by the user. This premise about cognitive resources












completed the return phase of the cycle to the cognitive foundation tha
t the
user built prior to performing a task.

Cognitive Interference

According to existing research, the construct of interference played an
important role in the cognitive loading of viewers. Interference occurred
when elements from the screen design eith
er distracted the user or interfered
with user reception of the overall message (Bradshaw, 1997; Sweller, 1990;
Sweller & Chandler, 1994). The distraction could take several forms, such as
the contrast between text and background colors making reading the
content
material impossible. Another distraction is the cognitive inability of the
student to process the content because of insufficient prior knowledge of the
subject. The distraction increased the amount of cognitive loading required
to process a task.

An example of interference occurred when the textual
content of a screen described one concept but the graphical content related to
a different concept. An example of Web
-
based interference is the distraction
caused by long download times required to deli
ver large video or graphic
files (Horton, 2000; Ebersole, 1997).













The media selected for delivery over the Internet must be carefully
chosen so as not to interfere with student comprehension of the information
presented. The goal of the designer must be to

create a Webpage with
minimal distraction. Such a selection process ensured that the presentation
did not influence, in a negative fashion, the cognitive loading of the viewer.

Hypermedia

History and Evolution

Bush (1945), in an
Atlantic Monthly
article,
proposed a new method
of storing and retrieving information. Instead of using a typical filing system
to store data in predefined categories, he believed that mimicking the mind’s
cognitive process of storing data using associative links with previous
cont
ent was more effective. Bush proposed creating a mechanical storage
device called the Memex machine. The Memex machine would have stored
and retrieved data by associative links created by mechanical levers.
However, the technology needed to fulfill his dre
am of an associative
database was not available until the computer revolution of the early 1980s.

Bush’s Memex machine was the prelude to hypertext. Hypertext
quickly evolved into hypermedia when computers became capable of
handling graphics, sound, and an
imation. The World Wide Web (Web) was












the manifestation of Bush’s vision: linking documents together creating a
database with access to resources from around the world. Prior research in
hypermedia such as computer
-
based instruction (CBI), screen design an
d
graphic user interfaces indicated that similar assumptions applied to Web
-
based delivery of instructional material.

The ability of instructors using CBI to provide more representations of
the knowledge through the use of text, graphics, audio, and video
offered the
students more opportunities to enhance learning (Jacobson, Maouri, Mishra,
& Kolar, 1996; Jonassen, 1989). While content delivered over the Internet
was a form of CBI, the research was based on instructional material located
on the hard drive,
stored on CD
-
ROM, or delivered over a local area
network.

In the meantime, unique problems inherent to the delivery of
instruction arose in the use of Web
-
based instruction. These problems have
included slow content delivery as well as the frequent occurr
ence of partial
unavailability of content due to network traffic. Consequently, hypermedia
research needs to be expanded to include factors involving the Web.













Website Usability

Website usability has emerged as a new field of research designed to
determine

what factors make a Website “easy to use” (Nielsen, 1997b, 2000;
Spool, Scanlon, Schroeder, Snyder, & DeAngelo, 1997). Research about
hypermedia, conducted by Nielsen (1997b, 2000), defined “easy to use” as
the ability to access or find Website informatio
n easily. Spool et al. (1997)
took the Nielsen research, completed at Sun Microsystems, and used the
same research criteria to evaluate several Websites on the Internet. The
summation of the findings of both research studies contained the following
implica
tions:

1.

Graphic design (use of graphics) neither helped nor hurt the
usability of a Webpage.

2.

Text links were vital to maintain the usability of a Webpage.

3.

Navigation and content were inseparable to Webpage design.
(The content needs to be lo
gically displayed through the
Website.)

4.

Information retrieval was different from surfing. (Making
information easily accessible on a Webpage is different from
making a Webpage visually appealing to a user.)













5.

Websites were not like software. (i.e. So
ftware design criteria
do not apply to Websites.)


The navigation advice outlined above is designed to lower the
cognitive load needed to process the information on the Webpage. Keeping
the navigational aides consistent, the Web designer allows more cognit
ive
resources to process the remaining information on the page
.

Hypertext evolved into hypermedia when computer system
capabilities developed to the level of simultaneous display of text, graphics,
sound, and animation. The addition of the new media gave d
esigners the
ability to create virtual environments for use in instruction. At that juncture,
media designers started to study the visual media’s ability to support or
distract from the cognitive processing of new information.

Visual Literacy

Visual litera
cy has been defined as “the ability to understand and
produce visual messages” (International Visual Literacy Association, 1995).
Existing research on the topic of visual literacy

specifically the topic of
screen design

provided a foundation from which the

design of a Webpage
could be evaluated.













Text, Graphics, and Animation in Electronic Media

Several researchers (Bradshaw, 1997; Morrison, Ross, & O’Dell,
1991; Hathaway, 1984;) reported that large amounts of text were likely to be
forgotten, especially w
hen displayed in an electronic format. Both Dual
Coding and Multimedia Generative theories suggested that retention of the
content material was enhanced by the appropriate use of media. Kemp and
Dayton (1985) stated that graphics added variety, thus making

the
information presented more interesting. The student’s increased interest
resulted in better attention and increased motivation to learn the subject
matter. In addition to improved attention, graphics created a visual metaphor
of content material that
helped build a mental model of the subject matter
(Soulier, 1988). If graphics help build a mental model, logically, animation
should be equally effective. Reiber (1991) confirmed that assumption by
finding that “animation was a significant aid to learning

as compared to
static graphics”(p. 12).

Graphical User Interface

The evolution of the personal computer operating system exemplified
the practical application of how graphics were effectively mixed with text to
increase the conceptualization of an abstra
ct concept such as the












organization of files on electronic media. The early versions of personal
computer operating systems were Microsoft Disk Operating System (DOS),
Apple DOS, and UNIX. A Graphical User Interface (GUI) that incorporated
text and graphic
al icons eventually replaced these text
-
based environments.
Mixing text and graphics together to form icons resulted from interface
research conducted by Xerox, Apple, and Microsoft

(
Lynch, 1994b). The
success of GUIs, as opposed to text
-
based systems, sug
gested that users
wanted to interact with graphical environments even though such
environments required more resources and were slower than text
-
only
environments
.
Schnotz (1996) and Mayer and Anderson’s

(1992) research
supported the need to use combined t
ext and graphics to make
communication and learning more effective. Research completed by Mayer
(1997) demonstrated that the “verbal and visual explanation of how a system
works will ensure that students will understand the explanation” (p. 7).
Lynch (1994
a, 1994b, 1994c) reported that the success of GUIs was
attributed to the ability of the user to create a conceptual model of what to
expect.

The GUI created the conceptual model by mixing visual and textual
metaphors that the user understood. An example
of the GUI visual metaphor












was the trashcan on the Macintosh operating system. When users wished to
delete files, they simply dragged the files into the trashcan and then chose
the command to empty the trash. The trashcan was a visual metaphor that
represe
nted throwing out files. Multimedia programmers’ mixed text and
graphics with sound to reinforce the messages of the programs. Although
use of this mixture suggested that using more media enhanced student
retention of knowledge, designers needed to avoid o
verwhelming the users
with excessive use of media.

Summary

The most efficient cognitive learning environment, according to
existing research, was the real world (Brown, Collins, & Duguid, 1989;
Saettler, 1968). Cognitive transfer was more likely when the s
ubject matter
was learned in an environment that represented actual conditions. Students
were able to build cognitive links from prior knowledge to real
-
world
practices. Findings demonstrated that media had several attributes that
enhanced the learning env
ironment.

Dual Coding and Multiple Coding theories explained why using
media to stimulate the senses increased the cognitive processing of new
knowledge by the students (Mayer, 1997; Paivio, 1991). The Generative












Theory of Multimedia stressed the positive

effect the inclusion of
supplemental media had on the ability of students to achieve successful
cognitive processing of content (Mayer, 1997).

Content retention was enhanced when the cognitive processing of the
new knowledge created scaffolds that linked

new information to prior
knowledge. By reducing the cognitive load, instructors created learning
environments in which students had more cognitive resources available to
process new information and build scaffolds into long
-
term memory
(Sweller & Chandler
, 1994). Consequently, the goal of Webpage designers
should be to lower the cognitive load of new content material by carefully
selecting the type and amount of media to be added to the Website.

Cognitive theories suggest that designers of hypermedia enco
untered a
point at which adding collateral representations of knowledge or subject
matter became counter productive. However, none of the theories identified
a method for determining a threshold point in hypermedia designs.
Moreover, the research demonstra
ted that the ability to process information
or to handle cognitive loading varied with each user. Web
-
based instruction
faced the problem of cognitive loading caused by the interference generated
with excessive download time. What remained necessary, yet a
bsent, was a












definition of the optimal balance between the cognitive load of processing
additional symbols and the need for media to enhance the building of mental
representations for knowledge transfer.

Chapter 3 is devoted to describing the research mod
el created from
the conclusions drawn from the literature review. The first section,
“Theoretical Assumptions,” discusses the need for media on Webpages
while the “Pilot Study,” highlights the importance of knowing the speed of
the network that delivers t
he content. The third section, “Assessment of
Networks,” defines the criteria for judging a network’s speed that is
incorporated in the “Model Selection.”













CHAPTER 3


WEBPAGE MEDIA MODEL


A Webpage research model was created from hypermedia, visual
lite
racy, cognition, and media research. The model used bandwidth (the
amount of data the Internet is capable of transferring at any time) as a factor
in determining the influence media could have on the cognitive processing
of Web
-
based content material.

The
oretical Assumptions


The Webpages that provide the lowest cognitive loading for
processing of the content material would have the highest retention rate of
all the pages. As described in Figure 1, the cognitive loading associated
with processing of cont
ent delivered over the Web was initially divided into
three factors:



The loading associated with processing the content regardless
of media



Type and number of media used to present the content



Download time required for delivery of the content













The Webpage’
s cognitive loading to process the content would vary
dependent upon the usage of media and its download time. In addition, the
cognitive loading associated with the student’s processing of the content
could be divided into four factors.



Prior knowledge o
f the content



Complexity of the new knowledge relative to prior knowledge



Prior experience using the Web and computer delivered content
(more experience results in lower cognitive loading)



Miscellaneous loading factor (stress, environmental
consideration,
mood, attitude towards content)

The model assumes that the cognitive loading to process the content
material will be reduced as media is added to the Webpages. Cognitive
loading is reduced because the media enhance the student’s ability to create
a mental

model of the content material. The creation of a mental model
results in more efficient content processing (Soulier, 1988).

Figures 2 through 5 represent the proportion of cognitive loading
required for the Webpage download time, media processing and con
tent
processing for a very fast network. The addition of media will add to the
Webpage’s total file size resulting in longer download times. The addition












of more media will also increase the amount of loading required to
understand the concepts depicted
by the media. The model assumes that the
addition of media will be complementary and will reduce the amount of
processing load for each media item.

Web usability studies suggested that text
-
only Webpages were
sufficient for finding or gathering informat
ion, and that the very fast
download times of text
-
only Webpages provided minimal cognitive
interference as can be seen in Figure 2.

Unfortunately, however, large amounts of text were likely to be
forgotten

especially when displayed in an electronic format

(Bradshaw,
1997; Hathaway, 1984; Morrison et al., 1991). Using graphics made content
more interesting (Kemp & Dayton, 1985) and enhanced the creation of
mental representations that helped the cognitive processing of new
knowledge (Soulier, 1988). The mo
del assumes that the addition of graphics
to Webpages would lower the cognitive loading needed to process the
content material. The cognitive interference generated by the longer
download times would not negate the cognitive gain made by the addition of
g
raphics as illustrated in Figure 3. The lower cognitive loading would result
in a higher content transfer rate.






































The model predicts that addition of graphics to text
-
only
Webpages would enhance content retention regardless of the speed of the
network.

Media research suggests animations would provide additional
cognitive support (Mayer, 1997; Mayer & Anderson, 1992; Soulier, 1988)
for the processing of content. Consequently, it was predicted that adding
animations would improve the cognitive transfer
of the content as depicted
in Figure 4.

According to Reiber (1991), animation had the ability to present
complex concepts and relationships that static graphics could not convey.
The drawback of using animation on Webpages consisted of the requirement
for

larger file sizes to be delivered over the Internet. The research findings of
Mayer (1997) suggested that cognitive processing of concepts presented in
animation was enhanced by the addition of sound. However, Mayer’s
findings might not apply because the
download time for sound may generate
cognitive interference that negates the cognitive gain of adding sound.























































The reduction in cognitive loading created by the addition of sound
might be offset easily by the interference generated by slow download
s.
Therefore, any cognitive model for Webpage design must factor in the speed
of the network that would deliver the content. Figure 5 assumes the content
was delivered using a very fast network.

The pilot study conducted in preparation for this study v
erified that
speed of the network does influence a Webpage’s cognitive loading.

Pilot Study

A pilot test of the delivery of the treatment and the data collection
procedures was conducted at a major southwestern university. The network
was slower than the

one used in the final study. Sixty graduate and
undergraduate students from computer classes participated in the pilot study.
The pilot study uncovered a potential problem with the delivery of the
Webpages with text, graphics, animation, and sound. The
long download
times for these Webpages resulted in a dropout rate of 50%. The students
expressed frustration with the study and they requested permission to
terminate the testing. The longer download times tended to confuse the
subjects, resulting in a
higher cognitive load and lower content transfer. The
subjects thought that the Webpages were corrupt when the pages did not












appear immediately. Consequently, the frustration level of the subjects
increased significantly, as did their cognitive loading.

The pilot study results
suggested that the bandwidth of the network should be considered carefully
when selecting media for inclusion in Webpage designs.

Assessment of Networks

During the investigation to determine the appropriate network to use
for the
study, several network characteristics were defined. A slow network
was defined as a connection of 14,400 bits per second to 28,800 bits per
second. A moderate connection was defined as 56,600 bits per second. A
fast connection was defined as a T
-
1 or a

T
-
3 data line to the college
combined with 10 megabits per second (Mbps) shared Ethernet connection
to the desktop of the user. Finally, a very fast connection was defined as a
T3 fiber optic connection with at least a 10 Mbps on demand switched
Ethernet

connection to the desktop of the user.

Model for Slow Network

The large file sizes for Webpages with animation or animation with
sound caused excessive cognitive interference. The model assumed that the
cognitive load from the excessive download time o
ffsets any gain in content
processing the media provided. Even at slow download speeds, the content












required graphics to help the processing of the content and add variety to the
instruction. Consequently, the most efficient method

to transfer content ov
er
a slow network was restricted to Webpages with only text and graphics as
shown in figure 6.

Model for Moderate Speed to Fast Network

Once again, text
-
only pages did not provide the cognitive support
required for processing complex concepts. Because th
e network connection
was fast, the model assumed that graphics would improve the cognitive
processing of the material but not as much as did Webpages with animation
added. However, Webpages with animation and sound required file sizes
that were sufficientl
y large to create the download interference issue. The
pilot study was conducted on a network that met the fast criteria and the
subjects experienced problems related to download interference.
Consequently, the results of the pilot indicated that cognitive

gain because of
the addition of sound was negated by the interference created by excessive
download times.



























Existing research suggested that Webpages had to balance the type of
media used against the interference caused by the download time of the
med
ia. As demonstrated in the examination of the different types of
networks, the measurable advantages of faster speeds became apparent as
seen in figure 7. Webpages with text, graphics, and animation were
predicted to provide the best content retention.

M
odel for Very Fast Network

Regardless of the speed of the network connection, text
-
only pages do
not provide the cognitive support required for complex concepts.
Furthermore, the excessive use of text in electronic form hinders the
processing of the conten
t material (Bradshaw, 1997; Hathaway, 1984;
Morrison et al., 1991).

Text and graphics increase the probability that a mental model was
created because the graphic provided a visual representation of that mental
model. Animation, in conjunction with text a
nd graphics, provides a more
detailed mental representation of relationships of the content material.

Mayer and Anderson (1991) suggest that users benefited from
animation combined with sound.


























Webpages with text, graphics, animation, and sound on a ver
y fast network
are not likely to experience the same amount of cognitive interference as
generated by a slower network’s longer download times. As a result, the
media’s download time is not expected to adversely affect the cognitive
loading of the student
s.

Assuming the appropriate use of media, the Webpage design model
predicted that for a very fast network connection, Webpages with text,
graphics, animation, and sound would provide the best cognitive support and
lowest cognitive loading. Consequently, We
bpages with those media would
result in the best content retention (see Figure 8)
.

Very Fast Network Model Tested

The network used in this study met the criteria for a very fast network
(T3 fiber optic connection with at least a 10 Mbps on demand switched
Ethernet connection to the desktop of the user). Because the network is
extremely fast, the interference generated by larger download times required
for sound will not outweigh the cognitive benefit to the user. Consequently,
it was predicted that Webpag
es with text, graphics, animation, and sound
would result in the best transfer of content. Webpages with animation will
provide more cognitive support and better content transfer than did the text
-












only Webpages or Webpages using text and graphics. Howeve
r, the text,
graphics, animation, and sound Webpages provided the most cognitive
support for the transfer of content.

Chapter 4, Method, includes a description of how the Webpages were
designed and details how this study was conducted on a very fast netw
ork.


















CHAPTER 4

METHOD


Research Purpose


The purpose of this study was to create a set of baseline data that identified
appropriate media for cognitive processing of the content on educational
Websites to ensure the greatest degree of student success in

content transfer.

Research Question


The research was designed to address the following question. Does
the Webpage design model described in Chapter 3 for a very fast network
correctly predict which Webpage design provides the best content transfer?
Th
is Webpage model predicts that Webpages with text, graphics, animation,
and sound would provide the best cognitive support resulting in the highest
degree of content transfer.

The Null Hypothesis

The Null hypothesis is: An analysis comparing pretest and
posttest
scores of groups of subjects will demonstrate that there is no significant
difference between the Webpage designs and the degree of content transfer.













Subjects

Ninety subjects from undergraduate classes in Computer Information
Systems at a communi
ty college in Arizona were selected as subjects for
this study. The classes consisted of first
-

and second
-
year students plus
adults returning for retraining who were taking introductory computer
classes in Web publishing, networking, or computers in busi
ness. The
subjects possessed varying degrees of expertise both in computer
technology and in Website access experience. The subjects required
between 45 minutes to an hour and a half to complete the test.

The research was scheduled at different times thr
ough out the day to
ensure that different networking congestion was encountered. Two classes
of subjects were tested in the morning, one at 9:30 a.m. on March 9 and the
second at 11:30 a.m. on March 7. Two classes of subjects were tested during
mid
-
day, on
e at 1:30 p.m. on March 9 and the second at 4:30 p.m. on March
9. One night class was tested at 6:30 p.m. on March 3.

A server error encountered during the testing sessions on March 9
(from 9:30 a.m. to 3:00 p.m.) resulted in the loss (deletion from the
electronic database) of pretest responses from 22 subjects. Six Webpages
with text, graphics, and animation in addition to the Webpages with sound












were lost, as were seven Webpages with text and graphics. Of the several
text
-
only Webpages, only three were

lost at the time of the server error. The
server delivered all the posttest responses successfully. The loss of data
marginally lowered the design’s sample size and the effect size.

Design of Experiment


This research study used a 4 x 2 factorial design

with one
between
subject
factor and one
within subject
factor. The
between subject
factor was
type of media, which contained four levels: (a) text
-
only; (b) text and
graphics; (c) text, graphics, and animation; and (d) text, graphics, animation,
and soun
d. The
within subject
factor contained two levels: (a) pretest and
(b) posttest. Given the design of the experiment, the following parameters
were established:

1.

The effect size was .4.

2.

The alpha level was .05.

3.

The sample size was 68.

4.

The power for this stu
dy was .85.













Instruments


The pretest and posttest developed by the instructor for this study
were used as performance tests. The performance test was comprised of
questions selected from previous tests and from networking textbooks. The
test was composed

of 19 items, of which the content for seven questions
was presented using only text (even on pages with media); and twelve
questions were presented using text and media. The pretest established the
subject’s prior knowledge of the content. The posttest
consisting of the
same questions was used to measure the effect of the treatment. The