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 I DATE DUE DATE DUE
5

W kw

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8/01 enemas-p.15

ADVERTISING IN INTERACTIVE TELEVISION:
HOW AUDIENCES’ INTERACTIONS WITH ADS AFFECT
PERCEPTIONS OF PROGRAMS AND BRANDS
By

Joo-Hyun Lee

A DOCTORAL DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Advertising

2003

ABSTRACT
ADVERTISING IN INTERACTIVE TELEVISION:

HOW AUDIENCES’ INTERACTIONS WITH ADS AFFECT
PERCEPTIONS OF PROGRAMS AND BRANDS

By

JOO-Hyun Lee

Consumers are passively exposed to most commercials in a conventional
television environment unless they take active steps to avoid them. On the contrary,
interactive product placements (iPPL) in an interactive television (ITV) environment
would deliver the advertising message only upon the consumer’s request. Traditional
product placements (PPL) aim to put the audience in contact with the advertiser’s brand
through simple exposures, so traditional PPL studies focus on the effect on the
consumer’s memory. This dissertation recognizes that the iPPL can generate actual
interactions from the consumer while watching the program, and examines how the iPPL
and its interactive natures can change the audience’s consumption patterns of programs
and the advertising effectiveness in an ITV environment.

The primary purpose of this dissertation is to predict the types of products,
programs, and audiences that would generate improved effectiveness of iPPL before ITV
becomes widely diffused. In particular, this dissertation examines the audience’s
involvement with product categories, involvement with the program, attitudes toward the
program, and attitudes toward the characters in the program in relation to the audience’s
attitude toward the brand and the interactions made to iPPLs. An experiment was
conducted in a computer lab with 396 participants, and an interactive ITV interface

created for the experiment was played on computer screens.

The results indicate that a consumer with higher levels of involvement with a
product category could recall more brands and demonstrated more active interactions
with the iPPL compared to the consumers with lower levels of product involvement. This
dissertation proposed that higher levels of involvement with a program will cause a lower
amount of interactions and recalled brands, but the results show that the program
involvement and the amount of interactions are positively associated. Consumers’
positive attitudes toward the program and a character in the program are found to be
transferred to the attitude toward the brands. Consumers who demonstrated more active
interactions with iPPLs showed a bigger increase in their levels of involvement with the
program compared to the consumers who demonstrated less active interactions.

Finally, structural equation models were analyzed to investigate the overall
relationships of the variables. The results were specific to each particular brand and
character. These findings have implications for advertisers and the current television

industry.

 

Copyright by
JOO-HYUN LEE,

2OO3

 

DEDICATION

TO the Lord, who started me on the way, led me to this place,

and guided me all the way to this very moment.

ACKNOWLEDGMENTS

I would like to thank the professors in my dissertation committee, Dr. Bonnie B. Reece,
Dr. Steven M. Edwards, Dr. Carrie Heeter, and Dr. Hairong Li, who have always been an
invaluable resource, taking turns at being tough critic, sage counsel, and reliable friend. I
have to deliver special thanks to Dr. Bonnie B. Reece, my advisor and mentor, for her
unflinching supports and for carefully guiding me through all phases of my doctoral
program. I thank Dr. Carrie Heeter for her invaluable comments and insightful
discussions about the nature Of interactivity. I owe a debt of gratitude to Dr. Steven M.
Edwards and Dr. Hairong Li, who, first as teachers and later as co-authors, taught me

much Of what I know about interactive advertising today.

I am grateful to my colleagues in the Mass Media Ph.D. Program for all their support and
friendship. In particular, I would like to thank James for the hours he spent discussing
life and research with me. I am also grateful to ZOO Hyun and his wife, Yoon-Keun, for

the countless meals and help they provided.

This research was financially supported by Dissertation Competition Grant from the
American Academy of Advertising (AAA) and by Research Grant from the Ad Research
in Korea. I would like to place on record my deep appreciation for Dr. Kristina
Frankenberger, Chair Of the AAA Research Committee, and for Dr. WOO-Hyun Won,
Editor of the Ad Renard) for the supports they provided, which made this dissertation

possible.

vi

I cannot thank my family enough for their continued love and encouragement. In
particular, I thank my parents and my brother, Sanghyun, for being patient cheerleaders
from halfway around the world. I am deeply grateful to my wife Jounghae for always
being there and pointing me forward when the going got tough, and for all the sacrifices
she made Of her valuable time in order to facilitate the completion of my doctoral

program.

Vii

TABLE OF CONTENTS

LIST OF TABLES xi
LIST OF FIGURES xiv
Chapter 1. Introduction 1

1.1. Interactive Television (ITV) 2

1.2. Television Commerce (T-commerce) 3

1.3. The Purpose of the Study 4
Chapter 2. Interactive TV: Description, History, & Advertising 6

2.1. What is ITV? 6

2.2. History of Development 10

2.3. Forecast 15

2.4. Ads in ITV 17

2.5. Product Placements 19

2.6. Product Placements in ITV 25
Chapter 3. Involvement 27

3.1. Involvement in General 27

3.2. Involvement with Product Categories 30

3.3. Involvement with Programs 32

3.4. Attitude toward the Program and Characters 36

3.5. Gender 40
Chapter 4. Interactivity and Interaction 42

4.1. Interactivity - The Construct 44

4.2. Interactivity - Other Effects and Antecedents 48

4.3. Interactivity and Interaction 50

4.4. Value of Interaction 56

4.5. Interaction — Antecedents 58

4.6. Interaction — Consequences 62
Chapter 5. Methodology 66

5.1. Analysis Plan 66

5.2. Design and Sample 66

5.3. Development of the Stimulus Material 68

5.4. Procedure 73

5.5. Measurement 75

viii

Chapter 6. Results

6.1.
6.2.
6.3.
6.4.
6.5.
6.6.
6.7.
6.8.
6.9.
6.10.
6.11.
6.12.
6.13.
6.14.
6.15.

6.16.

6.17.

6.18.

6.19.

6.20.
6.21 .

Manipulation Check

Sample Size and Composition

Scale Reliability

The Effects of Product Involvement on Interactions
The Effects of Product Involvement and Attitude toward
the Program on Recall

The Effects of Program Involvement on Recall and
Interactions

The Effects of Attitude toward the Program on
Interactions

The Effects of Attitude toward the Program on
Attitude toward Brands in the Program

The Effects of Attitude toward the Characters (of the
Program) on Interactions and Recall

The Effects of Attitude toward the Characters on

the Attitude toward Paired Brands

The Effects of Gender on Interactions

The Relationship between Recall and Interactions
The Effects of Interactions on Changes in the Level
of Product Involvement

The Effects of Interactions on Changes in the Level
of Program Involvement

The Interaction Effects of Product Involvement and
Program Involvement on Interactions (Study 1)

The Interaction Effects of Product Involvement and
Attitude toward Program on Interactions and
Attitude toward Brands (Study 2)

The Interaction Effects of Program Involvement and
Attitude toward Program on Interactions and
Attitude toward Brands (Study 3)

The Interaction Effects of Attitude toward Program
and Attitude toward Characters on Interaction and
Attitude toward Brands (Study 4)

The Interaction Effects of Product Involvement and
Attitude toward Characters on Interactions and
Attitude toward Brands (Study 5)

Test of the Hypothetical Model and Structural Relations
SEM: Results

ix

78
80
80
80
83
84
85
86
89
91
93
93
94
95

96

98

102

105

115
124
128

78

Chapter 7. Conclusions and Discussion

7.1. The Role of Involvement in Advertising Effectiveness
7.2. The Role of Attitude in Advertising Effectiveness

7.3. The Relationship between Attitude and Involvement
7.4. The Role of Interactions in Program Consumption
7.5. Limitations and Suggestions for Future Studies

7.6. Conclusions and Implications

APPENDIX A. Questionnaire for the Experiment
APPENDIX B. Instruction for the Experiment
APPENDIX C. Descriptive Statistics

BIBLIOGRAPHY

131
133
136
136
140
144

149

165

177

182

131

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 6.4.1.

Table 6.4.2.

Table 6.4.3.

Table 6.5.1.

Table 6.5.2.

Table 6.6.

Table 6.7.

Table 6.8.1.

Table 6.8.2.

Table 6.8.3.

Table 6.9.1.

Table 6.9.2.

Table 6.9.3.

Table 6.9.4.

Table 6.10.1.

LIST OF TABLES

Hypotheses

Program used for the Main Experiment

Summary of Product Information Embedded in iPPLs
Experimental Procedure

Manipulation Check of Program Involvement

Product Involvement

Descriptive Statistics and Reliabilities for Scales Used in the Study
Descriptive Statistics on Product Involvement (BG) and Interactions
Relationship between Product Involvement and Interactions
Descriptive Statistics on Brand Recalls

Relationship between Product Involvement, ApRog, and Recalls
Interactions and Recalls in Program Involvement Conditions
Relationship between APROG and Interactions

Calculation of Interacted- and Non-interacted A3: Example

AB Changes in Non-Interacted Brands: Statistical Significance
Regression: Relationship between A3") and Ammo

Descriptive Statistics on Character: ACHAR and Interactions
Relationship between ACHAR and Interactions

Descriptive Statistics on ACHAR and Recalls of Paired Brands
Regression: Relationship between ACHAR (combined) and Recalls

Descriptive Statistics on Character: ACHAR and Interactions

xi

67

71

72

75

78

79

81

82

83

84

85

85

87

88

88

90

90

91

91

92

Table 6.10.2.

Table 6.11.

Table 6.12.

Table 6.13.1.

Table 6.13.2.

Table 6.14.1.

Table 6.14.2.

Table 6.15.

Table 6.16.1.

Table 6.16.2.

Table 6.16.3.

Table 6.17.1.

Table 6.17.2.

Table 6.18.1.

Table 6.18.2.

Table 6.18.3.

Table 6.18.4.

Table 6.18.5.

Table 6.18.6.

Table 6.18.7.

Table 6.18.8.

Table 6.18.9.

Table 6.18.10.

Relationship between ACHAR and AB

Amount of Interactions Across Genders

Regression: Relationship between Interactions and Recalls
Descriptive Statistics on Product Involvement Changes
Relationship between Interactions and Product Involvement
Descriptive Statistics on Program Involvement Changes
Relationship between Interactions and Program Involvement
Interactions Effects: Product Involvement x Program Involvement
Attitude toward Program: Creating Conditions

Interactions Effects: Product Involvement x Program Attitude
Interactions Effects: Product Involvement x Program Attitude
Interactions Effects: Program Involvement x Program Attitude
Interactions Effects: Program Involvement x Program Attitude
Attitude toward the Character: Creating Conditions
Interactions Effects: APROG x ACHAR: Chandler

Interactions Effects: APRQG x ACHAR: Rachel

Interactions Effects: ApROG x ACHAni Ross

Interactions Effects: APROG x ACHAR: Joey

Interactions Effects: APROG x ACHAR: Monica

Interactions Effects: APROG x ACHAR: Phoebe

Interactions Effects: APROG x ACHAR: Chandler

Interactions Effects: ApROG x ACHAR: Joey

Interactions Effects: APROG x ACHAR: Monica

xii

92

93

93

94

95

96

96

98

99

100

101

103

104

106

106

108

109

110

111

111

112

112

113

Table 6.18.11. Interactions Effects: APROG x ACHAni Phoebe

Table 6.18.12. Interactions Effects: APROG x ACHAR: Rachel

Table 6.18.13.

Table 6.19.1.

Table 6.19.2.

Table 6.19.3.

Table 6.19.4.

Table 6.19.5.

Table 6.19.6.

Table 6.19.7.

Table 6.19.8.

Table 6.19.9.

Table 6.19.10.

Table 6.19.11.

Table 6.19.12.

Table 6.19.13.

Table 6.21.1.

Table 6.21.2.

Table 6.21.3.

Interactions Effects

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

Interactions Effects:

:ApROG x ACHAR: Ross

Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:
Product Involvement x ACHAR:

Product Involvement x ACHAR:

Summary of the Results

Initial Result: Inexpensive Appliance / Joey

Initial Result: Kitchenware / Monica

Initial Result: Dessert / Phoebe

xiii

Chandler
Joey
Monica
Phoebe
Rachel
Ross
Chandler
Ross
Joey
Monica
Phoebe

Rachel

113

114

114

116

116

117

117

118

118

119

119

122

122

123

123

124

129

130

130

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 6.18.1.
Figure 6.18.2.
Figure 6.18.3.
Figure 6.19.1.
Figure 6.19.2.

Figure 6.20.1.

Figure 6.20.2.

LIST OF FIGURES

History of ITV Development

Thumbnail of Stimulus Material: Step 1

Thumbnail Of Stimulus Material: Step 2

Thumbnail of Stimulus Material: Step 3

Interactions Effects: ApROG x ACHAR: Chandler

Interactions Effects: ApRog x ACHAR: Rachel

Interactions Effects: APROG x ACHAR: Ross

Interactions Effects: Product Involvement x ACHAR: Chandler

Interactions Effects: Product Involvement x ACHAR: Ross
Hypothesized Model

Tested Model

xiv

14

68

69

70

107

108

109

120

120

125

127

Chapter 1. Introduction

Television is a passive and one-way mass communication medimn that has been
providing entertainment and information for millions of people around the world.
However, television is changing and has recently started to become an interactive two-
way communication platform. The interactive television (“ITV”) industry started to take
shape in the late 1990’s.

There are several factors that have facilitated the emergence of this new medium
(or an “advanced” medium), but the important technologies that make up the medium’s
foundation include computer hardware technologies, developments of telecommunication
network atmosphere such as the Internet, large bandwidth that allowed the high speeds
necessary for processing digital video, innovations in software technologies that eased the
creation of digital contents, and digitization of broadcasting transmission (Swann, 2000).
Various trials and errors from experiments on the medium during the last decade have
provided many lessons. Furthermore, the content and applications developed early have
showed that the concept is viable.

Many ITV descriptions use various new technological innovations to illustrate the
medium. But ITV can be simply explained as a convergence of the two sectors — the
Internet, which is a telecommunication domain, and the traditional media-oriented
television sector. This convergence offers interactive services to the audience via the
television or computer. Consequently, the audience can use e-mail, browse the Internet,
shop, seek information they want, and play games with their TV. Also, they can enjoy the

very same services from their computer, bringing television into the monitor screen.

There is no universally agreed upon definition of interactive television. It is even
being called different names, such as enhanced television, advanced television,
synchronized television, and so on. Nevertheless, people Share an agreement over the
fiinctions that the medium may and will provide. In fact, some of them are already being
used. For instance, the subscribers of Microsoft WebTV can play along with TV game
shows. Time Warner offered its subscribers in New York City interactive ads (Swedlow,
2000). Little by little, the audience is becoming accustomed to the idea of interactive
television or interactivity in television.

The industry is also showing signs that it considers the medium to be serious and
lucrative. For instance, more companies are now building divisions dedicated to
broadband/ITV strategies, and Multiple System Operators (MSOs) began launching ITV
network services to test new technologies, content, and the potential for television
commerce (t-commerce) revenues (Swedlow, 2000). Even though they are not yet
available on a massive scale, positive changes are occurring around the world to make

this new medium a reality soon.

1.1. Interactive Television (ITV)

What exactly is interactiVe television (ITV)? Is it the Internet via television? How
is it different from the interactive services that are currently provided by the Digital
Broadcast Satellite (DBS) and digital cable services? Is it different from HDTV? What
benefits does it offer? These are the questions lingering over average consumers’ heads.

Verifying this, a study conducted by In-Stat Group reported that the biggest obstacle

facing ITV in North America was confusion over what exactly ITV is, and more
importantly, what consumers want it to be (in Pastore, 2002).

The issue regarding the confusion about the concept and definition of this new
medium will be addressed in depth in the following chapter. Meanwhile, the functions of
ITV may be simply described as any two-way, interactive services that are being offered
to the audiences through the TV sets. In the ITV environment, the television works as a
host of applications such as entertainment services (pay-per-view, video-on-demand, and
games), information services (electronic program guides, local information, Internet
access, and distance learning), t-commerce services (shopping, electronic catalogs,
insurance, and banking), and other services (e. g., online voting). Consequently, an
audience’s conventional experience with television will be greatly transformed. For
example, people will be able to read more about the topic presented during a show, watch
a show on the viewer’s own schedule, and purchase goods associated with a program

using the aforementioned features.

1.2. Television Commerce (T—commerce)

One reason that ITV is an attractive alternative is t-commerce (television
commerce or TV-based e-commerce). Along with subscription fees and advertising sales,
t-commerce is another important revenue generator for ITV. In fact, the industry believes
t-commerce to be very viable, and many recent studies are introducing results that will
only strengthen this belief. For example, Yankee Group (summarized in ITV Marketer,
2002) reported that digital TV is expected to surpass PC-based Internet penetration in

Western Europe by 2005, leading to $17 billion in t-commerce revenues in Europe by

2006. OVUM (2000, in Pastore, 2000) predicted t-commerce revenue would be worth
$45 million by 2005. Also, a study by TechTrends, Inc. showed that 46 percent of
consumers are interested in t-commerce. It continues to report that one in three
consumers showed an interest in TV-based banking and investing, and that the most
likely users for t-commerce include premium cable and DBS subscribers, active online
shoppers and frequent customers of home shopping channels (TechTrends, 2000). Also,
the fact that more than 80 percent of the active audiences of home shopping channels are
interested in t-commerce suggests positive ideas to QVC and HSN regarding their future

in t-commerce atmosphere (Pastore, 2000).

1.3. The Purpose of the Study

On the surface, the Viability of interactive television (ITV) seems to be obvious.
However, the problems that the industry faces in developing and advancing ITV include:
(1) lack of technology standards; (2) lack of research on the kinds of products and
services for t-commerce that work best in an ITV environment; (3) advertisers’ lack of
knowledge on ITV; and (4) difficulty in testing advertising due to a low number of
installed set-top boxes. The current study aims to provide answers to some of these issues.
Particularly, it examines how added interactivity to television will change (1) the way the
ads are presented to the audiences and (2) the way the audiences respond to the ads.

To examine these issues, the current study proposes a new ITV ad format —
Interactive Product Placement (“iPPL”). iPPL refers to a type of advertising that are
embedded in the programming. The biggest difference of iPPL from traditional product

placements (PPLs) is that the iPPL would be designed to generate immediate actions of

the audience (e.g., requesting more product information and purchasing). Traditional PPL
focused on being recognized and perceived by the viewers, and thus many related studies
examined the role of exposure on consumers’ memory. This dissertation examines the
role of products exposure on memory, attitude toward the advertised brand, and
consumers’ interaction with the ads. Other placement-related factors such as the
consumer’s involvement level with the product class and the program, and their attitude
toward the program or the elements in the program will be examined as well. Interactive
commercials are just like traditional television commercials except that they will be able
to provide fimher information upon request (i.e., click) or provide different versions
according to the audience characteristics.

Of course, there will be other types of ads in ITV, such as interactive commercials
and interactive infomercials. The ads (and programs) from home shopping channels will
also benefit from the added interactivity. However, the audiences are already
experiencing a certain degree of interactivity without the ITV features in today’s
commercials and home shopping channels (e. g., calling the number on the screen to
purchase the product immediately). Therefore, this study concentrates on the interactive
product placement (iPPL). Hence, the following chapter discusses ITV in more depth in
terms of its concept, history, and its connection to the traditional television media
environment. It will be followed by the discussions on ITV ads, including the analogy to

the traditional product placement practices.

Chapter 2. Interactive TV: Description, History, & Advertising

2.1. What is ITV?

Some interactive services have already become common on many pay-TV
networks, through digital cable TV operators, and to most DBS subscribers. However,
these applications (e.g., Electronic Programming Guides and Video On Demand) that are
currently available to audiences are insufficient to be considered an ITV.

The potential of ITV functions has been described in many academic and
industrial reports, but there is little agreement on the extent of interactivity that will be
demonstrated to its audience. The individual viewers will be able to choose the ending of
a program; or, for a lower level Of interactivity, the viewers will be able to select different
camera angles of sports replays, or they will be able to see on-demand textual
commentaries of certain players or plays. Likewise, it is believed that viewers will be
able to click and see information about a certain product the character is using in a
particular Show. The lowest level of interactivity will allow viewers to watch a program
afier it has already been broadcast (i.e., delayed watch), and they will be able to pause
and replay the Show, just as they can do using current digital recording technology, such
as TiVO. Furthermore, audiences will be able to display program-related information on
their TViscreen, as they can now using electronic programming guides (EPG) as a part of
digital TV services. Undoubtedly, the possible advantages that interactive television can
provide, or the priority of its features that audiences would seek, may vary greatly.

What does it take to be an ITV? Although the current study places emphasis on

the advertising in an ITV environment, information will first be presented on the

medium’s contents, technology, and business model in an attempt to better understand the
nature of advertising in ITV. Then, the history of developments and the technological

specifications will follow.

2.1 .1. ITV: Contents

The types of content in an ITV environment are endless. Prominent applications
include a technologically enabled digital video-on-demand (VOD) and electronic
programming guides (EPG). VOD services use digital Video server technology, lets the
audience access the program database, and allows them to watch on their own schedule.
Swedlow (2000) noted that VOD would also provide an attractive billing model to the
program providers as its business model resembles that of the current pay-per-view
industry. Other on-demand services besides VOD include digital Video recording (DVR).
DVR is also called “personal video recording (PVR),” and it allows audiences to pause,
rewind, and even digitally save programs in the storage device such as a hard disk to
watch them whenever it is convenient for the Viewers. Although VOD is not fully
available at the moment, DVR services are already available in the US. from providers
such as TiVo and ReplayTV.

The EPG is also available on digital cable and DBS systems. The EPG that
appears on the television screen allows the audiences to navigate, search, and sort the
programs by channel, time, type, and so on. Companies like Gemstar, TV Guide, GIST,
ReplayTV, and TiVo are currently providing EPG service. Currently, the information in
the EPG is presented only in a format set up by the service provider. However, it is
expected that the EPG would work as a “TV portal” once the viewers become able to

build their own program guide.

Viewers will be also able to interact with the programs (e.g., quiz shows) or other
audiences (e.g., via instant messaging, chat rooms, e-mails). They will be able to obtain
further program-related information (including shopping options) directly on the
television screen. Channels Specializing on various commercial services can be provided
(e.g., shopping, e-mail, games, advertising, etc.), and public services such as distance
learning and online voting will be also provided directly on the television screen (Jacobs
& Dransfield, 1998).

In terms of the Viewing experience, the menus and various options are designed to
be presented in graphical boxes as we see in sports broadcasts today, or in separate fields
such as those from CNN news where the main screen is reduced in one comer and
various pieces of information (e.g., weather, stock, etc.) are displayed in other places.

cc‘ss
I

WebTV now puts an icon on the screen to provide more information on an advertiser
or a content provider. Swedlow (2000) predicts that such an interface will be used for the

programs touting direct purchase of the related products.

2.1.2. ITV: Technologies and devices:

Remote control is expected to be the primary input device. Other devices such as
wireless keyboards or wireless mice are considered as inconvenient and thus expected to
be secondary devices. To improve the convenience, a rudimentary version of the voice-
activated remote control has been already introduced to the market (e. g., Hammacher
Schlemmer, InVoca). Touch screens on televisions might also be used. Using personal
data assistant (PDA)-types of devices to integrate remote control functions (and

telephony) is also being tested.

The chief device in an ITV environment will be the set-top box, which has been
associated with the cable industry for a long time. The new boxes will carry
microprocessors, memory, conditional access technology (i.e., storage device), and a
connection to the network (Swedlow, 2000). Some ITV operators may choose to use
server technology at the cable headend and let viewers download applications and
contents from the server in order to prevent the set-top boxes from becoming obsolete

because of technological developments in hardware.

2.1.3. ITV: Revenues:

Subscription fees are an important revenue source. For instance, a report from In-
Stat/MDR (in Barlett, 2002) predicted that the number of TV households using on-
demand services (e.g., VOD) worldwide would increase from 1.3 million in 2001 to over
33 million in 2005, and that the revenues in North America would increase from $86
million to over $1.75 billion in 2005, which all suggest an optimistic future for ITV. On
the contrary, some studies present less optimistic views on the subscription-based
business models. For example, a study by TechTrends (2000) showed that only 1 in 17
consumers is willing to pay more than $3 for ITV services. It means that additional
revenues will have to come from advertising. But for ITV, DVR may pose a threat to
advertisers, since the device allows the viewers to easily skip commercials.

T-commerce, on the other hand, will be another critical revenue generator. An
example can be found from the service in the UK. BSkyB offers the service, called
“Open,” to its DBS subscribers (Swann, 2000). During the 1999 holiday season, it
provided a special channel for interactive home banking, grocery shopping, e-mail,

games, and so on. As a result, Open reported more than $1 million in t-commerce sales.

Although the sales have declined after the holiday season, there are two factors that could
change that trend. First, its service did not include enhanced television capabilities by
which viewers would have far more purchasing options. Second, the number of
subscribers continued to increase afier the season, which forecasts a positive future of the
service. Considering that more purchase options are made available, it is evident that t-

commerce will play a pivotal role in an ITV business model.

2.2. History of Development

2.2.]. 1970’s — VBI and QUBE

In the early 70's, the Vertical Blanking Interval (“VBI”) was used to send analog
signals, which eventually carried closed captions in the US. and teletext in the UK.
Teletext content included news headlines, sports scores, travel reports, movie reviews,
weather, and so on. Later, US. companies used these signal to send out programming
information for EPGs. In the mid-90’s, early developers of ITV platforms were already
exploring new types of broadcasting over the VBI because of the low cost and an already
established standard. Thus, broadcast data today still have to be transmitted over the VBI
to digital set-top boxes or other data receivers. The history of these two-way set-top
boxes started from QUBE, the first commercial ITV network (Swedlow, 2000).

Developed by Warner cable in 1977, QUBE allowed Viewers to vote, select
movies to watch, and play along with game shows. Although the services were rather
basic, QUBE demonstrated that viewers not only wanted interactive features, bit would

pay for them.

10

2.2.2. 1980’s — HDTV and Digital TV

In the early 1980’s, HDTV was first proposed (MacInnes, 1994). Despite its
superior image, many problems such as the degree of digital adoption, transmitting
method, broadcast spectrum, and compression scheme arose. The FCC tried to forge an
agreement regarding the controversy, yet failed to establish an international standard. In
the meantime, the existing analog spectrum is being taken back from broadcasters and
used for other data services such as mobile communications and datacasting. Such
services may be added to digital signal transmissions, especially for HDTV. Currently,
some HDTV programming is available in the U.S., but the high prices of digital TV sets
are preventing them from being widely distributed (Swedlow, 2000). This lack of
standard and the slow growth of usage are holding back manufacturers and program
developers from making the hardware and sofiware for HDTV. AS a result, digital
broadcasting is experiencing slow growth. On the other hand, digital broadcasting is
experiencing rapid growth in Europe, and subscribers there can access the Internet, enjoy

digital teletext, and engage in t-commerce (Greenspan, 2002; Jacobs & Dransfield, 1998).

2.2.3. 1990’s — Signs for emerging ITV

Until the mid 1990’s, many developments occurred to make ITV a reality.
Narrowcasting began and diverse channels have appeared over the analog networks
around U.S. Computer technologies equipped the television and film industries the
capability to digitally edit and produce work. The Internet made a widely distributed
interactive multi-medium a reality, by which every individual became able to produce a
Web site with rich content including audio and video. As a result, a new interactive media

industry was born.

11

There have been many trials and errors as well. Among those, Time Warner
launched an interactive service in 1994 and provided VOD, shopping, games, and EPG
services to 4,000 households in Florida (Swarm, 2000; Swedlow, 2000). Although the
high operational cost caused the project to be abandoned, some lessons were learned —
the service itself should have been free to the audience, tiered pricing models did not

work, and VOD was found to be a highly popular application.

2.2.4. Intercast and WebTV

In 1995, Intel introduced “Intercast,” which was a TV tuner card bundled with
software and the contents from NBC (Miller, 2001; Tedesco, 1996). Although it
represented an example of interactive data and television content integrated under a
single medium (i.e., PC), insufficient content made the product unsuccessful (Swedlow,
2000). In 1996, WebTV introduced a standalone set-top box with Internet service. One
year later when WebTV was purchased by Microsofi, its three set-top box models not
only featured two-way VBI broadcasting, but its content producers were able to use the
services of various ITV databroadcasters to provide enhanced interactive services

(McClellan, 1997; Swann, 2000).

2.2.5. ReplayTV and TiVO

ReplayTV and TiVo launched their own standalone products in 1999 (Hale,
1999b). These products used hard drives to offer users DVR capabilities. Also, EPGS
were already included in the service. Although the products have not been very successful,
the DBS industry recognized the potential of these types of products, and has begun to

offer similar services using its own set-top boxes.

12

2.2.6. Industry’s reaction - Enhanced TV and Synchronized TV

How will the industry react to all of the technological developments, and what
will they present as a standardized form of ITV service? Problems exist, especially in
establishing an industry standard. However, Enhanced TV and Synchronized TV emerged
as leading contenders concerning how ITV will progress.

First, Enhanced TV represents a specific type of ITV. In its interface, the elements
and data are transmitted Via the TV signal and then overlaid (not integrated) on the video
broadcast. The Enhanced TV services currently provided by Wink Communications
offers limited but still interactive choices to the audiences. The audiences can see a small
icon on the TV screen like WebTV, and it can be clicked to provide further information
(Hale, 1999a). It ranges from product information offered by the advertiser to the data
from the Weather Channel (Swedlow, 2000).

Second, Synchronized TV represents an Internet application, which may be
described as an integration of the Web and TV (e.g., WebTV). Basically, anything in the
Internet environment can be done in the Synchronized TV environment while the
program is on air (Swedlow, 2000). Synchronized TV receives HTML data broadcasts
that are synchronized with television programming. Applications can be downloaded
from Web sites and its audiences can interact with other viewers, call for further
information on the program, and use e-commerce capabilities provided by the advertisers.
A company called WorldGate recently launched a relevant service - Go!TV — that allows
“channel hyperlinking. (Swedlow, 2002)”

A third technology called Hypervideo enables producers to embed “hotspots”

inside a program (Sun Microsystems, 2001). The hotspots are clickable spots in the TV

13

screen to provide links to the Web, another program/channel, or to call up other elements
in sync with the streaming video. However, they are different from the abovementioned
“icons” in Enhanced TV, in that the icons generally stay in the same spot during the
program while the hotspots can travel on the screen in sync with the program. Many
companies and institutions including Digital Renaissance, Apple, Veon (now a part of
Philips), and the MIT Media Lab have developed authoring tools for hypervideo, and

they all provide embedded hotspots that will jump to another video segment, piece of text,
audio clip, picture, or Web page (Swedlow, 2000). Altogether, the technology reinvents
the concept of television as an interactive medium and suggests attractive applications to

advertising and e-commerce industry. Figure 1 illustrates the history of ITV development.

Figure 1. History of ITV Development

 

Time Warner (F SN):
VOD, EPG, shopping,
and games in Florida

 
 
 
 

 

 

5%

‘<

'5'

H

O

.3. 5

‘ O
J
t

 

 

l4

2.3. Forecast

Technological possibilities of ITV have been verified thus far. Marketing research
has suggested that people would like to have ITV around them. Considering the
definition of ITV, it might even be said that ITV is already here to some degree. On the
other hand, the social adoption of the medium has been a different topic, and there are
questions about whether ITV would be easily adopted by the general public (Jacobs &
Dransfield, 1998; Lee & Lee, 1995).

Jacobs and Dransfield (1998) assumed that ITV would be an integration of TV
and the Internet, thus allowing more interactivity to the audiences. Although they
concluded that ITV would be gradually integrated with the current television industry to
reduce the consumer resistance and indifference, they noted that watching television is a
group (e.g., family) activity, and thus consuming television and the Internet involve
different patterns. For instance, unlike television, the Internet is considered as personal
medium, and because of this difference, they argued that the public would not be
interested in accessing the Internet using their TV sets. ITV industry also recognizes
surfing the Web on television as a failed concept (Greenspan, 2002).

Although not focused on browsing the Internet on television, Lee and Lee (1995)
also examined people’s pattern of television watching, and suggested that, in order for
ITV to be successful, people’s current patterns, motivation, and gratifications of
television watching must be taken into account. They predicted that the audience would
not always prefer interactivity in television because they usually would like to sit back
and relax when they watch TV. Specifically, Lee and Lee (1995) noticed that people

enjoy low-involvement as well as high-involvement viewing, and predicted that ITV will

15

hinder this pattern of TV watching to slow its adoption rate. Lee and Lee (1995) assumed
that ITV would always require interactions and extra attention from the audiences.
Considering that ITV provides only the capacity for more interaction and involvement, it
may be anticipated that whether audiences pay more attention or not would depend upon
the individual’s decision. That is, if they don’t want extra interaction, they can watch it
like a conventional television.

Second, Lee and Lee (1995) used the term “routinization” to describe people’s
loyalty to certain programs. The study Showed that over 50 percent of those who chose a
specific program responded that they almost always watched the program and knew that
the program would be on at a particular time. The researchers also indicated that this
pattern would be an obstacle to viewers exploring new types of programs that require
interactivity and also to the programs seeking new audience groups. However, this type
of challenge is not exclusive to ITV, but to all types of new programs. On the other hand,
it may be expected that heightened involvement due to interactivity will produce less
“grazing” activities, which generally describe the audience’s constant channel changing
using a remote control (Eastman & Newton, 1995). Eastman and Newton (1995) Showed
that most grazing occurred between programs rather than during programs.

Third, relaxation has been considered an important benefit that the audience may
get from traditional television viewing that does not demand interaction (Lee & Lee,
1995). It has been implied that it is for this reason that ITV will not be greatly appreciated.
However, it must be noted that such benefit may not apply to all program types. For
instance, the audience’s internal states (e.g., attitude, emotional state) toward watching

television may vary across different program types. It is expected that the added

16

interactivity to the program would provide different features to the audience of different
programs (e.g.., team statistics in sports programs, shopping options to dramas, etc.).

Finally, conventional television is also known as a “storytelling medium,” and
Lee and Lee (1995) doubt that ITV may have much to offer to this nature. This charge is
related to ITV’s interactivity that might distract Viewers who are involved in an
engrossing story and thus interferes with viewers’ need for low engagement uses of
television as well. Again, it might be expected that this unwelcomed distraction would
occur in only certain types of programs. Despite all these weakness, Lee and Lee (1995)
predicted that commercials and infomercials would make strong use of interactive
capabilities.

Not all charges against ITV appear to be justified. It is certain that technology
now advanced sufficiently for ITV to compete in the market. Determining whether
consumers will prefer ITV services to current DBS, digital cable, or other similar services
require further deliberation. Although the viability of ITV is an important topic that needs
further discussion, it is not the focus of this study, and it will not be discussed further. The
primary focus lies in examining the ITV ad effectiveness based on what has been known
about the medium ’8 characteristics. It should be noted that because advertising will be a
very critical revenue generator for the medium, it is important to understand how the ads

will attract the audiences.

2.4. Ads in ITV
Despite the facts that there is no pre-established universal standard for ITV

technology and that there are no ITV ad formats that fully demonstrate the current ITV

17

technologies, the above discussions indicate that the ad format will allow real-time
interactivity for the audience. However, there has been no academic research on the types
of ITV ads or their effectiveness.

Conventional TV commercials are expected to benefit from added interactivity.
Stroud (2002) showed that the implementation of the interactive commercial in the UK.
was successful in terms of the viewer’s evaluation or their response rates. However, one
threat to the TV commercials comes from DVR, which allows the audience to skip over
commercials in an instantly recorded program. Some people underestimate this threat
comparing it with the threat of VCRs in 19805 that turned out to be minor. But unlike the
VCR that has been mostly used to play rented videos, the primary purpose of DVR is
instant recording and replaying, by which the commercials may easily skipped and
“zipped (i.e., fast-forwarding through pre-recorded commercials)” Despite this
forewaming, Swann (2000) predicts that the ITV will actually be favored by the
advertisers because of the interactive shopping. Some examples of interactive shopping
and interactive commercials can be found from the industry’s past experiments with
commercials. For example, WebTV introduced the “Click-To-Video” ads, which are the
banner ads that could be clicked by the WebTV subscribers. This banner ad let the
audiences watch the TV commercial for the particular brand, and then eventually takes
the audience to the advertiser’s Web site. Although this example features a banner-
commercial-Web site direction due to the limitation of the WebTV’s unique characteristic,
it shows what can be provided to the audience by interacting with a commercial. In an
experiment in August 1999, RespondTV and Domino’s Pizza tested their interactive

commercial and t-commerce application while sponsoring a Star Trek marathon in San

l8

Francisco area. During the Domino’s commercials, a small icon was displayed on the
screen with which the audiences could order a pizza by clicking it. 14% of the total
subscribers actually ordered a pizza, and 96% of those who ordered responded that they
would be inclined to order a pizza through the television again (Swarm, 2000).
Commercials in the ITV atmosphere would provide audiences with further
product information, an option for instant purchase, and a link to the advertiser’s Web site
upon request. Regardless of the advertising formats, the overall direction of ITV
advertising is headed toward strengthening t-commerce opportunities. Considering that a
popular application of t-commerce in an ITV environment will be the audience
purchasing goods through a TV screen during a Show, it may be expected that product
information embedded in a program would work as an ad format in an ITV environment.
This is critical particularly because that DVR technology offers Viewers a way to skip
traditional TV commercials, which might decrease both the reach and frequency —
assumptions on which conventional advertising models are based. This new situation
calls for new types of advertising in ITV. This study presents interactive product
placements as a potential new advertising format for the new medium. Also, the

characteristics of this format will be discussed.

2.5. Product Placements

TV commercials can be categorized based on the location in comparison with the
programs (i.e., within a program, between programs). Widely used formats of advertising
and other promotional practices in television include 15-, 30-, or 60-second commercials,

infomercials, PPL, and sponsorships. Particularly, sponsorships (including end credits)

19

have not been considered as an advertising activity in a strict sense. But many
sponsorships now appear in the form of a short commercials these days. The current
study examines product placements in ITV.

Product placements (PPL) have been a popular advertising practice in movies. In
the recently released movie Minority Report, which depicts society in 2054, a number of
today’s popular brands appear (e.g., BVLGARI, Lexus, Nokia, Gap). Characters in the
Men in Black series have been wearing Ray Ban sunglasses, and the recent 007 movie
series featured BMW as the “Bond Car.” It must be noted at this point that the term
product placement is somewhat misleading as the practice always refers to a placement of
specific brands. In this regard, many studies use the term “brand placement,” but this
dissertation will use product placement as it is more commonly used in the industry.

Steortz (1987, in Karrh, 1998) defined product placement as “the inclusion of a
brand name product package, signage, or other trademark merchandise within a motion
picture, television, or music video (p.22).” Adding the paid nature of the practice to the
description, Balasubramanian (1994) defined product placement as “a paid product
message aimed at influencing movie (or television) audiences via the planned and
unobtrusive entry of a branded product into a movie (or television program, p.31).” Karrh
(1998) extended its boundary beyond movies and television programs and defined it as
“the paid inclusion of branded products or brand identifiers, through audio and/or Visual
means, within mass media programming (p.33).”

Product placement represents an advertiser’s attempt to overcome the difficult
environments surrounding the television industry including increasing costs of

commercials, cluttering messages due to competition, and audiences switching channels.

20

However, this practice is regulated by the Federal Communications Commission (FCC)’S
sponsorship identification rules that apply only to the television programs made for

television. Specifically, the rule states:

When a standard broadcast station transmits any matter for which money, services, or other
valuable consideration is either directly or indirectly paid or promised to, or charged or received
by, such station, the station shall broadcast an announcement that such matter is sponsored, paid
for, or furnished, either in whole or in part, and by whom or on whose behalf such consideration
was supplied: Provided, however, That “service or other valuable consideration” shall not include
any service or property fumished without charge or at a nominal charge for use on, or for an
identification in a broadcast of any person, product, service, trademark, or brand name beyond an
identification which is reasonably related to the use of such service or property on the broadcast

(47 CPR. §73.1212, 1996).

In short, television programs should reveal the sponsor if there have been any paid
placements of brands for “more than a nominal” cost. Also, the rules describe that the
placement can be allowed unless the placement appeared to be unreasonable. However,
the sponsorship identification rules do not apply to the movies or movies broadcast on
television, and thus advertisers can use a product placement strategy without being
regulated.

Consequently, product placement has been flourishing in the movie industry. The
top five movies in 1990 featured more than 160 product placements (Colford & Magiera,
1991). Academic studies on product placement have been concentrated on the context of
movies. Many studies indicated that audiences were positive toward product placement

practices (Babin & Carder, 1996; DeLorme & Reid, 1999; Gould, Gupta, & Grabner-

21

Krauter, 2000; Gupta & Gould, 1997; Karrh, 1998; Nebenzahl & Secunda, 1993). The
common reasons have been reported as enhancing realism, helping character
development, and providing a sense of familiarity (DeLorme, Reid, & Zimmer, 1994;
Nelson, 2002). That is, audience members validate their usage of the brand, their own
identity, their interpretation of the character in the movie (or program) by comparing their
brand usage to that of the character in the movie (or program).

Marketers have also been favorable to the practice as it offers a captive audience
with a greater reach than traditional advertisements, a way to show the brand in its natural
environment, and a means of creating familiarity and even (indirect) celebrity
endorsements for their product (Buss, 1998; Turcotte, 1995). Turcotte (1995) also noted
low advertising clutter as another advantage of product placement. Furthermore, product
placements represent an attractive source of financial support. As a result, the popularity
of the practice, along with the accompanying price, has been rising. For example, Philip
Morris paid $350,000 to place its cigarette brand in the 1989 movie License to Kill
(Miller, 1990). Exxon paid $300,000 for a placement in Days of Thunder released in 1990
(Kanner, 1993). But Grover (2002) reported that Lexus paid $5 million in order to put its
futuristic (and not even commercially available) vehicle in the recent movie Minority
Report. Separate from that financial investment, Lexus also helped design the vehicle for
the movie. Although the practice is not ubiquitous in the television industry, it is expected
that its adoption might greatly benefit the producers, especially considering that
television sponsorship was reported to support 25 percent of the total production costs of

a program (Des Roberts, 1994, in d’Astous and Seguin, 1999).

22

As a result, usage of PPL has become increasingly popular. One sign of the
growing popularity of the practice is the increasing size of the (product placement)
agency groups in the US. and UK. (Curtis, 1996). It should be also noted that, despite
the F CC’s regulation of television industry product placements, placements made through
an agency may be able to bypass the regulation (Warner, 1995, in Karrh, 1998).

McDonald (1996) reported that the television networks have received calls from
the audiences asking where and how they could purchase the products placed in the
programs. In 1999 NBC started using its Web site to let the audiences of Passions
purchase jewelry and clothing appeared in the soap opera (Swann, 2000).
Balasubramanian (1994) noted that product placement — a “hybrid” commercial practice
- could stimulate more persuasion than traditional advertising through its combined
capability of advertising and publicity, and implied celebrity endorsement. Generally,
studies on the effects of product placement have reported some effects on audiences’
memory from placement. Law and Braun (2000) and Vollmers and Mizerski (1994) found
that the placement increased the consumer’s memory of the placed brand. Law and Braun
(2000) found that prominence of placement was also positively associated with high
recall and recognition. Karrh (1994) found that brands in previously unfamiliar categories
became more memorable by placement than brands that were already familiar. Some
studies, on the contrary, have reported no impact on the consumer’s memory (e.g., Babin
& Carder, 1986). Baker and Crawford (1995, in Karrh, 1998) also found that product
placements affected consumers’ short-term purchase intention. However, Vollmers and
Mizerski (1994) found no impacts of product placements on consumer attitudes toward

the brand or the actor. Karrh (1994) also found that consumers’ evaluation of the brand

23

was not affected by the placement. Nebenzahl and Secunda (1993) noted that implicit
placements in movies could generate negative ethical consumer concerns.

Some studies examined the effect of product placement based on the modality.
Generally, audiovisual placements were found to be most effective for higher recall and
recognition rates, but the superiority between verbal and visual placement is uncertain.
Paivio (1986) noted that the visual mediators were superior to the verbal mediators when
they are not Simultaneously available. Verbal information is harder to be retained than
visual information, which would imply that visual placement would be better recalled.
Also, it may be related to the fact that visual-only placements are most common, and
audiovisual placements are most expensive and hardest to achieve (Gupta & Lord, 1998).
But the empirical evidence is inconsistent. Law and Braun (2000) showed that Visual
placements generated higher recall than audio placement and that audio placement
produced higher recognition than Visual placement. Avery and Ferraro (2000) found that
brands that appeared verbally and those that appeared both verbally and visually at the
same time (i.e., audiovisual) were portrayed more positively compared to those that
appeared only visually. Similarly, Fischer (1996) found that a verbal mention of a brand
generated higher recognition rate than visual placement. It is also supported by an
industry practice. The Toronto Star (1991, in Karrh, 1998) reported that Walt Disney
Company charged $20,000 for a visual-only placement, $40,000 for a verbal-only
placement. This overall inconsistency might suggest that there is some other moderating
factor in the effectiveness of product placements or that the effectiveness depends on the

consumers’ individual differences.

24

There are implementational difficulties for the practice. First of all, advertisers
have little control over the content of the program/movie, and consequently it is not
certain‘whether their brand would appear in a positive or negative light. Second,
measuring its effectiveness has been a problem. Along with the aforementioned studies,
other academic research tested the effectiveness of product placement with memory-
based measures such as recall and recognition (e.g., Brennan & Dubas, 1999; Pracejus,
1995; Russell, 1998; Weaver & Oliver, 2000). Law and Braun (2000) employed an
implicit measure (i.e., effect of exposure on product choice indirectly) and differentiated
it from explicit measures (i.e., common recall and recognition measures) to reduce the
truth effect, mere exposure effect, and the false familiarity effect. Nevertheless, they were
memory-based tests.

In summary, product placements are becoming prevalent. However, there is little
knowledge about whether they are effective and how to best measure their impact. The
simple recall and recognition measures by themselves would not provide sufficient
information regarding persuasion since they cannot predict attitude formation and change
(e.g., Cacioppo & Petty, 1979; Greenwald, 1968; lnsko, Lind, & LaTour, 1976; Petty,
Cacioppo, & Schumann, 1983). Moreover, the reliability and validity of attitude measures
are better than those of memory-based measures (Clancy & Ostlund, 1976; Gibson, 1983).
Therefore, potential moderating factors influencing the effectiveness of product

placement need to be explored, using different measures.

2.6. Product Placements in ITV — iPPL

25

A recent survey conducted by Cahners In—Stat/MDR (2002, in Pastore, 2002,
January 23) implies that most consumers are aware that direct purchase while watching a
TV program (e.g., buying Jennifer Aniston's sweater while watching Friends) is going to
be available in ITV (in Barlett, 2002). As such, embedding ITV ads along with other I-
commerce information is expected to be implemented using aforementioned “hotspot”
technology in HyperVideo and SMIL (Synchronized Multimedia Integration Language).
The iPPL is expected to provide a new paradigm to the current product placement
practices in television. Traditional studies on PPL have been concerned with copytest-
related outcomes such as consumers’ memory or attitude. But the iPPL will add another
important response as they can generate consumers’ direct purchase. Considering this
added feature, Baker and Crawford’s (1995, in Karrh, 1998) finding on traditional
product placement provided a critical implication. They found that a mere brand
appearance might impact short-term purchase intention. Consequently, it is expected that
the added options of instant purchase might lead to actual sales when combined with
increased purchase intention.

Some findings from past research product placement would be applicable to the
iPPL. This dissertation examines the audience’s memory of advertised brands, attitudes
toward the brands, and the actual interactions. As the potential factors, this dissertation
examines the audience’s interaction with the iPPL, involvement with the programs and
the product categories, attitude toward the program and the characters in the program.
Main and interaction effects on audience response outcomes will be examined in the
context of a sitcom. The following chapters will discuss previous studies with similar

scopes and applicability to ITV. Hypotheses will be provided.

26

Chapter 3. Involvement

3.1. Involvement in General

A number of definitions of involvement emphasize the concept of self-relevance —
the degree to which a person perceives that concept (or the object) to be personally
relevant (Celci & Olson, 1988; Petty & Cacioppo, 1981; Richins & Bloch, 1986;
Zaichkowsky, 1985). The definitions of involvement are categorized into two groups.
The first group of studies defines involvement as an individual state (e.g., Andrews,
Durvasula, & Akhter, 1990; Johnson & Eagly, 1989; Gardner, Mitchell, & Russo, 1978;
Mitchell, 1981). The second group views the construct as a part of the process (e.g.,
Greenwald & Leavitt, 1984; Krugman, 1967). Representing the first group’s View,
involvement has been defined as “the motivational state induced by an association
between an activated attitude and some aspect of self-concept” (Johnson & Eagly, 1989,
p.293), “a situational state measured by the depth and quality of message-evoked
cognitive responses” (Batra & Ray, 1983, p.309), and “an individual, internal state of
arousal with intensity, direction, and persistence properties” (Andrews et al., 1990, p.28).
On the other hand, Krugman (1965, 1967) defined it as the dimensions of a process that
occurs during exposure to stimuli, and operationalized involvement as the amount of
connections the person has between his/her own life and the stimulus. Greenwald and
Leavitt’s (1984, p.591) definition states “the allocation of attentional capacity to a
message source, as needed to analyze the message at one of a series of increasingly

abstract representation levels.”

27

Celci and Olson (1988) highlighted the role of personal goals and values in
determining the degree of personal relevance. Leigh and Menon (1987) defined
involvement based on the level of attention and depth of processing. Unlike many studies
that advocated unidimensionality of involvement (e.g., Evrard & Aurier, 1996;
Zaichkowsky, 1985), Laurent and Kapferer (1985) and Kapferer and Laurent (1993)
advocated that it consisted of several factors resulting in involvement profiles. However,
later research supported the unidimensionality of the construct (Evrard & Aurier, 1996).

Involvement is “one of the most important variables in consumer research” (Antil,
1984, p.203). It has been a popular topic as it has provided rich implications for the
persuasion process among consumers (e.g., Krugman, 1965; Petty & Cacioppo, 1983;
1986; Celsi & Olson, 1988; Hoffman & Novak, 1996; Cho, 1999). Studies that View
persuasion in terms of attitude change use involvement as a motivational factor in
explaining the underlying cognitive processes (Chaiken, 1980; Petty & Cacioppo, 1979,
1981). Describing the relationship between motivation and involvement, Andrews et al.
(1990) noted that, although the properties of involvement (i.e., intensity, direction, and
persistence) are also properties of motivation, motivation represents a broader construct.
Motivation facilitates involvement-related consequences (Petty et al., 1983; Wright,
1973)

Explaining people’s attitude formation process, Petty and Cacioppo (1981, 1983)
established the well-known elaboration likelihood model (ELM). This persuasion model
focuses on the process of attitude formation or change, not the attitude per se, and it
illustrates that the process is influenced by the level of involvement. The ELM suggests

two distinct routes to attitude change — central and peripheral routes — when consumers

28

encounter persuasive communication. Petty and Cacioppo (1986) noted that a person’s
motivation, ability, and opportunity to process message arguments determine the route Of
elaboration. Change of attitude Via the central route occurs through a person’s attentive
and deliberate processing of message-relevant information. It is consistent with the
findings that comprehension, learning, and retention of relevant information are
important factors in high involvement state (Bettman, 1979; McGuire, 1976). Attitude
changes via the central route are also regarded as enduring and predictive of the person’s
behavior (Cialdini & Petty, 1981; Petty & Cacioppo, 1980). On the other hand, the
change of attitude via the peripheral route occurs by peripheral cues that are less relevant
to the information. In the peripheral route, consumers do not process the message-
relevant information intensively, and the changes in attitude are regarded as temporary
and less predictive of behavior. In short, the ELM views attitude formation or change to
be a result of the influence of central and/or peripheral cues. The level of involvement
determines the importance of those central and peripheral cues.

A person’s involvement state influences the person’s motivation to process the
information (that is relevant to the stimuli). Consistently, involvement studies have noted
that, in a high involvement state, individuals would pay more attention to the message
(Gardner et al., 1978), experience deeper levels of information processing (Leigh &
Menon, 1987), elaborate on the ad’s message (Petty & Cacioppo, 1986), produce self-
generated thoughts (Greenwald & Leavitt, 1984), and have more “connections” to the
message (Krugman, 1965). Supporting this View, Petty et al. (1983) found that the
strength of argument quality had a greater impact in high involvement condition.

However, studies have found that the level of involvement would not influence the

29

valence of attitudes. Andrews et a1. (1990) explained that whether the attitude is positive
or negative would be determined by individual reactions to the stimulus elements.
Likewise, Chattopadhyay and Nedungadi (1990) found no relationship between
involvement and attitude toward the ad.

Zaichkowsky (1985, p.342) defined involvement as “a person’s perceived
relevance of the object based on inherent needs, values, and interests.” The object may
refer to anything under the person’s consideration including a product class, an ad
message, programs, or purchase intention. Most involvement studies in advertising focus
on consumers’ involvement with the product category (e.g., Bloch, 1981; Bowen &
Chaffee, 1974; MacInnins & Jaworski, 1989), advertising message (e.g., Greenwald &
Leavitt, 1984; Laczniak & Muehling, 1993), and the program (e.g., Lord & Bumkrant,
1993; Tavassoli, Schultz, & F itzsimons, 1995). This dissertation examines the role of
consumers’ involvement with product class and programs on advertising effectiveness.

The following sections discuss these involvement types in detail.

3.2. Involvement with Product Categories

Bloch (1982) defined product (class) involvement as a state in the relationship
between consumer and product that reflects the amount of interest, arousal or emotional
attachment evoked by the product in the consumer. Celci and Olson (1988) noted that the
consumer’s involvement with a product (i.e., personal relevance of a product) is
represented by the strength of the linkage between the consumer’s individual needs, goals,
and values, and his/her product knowledge. Product involvement is considered to be a

state that may affect or moderate “the means by which brand attitudes are formed or

30

changed” (Laczniak & Muehling, 1993, p.302). Evrard and Aurier (1996) found that
product involvement was the most predictive factor of purchase behavior, and placed it in
the center of the “person-object relationship.”

Product involvement is considered to be an enduring involvement (e. g., Laczniak
& Muehling, 1993). That means consumers’ level of involvement with the product is the
state that affects consumer behavior on an ongoing basis (Bloch, 1981). On its
relationship with advertising, Maclnnis and Jaworski (1989) described product
involvement as being central in determining whether a consumer may be motivated
and/or able to process the advertising message. Laczniak and Muehling (1993) explained
advertising message involvement as a motivational state related to message processing.
This process-related involvement, unlike product involvement, is considered to be
situation-specific and transitory in nature. Product involvement influences advertising
involvement and the individual’s processing of the message, and this process influences
the formation or change of attitude toward the ad (Laczniak & Muehling, 1993).
Consequently, consumers highly involved with a product find advertising messages for
this product more personally relevant (Greenwald & Leavitt, 1984). They have a greater
motivation to attend to and a heightened level of involvement with the advertising. In
turn, high advertising involvement will generate higher attention to the message and more
cognitive elaboration of the message (Gardner, 1985).

However, the levels of involvement with a particular product may vary for
different consumers (Bloch, 1981; Laurent & Kapferer, 1985; Longfellow & Celuch,
1993; Zaichkowsky, 1985). The difference in the involvement levels could be found from

different consumers for the same product category or from the same consumer across

31

different product categories. In a high involvement condition, advertising message
effectiveness is enhanced by central cues such as argument quality, relevance to the
product, and relevance to the consumer’s beliefs. In a low involvement condition where
consumers are less motivated or less able to exert much processing effort, use of
peripheral cues such as celebrity endorsement, music, and advertising execution styles
are believed to be more effective.

Consequently, it is expected that the consumers will pay more attention to the ad
messages and exert more efforts for comprehension when the product falls in a high
involvement product category. De Pelsmacker, Geuens, and Anckaert (2002) reported
positive relationships between the level of involvement and recall, which is consistent
with the findings from other relevant studies (e.g., Cannon, 1982; Maclnnis & Park,
1991; Perry et al., 1997). Therefore, ITV audiences will attend to information about high
involvement products more than that about low involvement products. Likewise, it is
expected that they will be more likely to notice the iPPL of high involvement products
than those of low involvement products. Therefore, it is hypothesized:

H1. Consumers’ level of product involvement will be positively related to the

interaction with the iPPL.

H2. Consumers’ level of product involvement will be positively related to the

recall of the advertised brand in the iPPL.

3.3. Involvement with Programs
It was noted earlier that, for certain types of programs, interactive features of ITV

may actually disturb the audience’s television watching experience because they would

32

demand unwanted interactions. For some program types, audiences will not want to be
interrupted, and they would not welcome anything that would distract them (Lee & Lee,
1995). Therefore, it may be anticipated that, for certain types of programs, iPPL will
generate fewer interactions since such actions will require a shift of focus from the
program. This dissertation uses the audience’s level of involvement with the program to
explain the effects of iPPL that might possibly vary across different program types.

The principles of the involvement construct would be applied to program
involvement as well. Like product involvement, audiences’ involvement with the
program might also be considered as enduring. When considering the program as a
product, the audience’s level of involvement with the program may influence the
audience’s behavior regarding program (= product) consumption. But program
involvement may Show different patterns of interaction effects on various consumer
response outcomes, and may produce entirely different results in advertising effectiveness.

Studies on the impact of program involvement on advertising effectiveness have
produced opposite results. Some studies have found positive effects of program
involvement on recall and attitudes (e.g., Singh & Churchill, 1987; Srull, 1983). On the
other hand, other studies reported that high involvement generated negative effects on
recall of commercials (e.g., Pavelchak, Antil, & Munch, 1988; Soldow & Principe, 1981)
and attitude toward the ad (e.g., Soldow & Principe, 1981). Singh and Churchill (1987)
focused on the concept of arousal in explaining a positive effect of program involvement.
Similarly, Srull (1983) argued that arousal generated more vigilant and acute processing

of information which in turn leads to an increase in memory.

33

Studies in the opposing position advocate negative effects of involvement on
recall and attitude toward the ad. Their basic assumption is that television audiences use
their cognitive capacity (i.e., attention and comprehension) in order to process the
program and the capacity is limited. In addition, the limited capacity for cognitive
processing also reduces the chance to form counterarguments (Petty & Cacioppo, 1986).
This position is consistent with the aforementioned involvement principles that highly
involved audiences would go through more intensive information processes (to process
the program), which would result in lower elaboration on the commercials, and thus
lower recalls and unfavorable attitudes. Maclnnis, Moorman, and Jaworski (1991)
explained that even though audiences watching a highly involving program would have
greater resources for attention in the beginning of a commercial, their opportunity (and
ability) to process the ad would be less, as the resources would be focused on the
program. Other studies have also found that high involvement programs led to less ad
processing (Gunter, Fumham, & Beeson, 1997; Lord, Lee, & Sauer, 1994; Norris &
Colman, 1992). Tavassoli et al. (1995) suggested the inverted U-shaped relationship in an
attempt to reconcile the differences of two positions. They found that ad memory and
attitude reached the peak when the consumers had moderate level of involvement with
the program. However, they found that the peak of the inverted-U curve at lower level of
program involvement, and it may be interpreted as an instance of the negative
relationship for the most part of the involvement level.

This dissertation uses the level of product involvement to reconcile the differences
in the opposite positions. Lord and Bumkrant (1993) explained that program involvement

must be viewed in the context of the ad’s inherent attention-engaging capacity and the

34

audience’s involvement in the message. Involvement with the advertising message is
related to involvement with the advertised product. Product involvement might have been
less important in previous studies on program involvement and recall, particularly
because the program and the ads were presented separately on television. Therefore, the
issue of divided capacity for information processes would have been less obvious in the
context of separated presentation of the program and the commercials. Unlike ordinary
TV commercials, the iPPL competes directly with the program for the audience’s
information processing capacity. Because they would require audiences’ simultaneous
processing of advertising and program information, the problem stemming from
consumers’ limited cognitive resources might be maximized. Consequently, it is expected
that audiences highly involved with the program would have less capacity to process the
embedded iPPL, and that both interaction and brand recall would be negatively
influenced. Therefore, it is hypothesized:

H3. Consumers’ level of program involvement will be negatively related to the

overall amount of interactions with the iPPLS.
H4. Consumers’ level of program involvement will be negatively related to

recall of the advertised brand in the iPPL.

As described earlier, the level of involvement is not expected to influence the
attitude valence. The interaction effect of program involvement and product involvement
on recall and interaction will be investigated as well (Study 1). As previously noted, both
the product involvement and program involvement are considered as enduring

involvement. They are both relevant to the consumer’s (or audience’s) personal goals and

35

values. But they are anticipated to yield opposite results on consumers’ recall and
interactions. This dissertation proposes that the strength of the involvement will
determine the ultimate impacts on recall and interaction. Therefore, when the product
involvement is stronger and perceived as more relevant to the consumer, they will pay
attention to the product information in the iPPL even though the level of program
involvement is also high. On the contrary, they will ignore the information of the highly
involved product when they feel the program to be more relevant (or more important)

than the product.

3.4. Attitude toward the Program and Characters

3.4.1. Attitude toward the program

Media context has been considered as an important factor that may influence a
person’s attention to and elaboration of advertising stimuli (De Pelsmacker et al., 2002)
and advertising effectiveness (Derks & Arora, 1993; Perry et al., 1997). In reconciling
conflicting theories and hypotheses regarding the effect of the appreciation on advertising
effectiveness (e. g., cognitive capacity theory and affect transfer hypothesis), Norris and
Colman (1992) explained the differences using media characteristics, and De Pelsmacker
et a1. (2002) employed involvement as a moderating factor. Particularly, Norris and
Colman (1992) recognized that, unlike the ads in print media, TV commercials could not
be skipped easily. Therefore, the appreciation of the print medium context led to less ad
processing, which resulted in reduced advertising effectiveness. On the contrary,
television commercials have been thought to benefit from the carry-over effect of context

appreciation. De Pelsmacker et al. (2002) also reported similar results with television

36

commercials that appreciation was positively related with recall and likeability of
commercials. Considering the “skippable” and “inseparable” nature of the iPPL (i.e., the
iPPL has to compete directly with the program content for the audience’s attention), it is
suggested that the attitude toward the program would demonstrate similar effects as
program involvement. Therefore, it is expected that the iPPL will receive less attention
when the audience has a positive attitude toward the program content. However, the
attitude toward the brand featured in the iPPL is expected to enjoy the carry-over effect,
because the audiences would be exposed to the brand information only when they chose
to View the information (i.e., voluntary exposure), and thus they will not be bothered by
the advertising information. Therefore, it is hypothesized:
H5. The attitude toward the program will be negatively related to recall of the
advertised brand in the iPPL.
H6. The attitude toward the program will be negatively related to interaction
with the iPPL.
H7. The attitude toward the program will be positively related to attitude

toward the advertised brand in the iPPL.

Audiences’ attitude toward the program will also be examined in relation to
product involvement (Study 2) and program involvement (Study 3). Positive attitude
toward the program is expected to lead to similar effects as high program involvement on
brand recall and interaction with the ad because both represent the condition where
audiences’ cognitive resources are occupied. As a result, audiences’ interactions with the

iPPL and brand recall are expected to reach a peak when the attitude toward the program

37

is less positive and the level of product involvement is high. On the other hand, both
recall and interaction are expected to reach the lowest level when the program generates a
positive attitude and product involvement is low. However, consumers’ attitude toward

the brand (AB) featured in the iPPL may show a different pattern, because the attitude
toward the program (APROG) can exert positive effects on A3 while the product
involvement would not influence AB. In short, ApRog may be regarded as a peripheral cue,
and the impact oprRog will be greatest for low involvement products. This means that
low involvement products placed in positive APROG programs would show most favorable
A3, while this condition is expected to generate lowest level of brand recall and
interaction with the iPPL.

When APROG is combined with program involvement instead of product
involvement, the consumer’s response outcomes are expected to show still another
different pattern. In this situation, high involvement with the program is expected to yield
a synergy effect on A3 when combined with a positive Apnog. Likewise, negative Apgog

will have the worst effect on A3 when the audience is highly involved with the program.

3.4.2. Attitude toward the characters in the program

Traditional communication theories such as balance theory (Heider, 1946) and the
congruity hypothesis (Osgood & Tannebaum, 1955) may suggest other interesting
relationships. Heider (1946) explained changes in attitude by people’s tendency to
maintain a balanced state in the relationship between themselves, the communicator, and
the message. For example, when a consumer has a positive attitude toward a character in
a program (ACHAR) and when the character Shows positive attitude toward a brand, the

consumer’s attitude toward the brand (AB) is likely to be changed or reinforced in a

38

positive direction. Given the attachment audiences may have with the program and its
characters, and that the iPPL is likely to be implemented only for the products that are
positively associated with the character(s), it is suggested that consumers’ AB, interaction
with the iPPL, and brand recall may be favorably influenced.
H8. The attitude toward the character will be positively related to the
interaction with the paired iPPL.
H9. The attitude toward the character will be positively related to the recall of
the brand advertised in the iPPL.
H10. The attitude toward the character will be positively related to the AB in the

iPPL.

The attitude toward the character is expected to demonstrate stronger impact on
A3 than APROG because the relevance to the product would be greater for the character
than for the program. For the audience, ACHAR is centered on the actor or the actress while
APROG is centered on the program itself. When the audience has positive (or negative)
attitudes toward both the program and the character, the recall and interaction are
expected to be the highest (or the lowest). However, when positive (or negative) ACHAR is
combined with negative (or positive) ApRoo, the strength of the attitude is expected to
play an important role in the interaction effect on A3. That is, when the directions of
ACHAR and Armor; are opposite, the AB will be affected by the stronger of the two other
attitudes (Study 4).

The attitude toward the character will be explored in relation with product

involvement as well (Study 5). As with APROG, ACHAR is expected to work as a peripheral

39

cue in low involvement situations, and thus generate more positive (or negative) A3 for
low involvement products according to the direction ofACHAR. On the other hand,
although APROG is expected to inhibit audiences’ attention to the iPPL for both high- and
low involvement products, ACHAR is expected to encourage attention for both high and
low involvement products. This is because ACHAR will exert its influence by individual
pairings with particular characters, and thus the more positive ACHAR will produce higher

attention to the paired products.

3.5. Gender

Different TV watching behaviors across the genders also suggest an interesting
topic. Many studies on people’s remote control use have examined (1) who uses it more
frequently, and (2) who has the control over the device (e.g., Copeland & Schweitzer,
1993; Greene, 1988; Krugman, Cameron, & White-McKeamey, 1995; Perse & Ferguson,
1993). Particularly, research suggested that men engaged in “zapping (i.e., changing
channels during programs and/or at commercial breaks)” more frequent than women
(Comwell et a1. 1993; F risby, 2001; Greene, 1988; Heeter, 1988; Perse & Ferguson,
1993). It might be interpreted that since men dominate the remote control (Copeland &
Schweitzer, 1993), women would have little chance to use the device. But other studies
showed that compared to women, men change channels more frequently between shows,
during shows, and during commercial breaks (Heeter & Greenberg, 1985; Perse &
Ferguson, 1993).

In addition to many studies on zapping that discuss the audience’s ad-avoidance

behaviors, Heeter (1988) implied that male audiences had less concentration on the

40

program(s). Particularly, males were (1) more likely to watch more than one program at
the same time, (2) more likely to change channels just to see what else is on, (3) less
likely to watch a program from beginning to end, and (4) less likely to watch the same
program every week (Heeter, 1988). Similarly, Comwell et al. (1993) also found that
males were more likely to change channels immediately after turning on the television set.
On the other hand, it was reported that women are more willing to watch a program from
beginning to end (Heeter & Greenberg, 1988), and that women tend to know better what
they want to watch and what they will watch before turning on the television set (Heeter,
1988)

Therefore, it seems reasonable to expect that males would less actively engage in
the program content, but they would more actively interact with the iPPLs. Therefore, it
is hypothesized that:

H11. Male audiences will demonstrate greater amount of overall interactions

with the iPPLs than female consumers.

It is possible that certain program types or contents will be more favored by a

particular gender (e. g., actions and sports favored by males). However, this dissertation

will conduct a pretest to select a program that is not gender-biased in terms oprRoo.

41

Chapter 4. Interactivity and Interaction

Although the construct “interactivity” has not always been clearly labeled as such,
the idea existed in the form of feedback processes in traditional communication studies.
The construct had been studied for some time (Weitz, 1978; Wright, 1973), but it was
after the advent of many new media when interactivity became a widely popular topic for
researchers. Those new media considered to have facilitated interactivity included teletext,
video games, the Internet, and so on. In particular, the emergence of the Internet and the
World Wide Web (“Web”) and the development of relevant technologies in the late 19903
have brought a variety of interactivity studies as the Internet has been understood to
possess the capacity to feature full interactivity along with the multimedia content.

When we focus on the interactivity concept from the perspectives of the (media)
features, it can be argued that rapid deployment of new technologies has been increasing
the level of interactivity within a medium. Furthermore, it might be argued that an
element or a feature that was once regarded as very interactive may lose its
innovativeness although they might be still interactive by definition. For example, use of
multimedia and hypertextuality was considered interactive when the Web was first
introduced (Newhagen & Rafaeli, 1996). But such elements are now very common in
today’s online environment, and the audiences are familiar with even more “controllable”
multimedia objects on the web such as volume controls or interactive flash animations.
Therefore, adhering to certain interactive features in examining the effects of interactivity
might be risky considering this rapidly changing environment, and it is suggested to

focus on the heart of the (interactivity) construct that might be applied to different media

42

in different times. Chen (1984, p.284) stated that “looking beyond the technology of each
new medium to its underlying content. . .will enable theoretical progress that does not stop
at the borders of each new medium.”

Some studies have pointed out that the interactivity features might be perceived
differently to audiences, and examined the impact of the individual’s perceived
interactivity (e.g., McMillan, 2000b; Newhagen, 1998; Wu, 1999). However, many
studies still focus on the feature and try to examine its impacts on audience responses
(e.g., Massey & Levy, 1999; Rice, 1984; Rogers, 1986).

Despite the different focuses on interactivity, most studies assume audiences’
actual interactions as a given, and do not investigate the true nature of interaction. Based
on previous studies on interactivity, this dissertation will examine the nature and the
impacts of interaction. The following sections of this dissertation (1) revisit the various
definitions, antecedents, and effects of interactivity described in previous studies, (2)
present interaction as a distinct concept that might supplement interactivity in explaining
various audience responses to the stimulus medium, and (3) propose some effects of
interactions. Based on the previous definitions of interactivity and its dimensions of
synchronicity, vividness, and social presence, Fortin (1997) classified various
communication media along a continuum of their potential for interactivity. He described
conventional television to be the least interactive among other media such as print,
telephone, Internet, and conventional mail. However, ITV should be evaluated differently

because of its various interactive features.

43

4.1. Interactivity -— The Construct

Many studies from a variety of disciplines have defined and explained
interactivity from different angles (Aldersey-Williams, 1996; Ha & James, 1998; Heeter,
1989, 2000; Hoffman, Novak, & Chatterjee, 1995; Neuman, 1991; Newhagen, Cordes, &
Levy, 1995; Pavlik, 1996; Rafaeli, 1988; Rafaeli & Sudweeks, 1997; Rogers, 1986;
Steuer, 1992; William, Rice, & Roger, 1988). However, some suggest that the construct
still needs clearer conceptualization (Brody, 1990; Heeter, 1989; Morris & Ogan, 1996;
Pavlik, 1996; Rafaeli, 1988).

Interactivity is generally regarded as a characteristic of a communication system
(Williams et al., 1988) or of a communication process (Chen, 1984; Rafaeli, 1988) or a
combination of both (Heeter, 1986, 1989). Most definitions of interactivity are divided
into two groups — feature-centered definitions and perception-centered definitions. In the
beginning, researchers focused on the interactive features of a medium and presented
feature-oriented definitions (e.g., Heeter, 1989; Rice, 1984), but later studies started to
notice the effect of individuals’ perceptions of interactivity, and described the construct
from the individual’s angle (e. g., McMillan, 2000b; Newhagen, 1998; Wu, 1999). This
distinction between feature- and perception-oriented perspectives is important not only
because they may have different effects but also because we need to keep in mind the fact
that interactivity may vary within the same medium for different programs and within the
same program for different users. For example, the Web is regarded as a highly
interactive medium but some sites do not offer as much interactivity as others. Television
is regarded as less interactive, but some audiences participate in interaction with

programs that Offer such chances as call-in discussions, ARS (Audience Response

44

System) quiz shows, and so on. Similarly, Rafaeli (1990) noted that traditional mass
media audiences are becoming more active in participation using letters to the editor and
on-the-air talk shows.

In short, feature-oriented descriptions cannot explain the individual audience’s
different perception of the same medium. Following this approach, televisions are always
less interactive than e-mails. Perception-oriented interactivity definitions would help
describe the differences in individual audiences, but it would not be very helpful in
categorizing the different media based on interactivity, and thus it would be less useful to
examine the (interactivity) potential of the each medium. Despite these shortcomings,
both perspectives provide some valuable insights in understanding the interactivity
construct and in establishing the interaction construct.

F eature-oriented perspectives have defined interactivity as the capability of new
communication systems to talk back to the user (Rogers, 1986), and the extent to which
communication reflects back on itself, feeds on, and responds to the past (Newhagen &
Rafaeli, 1996), within the context of an ongoing communication exchange (Rafaeli,
1988). Therefore, interactivity refers to the extent that the later transmission of the
message is related to the earlier transmissions (or exchanges) of messages. In other words,
communication roles between sender and receiver must be interchangeable in order for
full interactivity to occur (e.g., Williams et al., 1988). Also, synchronicity of exchanges is
another characteristic of interactivity, but there is a general consensus that synchronicity
alone does not make a necessary nor a sufficient condition for interactivity to occur
(Fortin, 1997; Heeter, 2000). Heeter (1989) employed the concept of control from the

feature’s perspectives, and pointed out that users with interactivity would have more

45

control over the information to which they wish to be exposed. Similarly, Neuman (1991)
defined interactivity as a quality of electronically mediated communications
characterized by increased control over the commtmications process by both the sender
and receiver.

In particular, Heeter (1989) regarded interactivity as a multi-dimensional concept
based in the functions of the medium, and suggested that it is characteristic of the
medium’s processes or feature. Using Heeter’s (1989) definition, Massey and Levy
(1999) examined the level of interactivity in a Web site based on the presence of
interactive features (e.g., e-mail links, chat rooms). McMillan (1998) and Ha and James
(1998) also used Heeter’s (1989) definition, and identified additional interactive features
of a Web site such as search engines, hit counters, games, registration forms, surveys,
toll-free numbers, and so on. Newhagen and Rafaeli (1996) defined it as the extent to
which communication reflects back on itself, feeds on, and responds to the past. Alba et
al. (1997) defined it as a multi-dimensional construct, the key aspects of which include
reciprocity in the exchange of information, availability of information on demand,
response contingency, customization of content, and real-time feedback.

Despite slight differences in feature-oriented interactivity definitions, they
emphasize exchange and mutuality. These shared concepts is well expressed in Rice’s
(1984) description of the role of new media and interactivity. He noted that new media
“facilitate interactivity among users or between users and information” (Rice, 1984, p.35)

On the other hand, Williams, Stover, and Grant (1994) emphasized the importance
of understanding individuals’ uses of new media in the theory-building process.

Newhagen, Cordes, and Levy (1995) highlighted the psychological dimension of

46

interactivity, which centered around the “sense” that communication participants have of
their own and of the receivers’ interactivity.

Some studies have explained interactivity using the concept of audience’s control
(Ku, 1992; Rafaeli, 1988; Steuer, 1992; Spalter, 1996; Williams et al., 1988). Steuer
(1992) emphasized the individual ’3 experiential aspect of interactivity, and defined
interactivity as the extent to which users can participate in modifying the form and
content. Spalter (1996) described interactivity as enabling users to control and choose the
content. Newhagen (1998) argued that although the medium’s features may be important
to facilitate interactivity, the way that individuals use a medium would explain the
interactive process better. In short, the perception-oriented approach recognizes and
emphasizes the possible differences in the level of interactivity perceived by different
audiences for the same medium.

Therefore, Wu (1999) focused on the perceived interactivity of web sites, and
found that it was significantly related with people’s attitude toward the web site (AST).
McMillan (2000b) also documented that interactivity resided largely in the user’s
perception. She employed both feature- and perception-oriented interactivity and
examined whether the features would influence user perception of interactivity but found
a very weak relationship. But she found that both interactivity features and perceived
interactivity had a positive influence on users’ attitude toward a web site (McMillan
2000b), yet the perception was a stronger indicator than the features (McMillan 2000a,

2000b)

47

4.2. Interactivity — Other Effects and Antecedents

Studies that manipulated level of interactivity indicated a weak effect on learning
(e.g., Bailey, 1992; Frazer & McMillan, 1996; Kettanurak, 1996;1(u, 1992; Shaw,
Amason, & Belardo, 1993) or knowledge gain (Jaffe, 1996). However, Hoffman et al.
(1995) argued that interactivity would let users actively engage in the communication
process, and would help build the consumer-marketer relationship. Cho and Leckenby
(1999) used the feature-oriented interactivity concept, and showed that higher
interactivity was associated with favorable attitudes toward the banner ad and the
advertised brand, and with the intention to purchase the brand. McMillan (2000b)
explored the role of interactivity in explaining consumers’ behavioral responses (i.e., send
e-mail to the Web site, tell others about the site, etc.), and found that only perception-
oriented interactivity had a significant impact on the users telling other people about the
Web site (McMillan, 2000b).

Focusing on the control aspect of interactivity has yielded interesting results and
implications. As mentioned earlier, many studies examined interactivity in terms of more
user control (e.g., Ku, 1992; Heeter, 1989; Rafaeli, 1988; Steuer, 1992; Spalter, 1996).
Considering that the individual’s feeling of being in control was found to increase self-
efficacy beliefs (Bandura, 1977; Gist & Mitchell, 1992; Phillips & Gully, 1997; Tafarodi,
Milne, & Smith, 1999), more controllability enabled by interactivity in a media
environment would produce higher self-efficacy beliefs in a user. Increases in self-
effrcacy level have been reported to result in better performance over a task (Bandura,
1982; Bandura & Adams, 1977; DiClemente, 1981; Lee & Edwards, 2002;Nah1, 1996).

And consistent with the previous findings on the interactivity — AgT relationship (e.g.,

48

McMillan 2000b; Wu 1999), increases in self-efficacy level were also found to generate
more favorable A51 (Lee & Edwards, 2002). Therefore, it might be understood that higher
level of the user control from higher interactivity leads to increased self-efficacy and
favorable attitude toward the web site.

The approach that focuses on the control aspect of interactivity, as well as on the
aspect of a two-way communication, also suggests a relationship between interactivity
and involvement. That is, by practicing an active control in a two-way communication
process, users experience higher interactivity, and they may feel higher involvement with
the object of interaction (i.e., person, medium, content, message). Participating in a two-
way communication process means that the audience would be sending and receiving
messages instead of passively receiving them. These exchanges of messages and facing
chances to make decisions (or choices) would require greater attention of the participants,
which would heighten the level of their cognitive involvement occurring in the
communication process.

However, it Should be noted that simply having a chance to interact and actually
participating in the interactive communication process are different. Furthermore, it
might be questionable whether only “having a chance” would increase the involvement
level, especially when considering the possible differences between those who
participated in the process by actively interacting and those who ignored the options and
did not interact at all.

In short, it is reasonable to anticipate that audiences who participated should
experience higher involvement level than those who did not. But when considering the

common descriptions of interactivity features and perception, it is noticed that neither can

49

actually explain this difference. The feature-oriented definitions will describe the
situation (or the medium) as “interactive” because the users had chances (or options) to
interact. The perception-oriented approach might describe that everyone in the situation
perceived a similar level of interactivity, failing to distinguish those who interacted from
those who did not. The only difference can be found from the people’s actual actions, and

this issue will be discussed in the following sections in greater detail.

4.3. Interactivity and Interaction

Regardless of different definitions and conceptualizations of interactivity —
whether it is feature-oriented or perception-oriented, the interactivity construct centers on
the basic notion of human actions, reactions, or interactions. Therefore, examining
interactivity without taking an individual ’5 interaction into account would far less useful.
Individual differences in the perception of interactivity are important, which is why many
studies have examined the role of perceived interactivity as differentiated from
interactivity features. But the individual differences in the level of engagement in actual
interactions are also important. For instance, comparing audiences who interacted with
the stimuli in a medium and participated in the communication process with those who
did not could yield considerable differences in terms of their response outcomes such as
degree of attention, comprehension, and resulting level of involvement. The reasons for
the importance of interactions include: (1) a medium presents different interaction
conditions where the different amounts of interactivity might be selected by its audiences
(e.g., Laura and George both wanted call the radio station to participate in a quiz, but

only Laura could call because George had to go to the bathroom. ), (2) the amount of

50

perceived interactivity might vary for different audiences with the same medium (e. g.,
George only knows how to send and receive e-mails, while Laura is running an online
virtual community), (3) different audiences may have different levels of tendency to
interact with the medium (e. g., George would never buy anything from the Home
Shopping Channel, but Laura would buy anything that seems reasonably priced. ).

In a similar vein, Heeter (1989) also noted that different media systems require
different levels of user activity. She pointed out that although users are always active with
media to some extent, some users are more active than others and some media are more
interactive than others. For example, e-mails are regarded as both highly interactive and
non-interactive depending on a user. Conventional television and radio are regarded as
non-interactive mass media, but some audiences enjoy interactivity by participating in
live discussions. Of course, this does not deny that there are differences in the level of
interactivity across different media.

However, few studies have tried to distinguish interaction from interactivity, or
examined the role of interaction in consumer information processing. Before proceeding
with the interaction conceptualization, let us briefly review another classification of
interactivity — person interactivity and machine interactivity. Steuer (1992, p.84)
explained machine interactivity as “the extent to which users can participate in modifying
the form and content of a mediated environment.”He also emphasized the role of media
(in a model of mediated communication) as a facilitator of person-to-person interaction
by noting that media serve as a “conduit” in which message senders and receivers could
interact. Hoffman and Novak (1996) Viewed interactivity in terms of “feedback,” and

explained that a computer-mediated environment enables users to communicate through

51

the medium (i.e., person interactivity) and to provide or interactively access media
content (i.e., machine interactivity). Hoffman and Novak (1996) state that interactivity
could be through the medium (emphasizing the hmnan communication process mediated
by machine — person interactivity) or with the medium (and users interact with the
content — machine interactivity). Media features are central in machine interactivity since
they would directly enable the interactivity. The machine would play the role of a
communicator. On the other hand, the features would be less important in person
interactivity, where they would only facilitate human interactions. The machine performs
only as a mediator. Excluding the unmediated interpersonal communication (that is not
the focus of this study), it can be said that the machine (or the medium) always plays a
certain role — a communicator or a mediator.

It can be noted that the above descriptions on person and machine interactivity are
not just about the medium’s interactive features or user perceptions. Although it was not
clearly stated, the descriptions focus on the aspect of actual interactions occurring among
users and between users and media. At this point, it would be worth pointing out the
difference between interaction and interactivity: interactivity features and perception
characterize the machine (i.e., medium)’s elements and hurnan’s feeling, respectively. But
interaction refers to a behavior-oriented communication process whether it is between
people or between people and media. In order to examine the role of interaction and
discuss the degree of a medium’s interactivity based on the medium’s potential (not
features) to generate interaction, a clearer conceptualization of interaction would have to

be presented.

52

Heeter (2000) conceptualized interactivity while taking interaction into account as
well, and it provided a valuable starting point for the interaction conceptualization.
Primarily, she suggested that the concept of “interaction” would encompass a wide range
of internal responses of the audience to include thinking, feeling, attention, interpretation,
intention, and so on. In the beginning, she included every human action with an object in
the interaction boundary, and as a result, Web users’ Simple mouse movement, data
inquiry, along with their cognitive/affective responses were interpreted as interactions
(Heeter, 2000).

Then, Heeter (2000) limited the interactions to the actions physically observable
to separate the concept from such internal processes as perception, motivation, emotions,
and so on. She noted that those internal dimensions of interaction were “not subject to
direct observation,” drew a line between interaction and other (internal) responses, and
defined interaction as “an episode or series of episodes of physical actions and reactions
of an embodied human with the world, including the environment and objects and beings
in the world (Heeter, 2000).”

However, this dissertation proposes to further refine Heeter’s (2000) interaction
definition. According to her definition, television audiences’ flipping channels can be
understood as an interaction. Also, a magazine reader’s particular reading habit can be
interpreted as an interaction since it is observable. But these types of interactions have a
limited capacity to explain the medium ’5 interactivity (or interaction-generating potential),
although they are related to personal characteristics and tendency (to interact). Therefore,
it would be helpful to find a way to systematically differentiate these types of interactions

from other types as calling or writing back to the message sender. Simply speaking,

53

channel flipping actions and particular reading habits can be said to reflect how an
audience consumes, processes, and reacts to the stimuli provided by the medium. These
activities might not be sufficient to be labeled as an interaction — rather, they are closer to
“reactions.” Thus, the attempt of the current study to refine Heeter’s (2000) interaction
concept starts from distinguishing interaction from reaction.

The Merriam-Webster dictionary describes interaction as a “mutual or reciprocal
action or influence” or “to act upon one another: ” Reaction is defined as “the act or
process or an instance of reacting (which is “to respond to a stimulus”); a response to
some treatment, situation, or stimulus, and; bodily response to or activity aroused by a
stimulus.” Interestingly, the heart of Heeter’s (2000) interaction conceptualization — the
observable nature — is found under the description of reaction. And it should be noted that
the interaction definition describes mutuality. A similar clue for the differentiation (of
interaction from reaction) can be sought from many interactivity definitions, which
emphasize the aspect of Mo-way communication. Rafaeli’s (1988) definition of
interactivity is based on the “responsiveness” of a counterpart in the communication
process. He noted that for a communication to be firlly interactive, the sender-receiver
roles must be interchangeable.

From this, the current study proposes a refined conceptualization of interaction
using the concept of interchangeability, and it is stated as “observable physical actions an
audience performs in response to messages (content) provided through a medium which
alter the content being provided and/or which communicate with the sender (publisher),

either synchronously or asynchronously.”

54

The mutuality in the context of communication exchange was established with the
new conceptualization, but there is still one more issue that calls for a further
investigation. As mentioned earlier, certain interactions (e.g., channel flipping) are
different from other interactions (e.g., writing back to a magazine), and the new definition
by itself falls short in fully explaining the difference as it includes both the publisher and
the medium for the communication counterpart to which the feedback can be sent. The
answer may be found from the aforementioned rationales of person interactivity and
machine interactivity. Talking back to a publisher or sending information in a Web site
may be understood as a function of person interactivity as the audience’s interaction
would reach the original message sender. The communication counterpart for this kind of
interactions would be a person or an organization. This type of interaction embodies
higher interchangeability, and can be labeled as the “human interaction (with person or
organization)” On the contrary, the interactions like channel flipping, reading habits,
recording a program, or increasing the volume represent the interactions that hardly ever
reach the sender, and they can be understood as a function of machine interactivity.
Usually, it involves no human communication counterpart, and the world is oblivious to
this interaction. This type of interaction illustrates interaction with the medium or content
and can be labeled as the “content interaction.” Both types of interactions share the core
of the interaction definition, roles interchangeability, that is provided in this study. The
only difference between these types of interactions is in the communication counterpart —

(medium-mediated) person versus the medium itself.

55

4.4. Value of Interaction

Stewart and Ward (1994) recommended that advertising studies should change the
focus from analyzing media stimuli (and their impact) to exploring the way audiences
interact with the media. The new definition presented in this study will provide a means
to more closely associate the concept of interaction with that of interactivity, and will
allow us to use interaction concept as a means to examine media interactivity and the
advertising effectiveness. For example, an advertisement’s simple exposure to the
consumers has been believed as one of the key objectives for advertisers. It has also been
echoed by current industry practices that employ popup ads and by current online
advertising pricing policies that are based on reaches and frequencies. However, it should
be noted that more fundamental goals of advertisers are to take the audiences to the
advertiser’s web site, or to generate sales from the ad efforts. In other words, it can be
argued that the more important goal is to generate consumers’ interactions with the ads.

Interaction represents the audience is goal of the as well. That is, an individual
audience member’s (series of) interactions in media use may be interpreted as (re)actions
to achieve his or her goal of the media usage. The interactivity of a medium must be
designed in a way that can help audiences achieve their goal, and the content providers
and designers of a medium should first consider the reasons for audiences’ medium usage.
As Heeter (2000) describes, “designers try to make obvious to the human what actions
are possible at any time, and what affordances are available within an application.” Of
course, the designers should also try to produce favorable responses from the audiences
even though the responses might be unobservable(e.g., positive A51). It should be noted

that interactions might be the results ofthose unobservable reactions. For instance,

56

continuing interactions might be a results of the positive A57, whereas exiting the web
site might be a Sign of the negative AST.

Audiences View affordances based on their own goals, and every audience
member has a goal. Although a person may appear to be browsing a Web site without any
specific purpose, the person’s behaviors and the goal can be understood as killing time
(which may be achieved by entertaining contents). A person’s goals may be classified into
cognitive-driven / affect-driven goals, or information-oriented / entertainment-oriented
goals. However, no interaction would be aimless. Cooper (1999) advocated this View and
emphasized the design focusing on individual goals.

Earlier in this chapter, interaction was defined as having physical observability
and interchangeability of the sender-receiver roles. For a concept to be used in
comparison with other constructs, it should be measurable. The unit of analysis for
interaction may have different forms for different media. For television and radio
audiences, writing a letter to the station can be an example of interaction. For Web site
Visitors, clicking toward or away from certain web elements could be regarded as an
interaction. Despite the differences, the behavioral patterns and accompanied goals would
be similar across different media. One possible way to categorize these patterns would be
position them in a avoidance—acceptance dimension. Interactions of complete avoidance
would include closing a web browser window, clicking away from a web site, ignoring an
e-mail, changing a channel, and turning off the equipment. The complete acceptance
would include clicking into the web pages, saving the content for later, bookmarking, and
increasing volume of the television set. Similarly, interactions could be classified based

on the time of media consumption — live consrunption of content (e. g., click/volume

57

increase), delayed consumption (e.g., save/record), and avoided consumption (e.g.,
closing a browser window).

Because the web is computer-based and generally considered to be more
interactive than other mass media, the interactions in the online environment have a
unique characteristic - ease of measurement. The interactions on the Internet can be
represented by the clicks. Chatterjee, Hoffman, and Novak (1998) used Visit duration and
the number of pages visited as possible measure of consumer interaction with Web sites
and banner ads. But, the Visit duration in Web sites may be problematic when used alone.
Audiences’ time spent in viewing Web sites encompasses the number of pages viewed.
Besides, this measure can easily suffer from confounding variables such as the speed of
connection, individual differences in comprehension rate, and the particular situation in
which the person is browsing the web sites (e.g., concentrating on the content vs. doing
something else at the same time). Although Visit duration might be suitable for some
experimental studies conducted in a computer lab, it would not be an appropriate measure
of interaction for most of the cases. Instead, the number of web pages Visited by an
audience member, the number of clicks made to the hyperlinks (including ads), or the
individual click made on a certain hyperlink may be recommended as safer measures of
online interactions. The next section will discuss (1) what makes audiences interact and

(2) what is caused by the interaction in the context of ITV.

4.5. Interaction — Antecedents

Examining the effect of interactivity perceptions in an interactivity process,

McMillan (2000a) found that consumers’ positive attitudes toward the Web site would

58

better predict their subsequent actions than their interactivity perceptions would.
Considering that McMillan (2000a) and Wu (1999) found that that consumers’ perceived
interactivity affected their AST, it means that the impact of perceived interactivity on the
actions is mediated by the Ag. McMillan (2000b) also found that the direct influence of
interactivity perception on consumers’ future actions was only partial and mostly limited.
However, the conceptual difference between consumers’ actions used in McMillan
(2000a, 2000b) and the interaction has to be noted. McMillan’s (2000a, 2000b) actions
referred to those that were favorable reaction to the Web site (e.g., telling about the Web
site and purchasing from the site), whereas the interaction in the current study is rather
neutral in its nature. It was discussed earlier that interactions may represent audiences’
avoiding or accepting tendency with the stimulus. The audience’s actions could be
predicted by attitude because both variables were measured based on their favorability —
good vs. bad. Therefore, a direct application of McMillan’s (2000a, 2000b) rationale
might be problematic as it did not include the negative (inter)actions. Also, consumers do
not have to be favorable to the Web site in order to interact. Although they might not like
the Web site (e.g., online shop with a bad interface), they will still interact (e.g., purchase
a product or browse for further information) when they find a reason to interact (e.g.,
cheap price). In short, the interaction will not occur only because someone likes the Web
site. Rather, it will occur when someone sees a certain benefit in making the interaction.
Other possible reasons that would make it difficult to use the attitude as an interaction
precursor is the fact that interactions occur on very specific elements within Web sites
(e.g., chat rooms, ads, contents in need, etc.). Each element can affect the overall level of

the audience’s A51 based on the audience’s purpose of the web browsing, but the overall

59

A51 will not be able to clearly explain whether the audience would interact with a specific
element. Although the role of attitude might be unimportant in explaining interaction,
examining the effect of interactivity perception and features on interaction may suggest
closer relationship because the interactivity construct originates from the basic principle
of interaction. Specifically, it is anticipated that the audience’s interactivity perception
and interactivity features in media would increase the chance that the audience would
interact. However, it is unlikely that the features and perception would cause the
interaction behaviors.

If there is a well, people will come and drink. But it is difficult to say that the well
itself is the reason for people’s drinking. Its presence will increase the chance of drinking
from that site, but few will drink water only because there is a well. In other words, the
well does not represent the cause of the action. It is thirst that drives the action.
Interactivity features and perception only function as a well. They may increase the
chance of interaction, but they are not the cause. Why do people interact? It is to fulfill
their needs. Bumkrant and Sawyer (1983) recognized that the level of involvement is
determined by people’s need for information and the meaningfulness of the message
content. Therefore, this study recommends that it would be safer to assume that the
interactivity features and perception facilitate interaction and increase consumers’ chance
to interact, rather than representing the cause of interaction. Other factors that would
possibly increase the chance to interact are consumers’ cognitive intensity in processing
information. For example, it is reasonable to expect that the more attention the consumer

pays to the stimulus or the medium, the more likely the consumer is to show interactions.

60

As mentioned earlier, Novak, Hoffman, and Yung (2000) used the concept of flow
and described that consumers’ experience of flow under Web environment would make
the consumers involved in their navigation activity. They further explained that it would
let them more focus on the interaction, which they conceptualized as the “exploratory
behavior.” They found a significant influence of consumer experience of flow on their
exploratory behavior in their initial model establishment (Novak et al., 2000). However,
the measures and operationalization of interaction used in the studies examining flow and
Web site interaction (Berthon & Davies, 1999; Cho, 1999; Novak etal., 2000) seem to be
different from the interaction conceptualization in the current dissertation. They measured
people’s intention to click (Cho, 1999) and intention to revisit the Web (Berthon &
Davies, 1999). Also, one might argue that the exploratory behavior used in Novak et al.
(2000) may not correctly represent interaction as they measured users’ general tendency
in online navigation. None of them employed an action-based interaction measure. The
intention-based measures even differ from their likeliness to click. The intention to click
would partly explain the likeliness to interact, but it must be noted that intention-based
measures were mostly used to measure the consumer’s intention in a direction that is
favorable to the advertiser/publisher. Considering the neutral nature of the proposed
interaction concept, the intention measures would not provide a perfect fit either for the
chance of interaction or interaction per se.

Finally, Cho and Leckenby (1998) attempted to explain consumers’ banner-
clicking activity by investigating its underlying motivation, and presented advertising
values motivations (i.e., information/entertainment/usefulness), advertising

characteristics motivations (i.e., attention-Icuriosity-generating), and user characteristics

61

motivations (consumer needs/involvement/learning motivation). From the above
discussion, it is reasonable to expect that the consumers’ involvement with the product
category would be a factor that would increase the likeliness of interaction. This is also

consistent with the hypotheses described in H1 and H2.

4.6. Interaction — Consequences

When consumers are aware of the advertiser’s Web site, interaction with Web
sites was found to generate positive images for brands (Consumer Experience Probe,
1996, in Chaterjee etal., 1998). Similarly, Cho and Leckenby (1998) argued that
consumers’ interaction with messages or advertisers was likely to generate active and
intensive information processing, which in turn would result in more favorable consumer
attitudes and behaviors. They showed that consumers’ intention to click was positively
related to the attitude toward the banner ad and the advertised brand (Cho & Leckenby,
1998)

Generally, interactivity studies assume that a reciprocal and two-way
communication is a commonly desired trait in media. Interactivity definitions also
assume the audience desires interacting with others (e.g., people, media, etc.),
emphasizing exchange and mutuality. However, these assumptions are not shared by
everyone. For example, Ha and James (1998) criticized them to be unrealistic, and
proposed that individual differences in communication needs should be considered. Also,
while assessing the potential of interactive television, Lee and Lee (1995) pointed out that
individuals’ different needs in using a medium must be considered before providing

interactivity and noted interacting with a medium might be considered disturbing for

62

certain content contents or audiences. Likewise, Neuman (1991) argued that audiences
might prefer not having to interact although having a choice of interactivity would be
beneficial. It might be true — interaction (or more specifically, having to interact) may be
annoying. This may be related to the consumers’ involvement with the program, which
was accordingly hypothesized to have negative relationship with their interaction (H3).
Interactions may intensify a person’s information processing (Cho & Leckenby, 1998).
Then, how will interaction operate to make the process more intense? This dissertation
investigates the change in individual’s level of involvement as a possible consequence of
interaction. How can a person’s interaction with an object make him or her more
involved? Does anyone experience increased levels of attention, interest, and
involvement with an object after making an interaction? It is possible. For example, when
someone picks out for his/her favorite contestant during watching Fox’s American Idol,
and votes for the contestant using the provided 1-800 number, it would generally make
the person pay more attention to the result and more involved with the program (e. g.,
more wanting that contestant to win the match) compared to those who did not make such
an interaction. Similarly, voting on an issue in a Web site might generate similar effects.
These can be also explained by the audience’s investment of the time, money, or
cognitive resources.

Involvement studies have recognized that the high involvement state produces
higher level of attention, deeper information process, and more self-generated thoughts
(Gardner et al., 1978; Greenwald & Leavitt, 1984; Leigh & Menon, 1987).
Zaichkowsky’s (1985) definition states “a person’s perceived relevance of the object

based on inherent needs, values, and interests (p.342).” An object refers to anything under

63

the person’s consideration including a product class, an ad message, or purchase intention.
When regarding interaction (i.e., physical action) as an object, the above definition will
read that a person highly involved with the interaction will feel the interaction to be more
relevant to his or her needs and values. Likewise a person will feel the interaction to be
less relevant to their needs and values when the person has low involvement with the
interaction. Considering that a person’s most interactions would be generated from
recognizing his/her own needs and that it is an effort to achieve own goal, most
interactions can be described as the outcomes of at least medium level of involvement. In
addition, these high involvement interactions would reflect high level of consumer
attention and more intensive information process.

From this, one can speculate that consumer’ high involvement interactions will
occur more often when the interactions are with high involvement the product (in H1). In
addition, when considering an interaction reflects an effort to fulfill ones need, the
(continuing) interactions would not only reflect the involvement levels of an object, but it
might also reinforce the level of involvement when the course of interaction is not
significantly interrupted by other factors such as unsatisfactory results. It may be justified
by the following two rationales. First, the (series of satisfactory) interactions will (I)
produce self-generated thoughts because of the nature of interaction in two-way
communication, (2) require more attention to the stimuli and the communication process
as the individuals practice active control, and thus (3) consumers will elaborate on the
messages provided and experience deeper levels of processing. Second, flow studies note
that flow is characterized by a sequence of responses facilitated by machine interactivity

(Novak et al., 2000) and describe that users will experience flow when they perceive a

balance between their skills and the challenges of the interaction (Novak et al., 2000).
Novak et al. (2000) explained that consumers experiencing flow during online navigation
are “acutely involved in the act of online navigation (p.6).” As a result, it is expected that
the online audience’s interactions will increase the level of product involvement. Also,
because the interactions with the iPPL represent the interactions with the program itself,
it is expected to increase the level of program involvement (Study 5). However, the
increased involvement level of the program will be situational involvement because the
interactions with the iPPL or the information gathered by the interactions are not intrinsic
to the program or the program information. Therefore, it is hypothesized:
H12. Interaction with an iPPL will be positively related with the recall of the
advertised band in the iPPL.
H13. Interaction with an iPPL will increase the consumer’s involvement with
the product featured in the iPPL.
H14. Interaction with the iPPL will increase the consumer’s situational

involvement with the program.

65

Chapter 5. Methodology

5.1. Analysis Plan

The six independent variables in this dissertation include the audience’s product
involvement, program involvement, attitude toward the program (ApRoo), attitude toward
the character (ACHAR), gender, and number of interactions. The five dependent variables
are brand recall, interactions, attitude toward the advertised brand (AB), (enduring)
product involvement, and (situational) program involvement. The hypotheses are tested
using regression analyses and t-tests, and these analyses are followed by the tests of
interaction effects of the independent variables in five separate studies. Table 1 illustrates
the list of hypotheses to be tested, the interaction effects to be examined, and the
associated analytical techniques.

Each study will used a 2 x 2 factorial design. Specifically, Study I examined the
impacts of product and program involvement. Study 2 and 3 examined the effects of
attitude toward the program, combined with product and program involvement,
respectively. Study 4 and 5 examined the effects of attitude toward a character, combined
with attitude toward the program and product involvement, respectively. Finally, Study 6

will examined the effects of gender and attitude toward a character.

5.2. Design and Sample
Computer lab experiments were conducted for this study. For the experiment, a
total of 396 undergraduate college students were recruited from courses at a large

midwestem university in the US. The courses were campus-wide electives so that the

66

participants could represent a variety of majors. Participants were randomly assigned to

one of two program involvement conditions (high vs. low), and each condition was

arranged to contain similar number of male and female participants to avoid uncontrolled

 

 

 

gender effects.

Table 1. Hypotheses
Independent Variable Dependent Variable Method

H1 Product Involvement Amount of Interactions Regression

H2 Product Involvement The Number of Recalled Brands Regression

H3 Program Involvement Amount of Interactions Hotelling’s T2

H4 Program Involvement The Number of Recalled Brands Hotelling’s T2

H5 Attitude toward Program The Number of Recalled Brands Regression

H6 Attitude toward Program Amount of Interactions Regression

H7 Attitude toward Program Attitude toward Brand Regression

H8 Attitude toward Character Amount of Interactions Regression

H9 Attitude toward Character The Number of Recalled Brands Regression

H10 Attitude toward Character Attitude toward Program Regression

H11 Gender Amount of Interactions T-Test

H12 Amount of Interactions The Number of Recalled Brands Regression

H13 Amount of Interactions Changes in Product Involvement Regression

H14 Amount of Interactions Changes in Program Involvement Regression

31 Product Involvement x Amount of Interactions Two-Way Mixed
Program Involvement ANOVA

52 Product Involvement x Amount of Interactions Two-Way Mixed
Attitude toward Program Attitude toward Brand ANOVA

33 Program Involvement x Amount of Interactions Two-Way
Attitude toward Program Attitude toward Brand Between ANOVA

S4 Attitude toward Program x Amount of Interactions Two-Way
Attitude toward Characters Attitude toward Brand Between ANOVA

SS Product Involvement x Amount of Interactions Two-Way Mixed
Attitude toward Characters Attitude toward Brand ANOVA

 

67

5.3. Development of the Stimulus Material

A 25-minute episode of a popular sitcom — Friends — was used for the experiment.
Interactive television (ITV) interface was established on the computer screen to enable
interactivity of the program. First, the program was digitized, and optimized to fit the
resolution of the computer screen (800 by 600 pixels). Interactive Product Placements
(iPPLs) used in this study and accompanying interactive features were produced and
embedded using Macromedia Flash. The digital video recording (DVR, or personal video
recording — PVR) feature that allows audiences to record and replay the program was not
established due to technical limitations. However, pause and replay functions were
included. To demonstrate iPPL functions, a small icon was placed in the bottom-right

corner of the screen (Figure 2).

Figure 2. Thumbnail of Stimulus Material: Step 1

 

Under normal viewing conditions, a clickable icon is
placed in the comer of the screen.

68

A click on the icon displayed multi-tiered product information. For example, a
participant who was interested in the jacket worn by a character could click the icon to
display a small transparent menu that contained the list of the available products within a

particular scene (Figure 3).

Figure 3. Thumbnail of Stimulus Material: Step 2

i

l a

, ' 1:0,.
mm --«
in v...
M] ......
1.1 a...

ifin Iva-lo!
80 .0 ‘9'” IMO“

can hummus:

 

When the icon is clicked, the list of available products
(along with the pictures) in the scene is displayed.

When the participant found the item of interest was included in the list and
clicked the item, further product information was displayed in a new window (Figure 4).
A purchase button was included in the interface design. However, clicking the button
would display a small dialogue box in which the viewer was told that the button is not

fully functional in the experiment.

69

Figure 4. Thumbnail of Stimulus Material: Step 3

D

”O. no". came to I
man porn. in! hack-d
“In

 

When a particular item is clicked, the detailed product
information is displayed.

The products available for the iPPLs were changed as the scenes (e.g., living
room, restaurant, etc.) changed. Table 2 shows the detailed information from the episode
used for the experiment, including its scenes and embedded product information.

To select the products for the experiment, all items appearing in the program were
listed. The final products for the experiment were selected using two criteria. First,
products paired with a character were clearly being used or held by a single character.
Second, brand information such as a brand name or a logo was not visually or verbally
available in the program. This was to control possible effects of verbal or visual

endorsement.

70

Table 2. Program used for the Main Experiment

Episode #408: Chandler in a Box

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8:13.16 Dur.l Place Character Product Brand Price
1 0:56 Joey’s Apt Joey Sweater Gap $35.00
Jeans Arizona Jeans $43.00
Phone Panasonic $19.99
2 0:45 Title NO iPPLs
3 4:03 Rachel’s Apt Rachel Tableware Target $4.50
Sweater J .Crew $58.00
Skirt Eddie Bauer $42.00
Ross Sweater Polo Ralph Lauren $109.00
Monica Kitchenware Crate &Barrel $15.00
Chandler Shirt American Eagle $39.00
4 1:36 Rachel’s Apt 2 Phoebe Shirt The Limited $60.00
Ross Beverage Impulse $9.00 3
Pants Tommy Bahama $55.00
Chandler Jeans Calvin Klein $49.50
5 1:07 Joey’s Apt No iPPLs
6 0:50 Cafe Chandler Coffee Starbucks $1.00
7 0:50 Eye Doctor No Products
8 1:41 Joey’s Apt Background Sofa IKEA $649.00
Recliner La-Z-Boy $430.00
Board Office Max $29.99
CD Rack WalMart $19.95
9 1:15 Eye Doctor No iPPLs
10 6:44 Rachel’s Apt No iPPLs
11 1:51 Balcony No iPPLs
12 3:00 Rachel’s Apt Background Sofa Art Van $350.00
Background Tableware Pottery Barn $14.99
Phoebe Dress DKNY $1 89.00
Phoebe Dessert Sara Lee $3.50
13 0:30 Balcony No iPPLs
1 Duration

2 Same place in different time frames
3 Price is for a 6-pack

Next, the level of involvement for each product was considered. Existing

literature on product involvement (e. g., Ratchford, 1987; Weinberger & Spotts, 1989)

were used to categorize general involvement levels for each product. Consequently, 24

71

products were selected, 15 of which represented high-involvement products and 9 of

which represented low-involvement products. Table 3 has more information regarding the

brands and product categories.

Table 3. Summary of Product Information Embedded in iPPLs

 

 

 

 

 

 

 

 

Product Brand Product Product Character
Category Involvement
Shirt American Eagle Clothes High Chandler
Pants Calvin Klein Clothes High Chandler
Coffee Starbucks Beverage Low Chandler
Shirt Abercrombie & Fitch Clothes High Monica
Jeans Banana Republic Clothes High Monica
Mug Cup Crate & Barrel Kitchenware Low Monica
Sweater Gap Clothes High Joey
Jeans Arizona Jeans Clothes High Joey
Telephone Panasonic Electronics Low Joey
Shirt The Limited Clothes High Phoebe
Dress DKN Y Clothes High Phoebe
Dessert Sara Lee Food Low Phoebe
Shirt J .Crew Clothes High Rachel
Skirt Eddie Bauer Clothes High Rachel
Mug CuL Target Tableware Low Rachel
Sweater Polo Ralph Lauren Clothes High Ross
Pants Tommy Bahama Clothes High Ross
Energy Drink Impulse Beverage Low Ross
Sofa (Big) IKEA Furniture High Background
Sofa (Small) Art Van Furniture High Background
Recliner Chair La-Z-Boy Furniture High Background
CD Rack WalMart Home Accessory Low Background
Bulletin Board Office Max Home Accessory Low Background
Pasta Bowl Pottery Barn Tableware Low Backggound

 

As mentioned above, no particular brand was visually appeared or verbally

mentioned in the program. In order to increase the external validity of the study, brand

names were selected from existing brands instead of assigning artificial brand names. The

individual image of each product (as shown in Figure 3 and Figure 4) had to match its

72

actual appearance in the program. Therefore, each image was carefully created with
computer graphic software to make it look exactly the same as the one that was shown in
the program. Finally, each product was priced based on the actual prices of similar
products in the market.

Currently, a participant has to click twice to View the product information (i.e.,
icon and list). Instead of clicking an icon, Viewers should be allowed to click directly on
the products as they are appearing in the program. However, such an interface has not
been fully developed in the ITV industry, and due to the technological limitation, the use
of an icon is reported as being a more viable option in the industry for the time being

(Swedlow, 2000).

5.4. Procedure

Since existing brands were used, this study employed a pretest-posttest
experimental design to measure changes in participants’ attitude toward the brands. To
avoid the priming effect and reduce their fatigue, participants’ existing attitude toward the
brands was measured two to three days prior to the main experiment. In the beginning of
the posttest experiment, participants received a brief introduction to the overall
experiment, which was followed by the measures of their initial (enduring) product
involvement and (enduring) program involvement. Participants were then led to the
computer screen where they were instructed about the use of the ITV interface created for
the experiment.

To reduce the novelty effect, participants were forced to go through a short

practice session. The material for the practice session was very similar to that of the main

73

stimulus material, except that it was made from a different episode of Friends. Six
products were embedded in the 4-minute, single-scene practice material. Brand names
were selected deliberately so that they would not overlap with the brands appearing in the
main program. During the practice, participants were encouraged to click the icon and
buttons to make themselves familiar with the ITV interface.

After the practice session, the participants in the treatment condition received
information intended to increase their program involvement, whereas the participants in
the control group received none. A modified version of Wright’s (1973, 1974)
manipulating was used. First, some background information about the program was
provided to the participants in the treatment group. Second, they were told that large
cable companies (i.e., AT&T Broadband and Cox Communication) were about to launch
a test market project in their local area in which the participants would find the exact
same interface being used. Finally, the participants were told that they would be asked to
answer to certain questions regarding the storyline of the program when the program was
over. In order to maximize the effect of manipulation, the participants were told that two
participants who could provide most correct answers would win a cash prize.

As another device to minimize the novelty effect, all participants were strictly
instructed that they should interact only with products they find to be of interest. Also,
they were told that their activities on the computer screen (e.g., clicking a product,
moving a mouse pointer) would be recorded by computer software. Although their
activity was not “recorded,” the information about the products clicked by each

participant was saved in a local database for later analysis.

74

Participants watched the program wearing headsets so that other participants
would not interrupt or distract them. When the program was over, participants went
through a brief distracting task, which will be followed by the measure of brand recall.
Next, in a separate questionnaire, participants were measured for other variables. Upon
completing the questionnaire, they were debriefed and dismissed. Table 4 illustrates the
experiment’s overall procedure.

Table 4. Experimental Procedure

 

 

Steps Treatment Group Control Group

 

1 Initial measure of attitude toward brand (2 to 3 days prior to the main experiment)
Introduction to the experimental procedure
Measure of initial product & program involvement
On-screen instruction on the experimental interface
Practice session (4 minutes)

 

On-screen treatment of

. None
program Involvement

 

 

Main stimulus material (25 minutes)
Involvement and Brand Recall measure

©00QQUIJ>UJN

Other measurement including second measure ofAB (in a separate questionnaire)

 

 

5.5. Measurement

5.5.1. Measure of Involvement

All measures used in this study were adopted from previous studies in similar
domains, and they used seven-point scales to facilitate further analysis in structural
equation modeling. To examine the changes in the involvement levels, audiences’
enduring involvement with a product category was measured twice — before and after

watching the program. Product involvement was measured with 7-point semantic

75

differential scales, which were modified from the personal involvement inventory

developed by Zaichkowsky (1994). The scales had five items that were anchored by

,3 6" ’9 66 ,9 ‘6

“important/unimportant, Interesting / boring, relevant / irrelevant, exciting/

unexciting,” and “means a lot to me / means nothing.”

5.5.2. Measure of Attitude

Audience members’ attitudes toward the program and toward each character were
measured with a semantic differential scale from MaeKenzie and Lutz (1989). The five
items were anchored by “good / bad,” “pleasant / unpleasant,” “favorable / unfavorable,”
“appealing / unappealing,” and “attractive / unattractive.” The same scale was used to
measure the audience’s attitudes toward each brand appearing in the study. However, as
there are 24 brands that needed to be measured, the attitude toward the brand (AB) scale
was modified to a 3-item scale to maintain the overall length of the questionnaire at the
reasonable level. Attitudes toward each brand and the program were measured twice in
order to see the changes in their degrees before and after the experiment. As mentioned

earlier, the attitude toward each brand was measured two to three days prior to the main

experiment.

5.5.3. Measure of Recalls and Interactions

Participants’ unaided recall of advertised brands was measured using an open-
ended question, which asked them to indicate for which brands they had seen an iPPL
during the program. Finally, the data for the audience’s actual interactions with a
particular iPPL was collected with computer software. The interaction with an iPPL was

operationalized as an event in which a participant opens the final product information

76

window (as shown in Figure 4). Therefore, clicking the icon to open the list of available
items (as in Figure 3) was not regarded as an interaction with an iPPL. The stimulus
material was programmed to save each interaction made by a participant for the brand

name and the order of interaction(s).

5.5.4. Demographics

At the end of the questionnaire, participants were asked to provide some personal

information, such as gender, age, race, and class level (i.e., freshman, sophomore, etc.).

5.5.5. Manipulation Check

For a manipulation check, participants’ level of involvement with the program
was measured to examine whether the participants in the treatment group showed a
higher level of involvement. This scale, with five 7-point semantic differential items, was
identical to the scale that was used to measure participants’ level of product involvement.
Program involvement was also measured twice — once before the experiment and again

after the program ended.

77

Chapter 6. Results

6.1. Manipulation Check

A t-test was performed to examine whether the participants in the treatment group
demonstrated higher levels of program involvement than those in the control group.
Unexpectedly, the t-test indicated that there were no significant differences between the
two groups (t (393) = 1.210, p > .05). The results, in Table 5, show that both groups were
above the midpoint of the scale. As the variance of the program involvement was small,
the program involvement was trichotomized using a median split and the middle third
was removed. In further analyses, only the responses that fell into the high— and low-

involvement group were used (N = 289).

Table 5. Manipulation Check of Program Involvement

 

 

 

 

 

 

Initial Analysis

N Mean Std. Dev. t (if p
Treatment Group 198 5.4061 1.2947 -1 .210 393 .227
Control Group 197 5.2475 1 .3108
Trichotomized

N Mean Std. Dev. t df p
High Involvement 149 6.5171 .3667 -27.664 287 .000
Low Involvement 140 3.9386 1.0731

 

A total of 24 brands, in eight different product categories, were employed for the
iPPLs embedded in the stimulus material. The product categories included clothes,
furniture, inexpensive electronic goods, home accessories, tableware, kitchenware,

beverage, and dessert. As mentioned earlier, clothes and fumiture were selected to

78

represent high involvement products, while the rest of the categories were representing
low involvement products. Although this classification was based on previous literature
(e.g., Ratchford, 1987; Weinberger & Spotts, 1988), participants’ level of involvement
with the above product categories were measured because the typologies from the
aforementioned literature were based on non-student samples. One-way within-subjects

analysis of variance (ANOVA) was performed and the result is shown in Table 6.

Table 6. Product Involvement

 

 

 

 

Std.
Product Categories Mean Deviation N
Clothes 6.143 .900 288
Furniture 5.237 1.067 288
Telephones (Inexpensive electronics) 5.015 1.328 288
Beverages 4.993 1.124 288
Home Accessories 4.937 1.295 288
Kitchenware 4.369 1 .249 288
Tableware 4.176 1.185 288
Desserts 3.846 1 .473 288
df MS F p

Product Involvement/Within-Subjects 7 146.90 144.61 .00

 

Although the differences in product involvement across the categories were
significant, the result showed a different pattern from Ratchford (1987) and Weinberger
and Spotts (1988). Specifically, inexpensive electronic goods, home accessories, and
beverages were found to have moderate-to-high levels of involvement; and kitchenware
and tableware showed medium levels of involvement. Desserts were found to have a low
level of involvement as expected. Based on the above results, product involvement was

treated as a continuous variable in further tests of hypotheses.

79

6.2 Sample Size and Composition

A total of 396 usable questionnaires were collected for data analysis. However,
following the trichotomization from the above manipulation check on program
involvement, 107 responses in the middle were discarded, and 289 were used for further
analyses. The final sample consisted of 192 females (66.4%) and 97 males (33.6%). The
average age of the participants was 20.5 years, and whites (non-Hispanic) represented
72.3% of the overall participants. Finally, they consisted of 51 freshmen (17.6%), 61
sophomores (21.1%), 115 juniors (39.8%), and 62 seniors (21.4%). One hundred forty-
nine participants were assigned to the high program-involvement group and 140 were

assigned to the low-involvement group.

6.3. Scale Reliability

The reliabilities of each construct were assessed using Cronbach’s alpha. All of
them indicated high levels of reliability ranging from .86 to .97 to indicate that the scales
used in this study had adequate internal consistency and were appropriate for further

analyses. Table 6.3 summarizes each scale with its source and reliability.

6.4. The Effects of Product Involvement on Interactions

The first two hypotheses (H1 and H2) tested the impact of participants’ level of
product involvement on the amount of interactions and recalled brands, respectively.
Separate bivariate regression analyses were employed to test the hypotheses. Here only
background products and brands that are not paired with particular characters were

selected for the analysis because participants’ attitude toward certain characters might

80

Table 6.4.1. Descriptive Statistics and Reliabilities for Scales Used in the Study

 

 

 

 

 

 

 

 

 

Variables Subscale Alpha Subscale Alpha Overall . Items
Alpha 1n Scale
Program Time 1 .9476 Time 2 .9472 .9501 5
Involvement
Attitude Time 1 .9636 Time 2 .9661 .9649 5
toward the Program
Product Tableware .861 5 Clothes .885 1 .8911 5
Involvement"I Beverages .8751 Home Acc. .9237
Telephones .8922 Desserts .9314
Furniture .871 7 Kitchenware .8885
Attitude Chandler .9 l 99 Joey .9089 .9180 5
Toward Character Monica .9200 Phoebe .9173
Rachel .9120 Ross .9296
Attitude J.Crew .9530 IKEA .9705 .9492 3
toward the Brand* Gap .9361 Art Van .9499
Eddie Bauer .9177 La-Z-Boy .9512
Banana Republic .9582 Target .9502
American Eagle .9581 WalMart .9694
Arizona Jeans .9112 Office Max .9387
Tommy Bahama .9324 Pottery Barn .9682
The Limited .9627 Crate & Barrel .9641
DKNY .9455 Panasonic .9298
Sara Lee .9371 Abercrombie .9679
Impulse .9423 Calvin Klein .9439
Starbucks .9649 Polo .9576
Mean Std. Deviation
Interactions 3 .9343 3 .73 1 3
Recalled Brands 1.6125 2.1234

 

* Product Involvement and Attitude toward a brand were also measured twice. Listed alphas for
individual brands and product categories Show the average alpha scores of Time 1 and Time 2.

have influenced the amount of interactions and recall along with their product

involvement. Therefore, the product involvement with fumiture, home accessories, and

tableware were examined in relation to the interactions and recall of those categories.
(See Table 3 for information on background products and brands.)

Hypothesis 1 posited that participants who have higher level of involvement with
a particular product category would demonstrate a higher amount of interactions with
iPPLs featuring brands of that product category. To test this, the amount of interactions
with the brands in each product category (i.e., fumiture, home accessories, tableware)
was regressed on the product involvement on each category. The results are summarized
in Table 6.4.2 and 6.4.3. They indicate that participants’ level of involvement was
positively associated with the amount of interactions in all three product categories, and

thus Hypothesis 1 was supported.

Table 6.4.2. Descriptive Statistics on Product Involvement (BG) and Interactions

 

 

 

Involvement Interactions“
Product Category Mean Std. Deviation Mean Std. Deviation
Furniture 5.2433 1.0698 1 .5017 1 .1400
Home Accessories 4.9375 1.2926 .9481 1.0479
Tableware 4.1813 1.1868 .4429 .5181

 

 

 

 

 

" Three furniture brands, two home accessory brands, and one tableware brand were used in the study, and
the listed amount of interactions represent the average amount of interactions per each brand to ease

comparison.

Table 6.4.3. Relationship between Product Involvement and Interactions

Independent Variables: Product Involvement
Dependent Variable: Amount of Interactions

 

 

Std. Adjusted Std. Error of
Product Category Coefficients t df R2 R2 the Estimate p
Furniture .509 10.007 287 .259 .256 .9832 .000
Tableware .540 10.864 287 .291 .289 .4369 .000
Home Accessories .385 7.071 287 .148 .145 .9687 .000

 

 

82

6.5. The Effects of Product Involvement and Attitude toward the Program on Recall

Similarly, only background products and brands were analyzed in testing the
relationship between product involvement and the number of brands recalled (Hypothesis
2). However, the levels of product involvement in three product categories were averaged,
and tested in relation to the combined amount of recalls of the six brands that belong to
the three product categories. As seen in Table 6.5.1, the overall amount of recall was

small and most participants reported no recalls of those background brands.

Table 6.5.1. Descriptive Statistics on Brand Recalls

 

 

 

 

 

Mean Std. Deviation
Recall: Background Brands .8028 1.0438
Recall: All Brands 1.6125 2.1234
Product Involvement: BG Catigories Combined 4.7874 .9338

 

Hypothesis 2 posited that participants who have higher level of involvement with
a particular product category would recall more brands that appear in iPPLs of that
product category. Also, Hypothesis 5 posited that participants with more positive attitude
toward the program would recall more brands. These hypotheses were tested with
multiple regression, and the overall number of recalled brands in the three baCkground
product categories was regressed on the combined product involvement of each category
and participants’ attitude toward the program. The results are summarized in Table 6.5.2.

Participants’ level of involvement was found to be positively associated with the
number of brands recalled (R2=.079; p<.01). Although the strength of the association is
weaker than that of Hypothesis 1, the result shows that Hypothesis 2 was supported. On

the contrary, the results indicate that participants’ attitude toward the program is not

83

significantly related with the recall (i.e., significance of coefficient >.05). From this,

Hypothesis 5 was rejected.

Table 6.5.2. Relationship between Product Involvement, Ammo, and Recalls

Dependent Variable: Amount of Recalled Brands

 

 

Independent Std. Sig. Adzj. Std. Error of
Variable Coefficients t (if Coeff. R2 R the Estimate p
Involvement“ .269 4.498 286 .000 .079 .073 1 .0494 .000
Attitude-Program .036 .602 .548

 

 

* Involvement with the product categories that incorporate background brands

6.6. The Effects of Program Involvement on Recall and Interactions

Hypothesis 3 and Hypothesis 4 posited that participants who report higher levels
of involvement with the program would show fewer interactions and brands recalled,
respectively. As mentioned earlier, program involvement was trichotomized and the
conditions at the two ends (high vs. medium) were compared to test these hypotheses.
Again, only the background brands were taken into the analyses.

Multivariate analysis of variance was performed on the amount of interaction and
recalled brands, and Hotelling’s T2 was examined as it is not acceptable to use separate t-
tests on possibly correlated dependent variables (Tabachnick & F idell, 2000, p, 20). The
Hotelling T2 statistics provide a single overall test of the group differences on two
separate dependent variables — amount of interactions and recalled brands. The results are
summarized in Table 6.6.

The results indicated that the differences were significant. However, they were not
in the predicted direction. Participants in the high program involvement condition

demonstrated higher levels of interactions (M=3.41) than those in the medium-level

84

condition (M=2.34). Likewise, participants in the high program involvement condition

recalled more brands (M=.93) than those in the medium-level condition (M=.66).

Therefore, both Hypothesis 3 and Hypothesis 4 are rejected.

Table 6.6. Interactions and Recalls in Program Involvement Conditions

 

 

 

Dependent Program Std. Wilks’ Approx
Variables InV. Mean Dev. t Sig.t Lambda F df Sig. F
Interactions High 3.409 2.125 -2.2009 .029 .923 11.937 286 .000
Medium 2.343 1 .700
Recall High .933 1 .101 -4.6938 .000
Medium .664 .964

 

 

 

 

6.7. The Effects of Attitude toward the Program on Interactions

Hypothesis 6 posited that participants with more positive attitude toward the

program (Apaoo) would show fewer interactions with iPPLs. To test the relationship, the

amount of interactions with background brands was regressed on the attitude toward the

program. Table 6.7 summarizes the results.

Table 6.7. Relationship between Apnoc and Interactions

Independent Variables: Attitude toward the Program

 

 

Dependent Std. Adjusted Std. Error of
Variable Coefficients t df R2 R2 the Estimate p
Interactions .175 3.009 287 .031 .027 1.9723 .003

 

 

The results indicate that participants’ attitude toward the program was positively

associated with the amount of interactions with background brands. However, the

direction of the association indicates that it is positively related, which means that a more

85

positive attitude toward the program resulted in a higher number of interactions.
Furthermore, the ApRog was not found to be a major predictor for the amount of
interactions (Adjusted R2=.027), which implies that the relationship is not highly
meaningful. Although the association was statistically significant, Hypothesis 6 could not

be supported.

6.8. The Effects of Attitude toward the Program on Attitude toward Brands in the
Program

Hypothesis 7 tested the affect transfer hypothesis (Mitchell & Olson, 1981; Shimp,
1981) — particularly the influence of participants’ attitude toward the program (Apnoo) on
the attitude toward the brands (AB) appearing in the program. Unlike the previous
analyses that used only background brands, this analysis was conducted on all 24 brands
appearing in the program. Because existing brand names were used in the stimulus
material, the changes in A3 (A32 - Am: the difference between the second measure ofAB
and first measure ofAB) were examined for the analysis. Also, the changes in A3 of the
non-interacted brands were not taken into the analysis. It should be noted that the
participants were exposed to the brand information only when they interacted with the
corresponding iPPLs.

This means that the participants who did not interact with the iPPL of “Brand A”
had no means to find out whether the “Brand A” appeared in the program. Consequently,
it is reasonable to expect that the difference between the A32 and A31 of non-interacted
brands should not be statistically significant. To test this relationship and hypothesis,

participants’ interaction responses had to be recoded. Interacted brands and non-

86

interacted brands varied for every single participant. That is, participant A could have
interacted with brands A, C, and D, while participant B could interact with brands C, E,
and F. Therefore, for each participant, all brands were grouped into either “interacted” or
“non-interacted” brands. Next, the first A3 measures of those “interacted (A31 1)” and
“non-interacted (A3Nu)” brands were calculated one by one for each participant, and the

same was done for the second measure of A3 (i.e., A312 and A3332) Finally, the difference

between A32 and A131 (i.e., changes in A3) for both interacted brands (A313) and non-

interacted brands (A3333) were calculated based on those newly calculated values, and

used for further analyses. The Table 6.8.1 illustrates this process.

Table 6.8.1. Calculation of Interacted- and Non-interacted A3: Example

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Brand A Brand B Brand C Brand D Mean
A31 A132 A80 A131 A82 ABD A31 A32 A13D A81 A32 A80 A31 A32 ABD
Participant 1 2 3 5 2 4 4 0 6 7 l 5 6 7 1 4.5 5.5 1.0
Interacted Interacted Non-interacted Non-interacted
Participant 2 4 l 6 .. 6 0 ,, 5 5 0 5 7 ,2 2 4 g 2 4.5 5.5 1.0
Non-interacted Non-interacted Interacted Interacted
A}: Measure 1 A9: Measure 1 Ag: Difference
Participant 1 AB" = (3 + 4) / 2 = 3.5 A312 = (5 + 4) / 2 = 4.5 A3“) = 4.5 - 3.5 = 1.0
ABN11=(6 + 5) / 2 = 5.5 ABNIZ = (7 + 6) / 2 = 6.5 Age") = 6.5 - 5.5 =1.0
Participant 2 AB“ = (5 + 2) / 2 = 3.5 A312 = (7 '1' 4) / 2 = 5.5 ABID = 5.5 - 3.5 = 2.0
ABNII =(6+5)/2= 5.5 ABh_ll_2=(6+ 5)/2= 5.5 Am =5.5-5.5=0.0

 

 

 

In the above example, the mean scores of all A3 changes (A33) are equal for both

participants. However, the difference between the two participants can be found when

calculating interacted and non-interacted brands separately, and it turns out that

Participant 2 showed more improvements in A3 over interacted brands compared to

Participant 1.

87

As mentioned earlier, A31 is assumed to be equal to A32 if the brand was not
interacted with. To test this assumption, a one-sample t-test was performed using the
changes in A3 scores for the non-interacted brands, and the result (Table 6.8.2) indicates
that the difference was not significant.

In testing Hypothesis 7, participants who did not make any interactions at all were
eliminated from this analysis. It is because, by definition, the participants with no
interactions are not expected to demonstrate any changes in A3. This decreased the total
number of participants to 242. Finally, to test Hypothesis 7, the differences in attitude
toward the interacted brands (A313) were regressed on the participants’ attitude toward
the program (A3303). Table 6.7.3 summarizes the results.

The results indicate that participants’ attitude toward the program was positively
associated with the changes in their attitude toward the interacted brands (A313).

Therefore, Hypothesis 7 was supported.

Table 6.8.2. A3 Changes in Non-Interacted Brands: Statistical Significance

 

 

Std.
Min" Max“ Mean Dev. Median Mode t (If 1
A33") -1.3667 1.2698 .0402 .3799 .0401 .0000 1.606 229 .1 10

 

 

 

 

 

Table 6.8.3. Regression: Relationship between A313 and Amos

Independent Variable: Changes in Attitude toward the Interacted Brands (A3313)

 

 

Dependent Std. Adjusted Std. Error of
Variable Coefficients t df R2 R2 the Estimate p
APROG .491 8.660 236 .241 .238 .6148 .000

 

 

88

6.9. The Effects of Attitude toward the Characters (of the Program) on Interactions and
Recall

The next two hypotheses (H8 and H9) tested the impact of participants’ attitude
toward each character on the amount of interactions and recalled brands. Separate
bivariate regression analyses were employed to test the hypotheses. However, in these
analyses, only the brands that are paired with particular characters were selected because
the interactions or recalls of unpaired brands (i.e., background brands) would be free
from the effects of participants’ attitude toward characters. (See Table 3 for information
on brands paired with particular characters.)

Hypothesis 8 posited that participants who have a positive attitude with each
character (ACHAR) would demonstrate a higher number of interactions with iPPLs paired
with each character. To test this, the number of interactions with the iPPLs paired with
each character (i.e., Chandler, Joey, Monica, Phoebe, Rachel, and Ross) was regressed on
the attitude toward each character. The results are summarized in Table 6.9.1 and 6.9.2.

The results indicate that in most cases (except for one character, Joey)
participants’ attitude toward the character was positively associated with the amount of
interactions. The results partially support Hypothesis 8. However, it should be noted that
the effect sizes were relatively small. The adjusted R2 ranged from .014 (Ross) to .218
(Rachel). But when excluding Rachel, the range of the adjusted R2 reduces to .014 to .099
(Chandler). Although these are weak, they are still stronger than the relationship between
the attitudes toward the program and the amount Of interactions (Adjusted R2 = .027, see
Table 6.7). This illustrates that although both attitudes could not explain much of the

audience’s amount of interactions, the attitude toward the character is a better predictor.

89

Table 6.9.1. Descriptive Statistics on Character: A03“; and Interactions

 

 

 

 

 

 

Number of ACHAR Interactions
Characters Paired iPPLs Mean Std. Dev. Mean Std. Dev.
Chandler 3 5.9856 1.0807 .2215 .6662
Joey 3 5.7780 1.0984 .5848 .9170
Monica 3 5.8047 1 .0755 .4583 .8462
Phoebe 3 5.5356 1.2469 .6159 .9545
Rachel 3 6.3453 .7980 1.1869 1.3918
Ross 3 5.6287 1.1895 .4948 .9687

 

Table 6.9.2. Relationship between ACHAR and Interactions

Independent Variables: A3333

Dependent Variable: Amount of Interactions

 

 

Std. Adjusted Std. Error of
Characters Coefficients t df R2 R2 the Estimate p
Chandler .320 5.714 287 .102 .099 .8704 .000
Joey -.114 -1.947 287 .013 .010 .6630 .052
Monica .204 3 .531 287 .042 .038 .8298 .000
Phoebe .277 4.879 287 .077 .073 .9188 .000
Rachel .469 9.006 287 .220 .218 1.231 1 .000
Ross .130 2.224 287 .017 .014 .9621 .027

 

 

Hypothesis 9 posited that participants with a positive attitude toward each

character would show more recalled brands paired with each character. Similarly, only

paired brands were analyzed in testing Hypothesis 9, but the number of recalled brands

was collapsed into a single score (i.e., the number of recalled brands paired with any

characters) because most participants could recall nothing when each character’s brands

were separately examined. (see Table 6.9.3.) Also, as in testing Hypothesis 7, participants

who did not make any interactions at all were eliminated from this analysis. It was

because the participants who did not make any interactions were not expected to recall

any brands. This reduced the total number of participants to 242.

90

To test Hypothesis 9, the combined amount of recall was regressed on the
participants’ combined attitude toward the characters. The results are summarized in
Table 6.9.4, and it was found that the relationship was not significant. Therefore,

Hypothesis 9 is rejected.

Table 6.9.3. Descriptive Statistics on A33“; and Recalls of Paired Brands

 

 

Mean Std. Deviation Median Mode
ACHAR (Combined) 5.9491 .7388
Recall: Paired Brands .9545 1.9201 .00 .00

 

 

 

Table 6.9.4. Regression: Relationship between Acmn (combined) and Recalls

Indgpendent Variable: ACHAR (combined)

 

 

Dependent Std. Adjusted Std. Error of
Variable Coefficients t df R2 R2 the Estimate p
Recall .000 .003 240 .000 -.004 1 .9241 .998

 

 

6.10. The Effects ofAttitirde toward the Characters on the Attitude toward Paired Brands
As with Hypothesis 7, Hypothesis 10 tested another instance of the affect transfer
hypothesis, but in this case the influence of participants’ attitude toward each character
(ACHAR) on the attitude toward the paired brands (A3) was examined. As mentioned
earlier, the changes in attitude toward brands were used because existing brands were
used in this study. Also, paired brands (excluding background brands) were taken into the
analysis again, and participants who did not interact at all were not examined. A series of

bivariate regression analyses were performed to test the hypotheses. Table 6.10.1 shows

91

the descriptive information regarding participants’ attitude toward brands paired with

each character, and Table 6.10.2 shows the result of the hypothesis testing.

The results indicate that in most cases (except for one character, Phoebe)

participants’ attitude toward the character was positively associated with the attitude

toward the paired brands. Therefore, Hypothesis 10 was partially supported. When

comparing the effect sizes in Table 6.10.2 (i.e., Adjusted R2 ranging from .023 to .084) to

that in Table 6.8.3 (Adjusted R2 = .238), it can be found that the consumers’ A3303 is a

better predictor than ACHAR in explaining the changes in their A3.

Table 6.10.1. Descriptive Statistics on Character: Acrmr and Interactions

 

 

 

Number of A3*

Characters Paired iPPLs Mean Std. Dev.
Chandler 3 .2824 .6819
Joey 3 . 2679 .6862
Monica 3 .2006 .6162
Phoebe 3 .0744 .7847
Rachel 3 . 1646 .6514
Ross 3 .3147 .7279

 

 

 

* A3 represents the changes in A3 between the first measure and second measure.

Table 6.10.2. Relationship between A0333 and A3

Independent Variables: ACHAR
Dependent Variable: Changes in A3

 

 

Std. Adjusted Std. Error of
Characters Coefficients t df R2 R2 the Estimate p
Chandler .297 4.816 240 .088 .084 .6526 .000
Joey .163 2.564 240 .027 .023 .6784 .011
Monica .199 3.145 240 .040 .036 .6051 .002
Phoebe .108 1.683 240 .012 .008 .7817 .094
Rachel .257 4.1 1 8 240 .066 .062 .6309 .000
Ross .218 3.464 240 .048 .044 .71 18 .001

 

 

92

6.11. The Effects of Gender on Interactions

Hypothesis 11 posited that male audiences would show more interactions with the
embedded iPPLs than female participants. To test this, all interactions (i.e., interactions
with paired brands and background brands) were examined, and a t-test was performed to
see if male participants and female participants showed different amounts of interactions.
Table 6.11 shows the result, and it shows that male participants and female participants
did not show significant differences in terms of the amount of interactions. Therefore,
Hypothesis 11 was rejected. (Interestingly, females actually had a higher mean than males,

so if the differences had been significant, Hypothesis 11 would still be rejected.)

Table 6.11. Amount of Interactions Across Genders

 

 

Mean Std. Deviation t df p
Male 3.3505 2.6287 -1.899 287 .059
Female 4.2292 3.7570

 

6.12. The Relationship between Recall and Interactions

Hypothesis 12 posited that participants’ amount of interactions and the number of
recalled brands would be positively correlated. To test this, interactions and recall on all
brands (i.e., background and paired) were examined, and the number of recalled brands
was regressed on the amount of interactions. The results show that the Hypothesis 12
was supported (R2 = 239, p=.00). (see Table 6.12.)

Table 6.12. Regression: Relationship between Interactions and Recalls

 

 

 

Std. Adjusted Std. Error of
DV Coefficients t (If R2 R2 the Estimate p
Recall .489 9.506 287 .239 .237 1.8551 .000

 

93

6.13. The Effects of Interactions on Changes in the Level of Product Involvement

Hypothesis 13 posited that participants’ interactions with iPPLs would increase
their level of involvement with product categories with which they interacted.
Interactions with all brands (i.e., background and paired) were examined. Also, since the
relationship focuses on the effect of interactions on involvement change, the changes
between two product involvement measures were examined. (Note that the product
involvement was measured twice.)

Each of eight product categories was tested. Table 6.13.1 shows the descriptive
nature of changes in product involvement in those categories, and the results are in the
opposite direction from what was expected in the hypothesis. That is, most categories
except tableware show decreased levels of involvement. Eight separate bivariate
regression analyses were performed to examine their significance, and the results are

found in Table 6.13.2.

Table 6.13.1. Descriptive Statistics on Product Involvement Changes

 

 

Product Categories Measure 1 Measure 2 Difference (Mean) Std. Deviation
Clothes 6.1459 6.1400 -.0059 .5278
Furniture 5.2433 5.2275 -.0157 .8649
lnexp. Electronics 5.0146 4.7640 -.2507 .9002
Beverages 4.9933 4.8907 -.1026 .8048
Home Accessories 4.9375 4.9152 -.0223 1.0618
Kitchenware 4.3779 4.2794 -.0984 1.0045
Tableware 4.1813 4.3910 .2097 .9343
Desserts 3.8457 3.7162 -.1295 .5544

 

 

 

The results indicate that the changes in product involvement were significant in
only three categories (i.e., clothes, furniture, and tableware). However, clothes and

furniture displayed decreases in involvement. Besides, considering the increase in

94

tableware was only marginal (p = .041), it can be interpreted that Hypothesis 13 is

rejected.

Table 6.13.2. Relationship between Interactions and Product Involvement

 

 

Product Std. Adjusted Std. Error of
Categories Coefficients t df R2 R2 the Estimate p
Clothes .161 2.766 287 .026 .023 .5218 .006
Fumiture .156 2.675 287 .024 .021 .8558 .008
Inexp. Electronics .006 .104 287 .000 -.003 .9018 .918
Beverages .056 .952 287 .003 .000 .8049 .342
Home Accessories .045 .728 287 .002 -.002 1.0626 .467
Kitchenware .060 l .013 287 .004 .000 1 .0044 .3 12
Tableware .120 2.054 287 .014 .01 1 .9292 .041
Desserts .091 1.545 287 .008 .005 .8737 .123

 

 

6.14. The Effects of Interactions on Changes in the Level of Program Involvement

Finally, Hypothesis 14 posited that participants’ interactions with iPPLs would
result in increased level of involvement with the program. The difference between two
program involvement measures (i.e., Program involvement measure 2 - Program
involvement measure 1) was regressed on the amount of interactions. Again, all
interactions within the program (i.e., interactions with background and paired brands)
were examined.

Table 6.14.2 shows the result, which indicates that the amount of interactions was
significantly related with the increase in the program involvement (R2 = 210, p=.00).

Therefore, Hypothesis 14 was supported.

95

Table 6.14.1. Descriptive Statistics on Program Involvement Changes

 

Measure 1 Measure 2 Difference (Mean) Std. Deviation
Program Involvement 4.9232 5.2680 .3448 .4913

 

 

 

 

Table 6.14.2. Relationship between Interactions and Changes in Program

 

 

Involvement

Dependent Std. Adjusted Std. Error of
Variable Coefficients t df R2 R2 the Estimate p
Program .458 8.734 287 .210 .207 .4374 .000
Involvement

 

 

6.15. Interaction Effects of Product Involvement and Program Involvement on
Interactions (Study 1)

In addition to the above hypotheses, interaction effects were tested as well. First,
the roles of two different involvements (i.e., involvement with product categories and the
program) on interactions were examined. Based on the hypotheses, involvement with a
certain product category was expected to increase the amount of interactions with iPPLs
of the corresponding category, whereas the involvement with the program was expected
to decrease the overall amount of interactions. The purpose of examining the interaction
effect of both involvement types was to test which involvement type would have a greater
effect on the amount of interactions. However, the results from the hypothesis test
showed that participants’ involvement with the program also had a positive relationship

with the amount of interactions. (See Hypothesis 3.)

96

Nevertheless, the interaction effect of program involvement x product
involvement was examined. A mixed analysis of variance (ANOVA) was conducted
because program involvement was a between-subj ect variable and product involvement
was a within-subject variable. Interactions were examined only for background brands,
excluding interactions with brands paired with a particular character.

First, the amount of interaction with particular product categories had to be
calculated based on the product involvement. As mentioned earlier, three product
categories were used for background brands, and they were furniture (3 brands), home
accessories (2 brands), and tableware (1 brand). Among these, furniture (M = 5.24) and
home accessories (M = 4.94) were combined into high involvement product, whereas
tableware (M = 4.18) was categorized as a low involvement product. The high
involvement product category had five brands, thus, to ease the comparison, the amount
of interactions per brand was calculated for high involvement products. Next, mixed
ANOVA was conducted. Table 6.15 shows the results.

Interaction effects were not found between program- and product involvement.
Moreover, the results show that the main effect of product involvement was not
significant either. This is an interesting finding because the earlier test of Hypothesis 1
indicated a positive relationship between product involvement and the amount of
interactions (Table 6.4.2). It should be noted that simple regression analysis was
employed to test Hypothesis 1 (i.e., Effects of Product Involvement on Interactions).
However, in examining the interaction effect the amount of interactions was
dichotomized to those in high involvement products and low involvement products, and

the mean scores were compared. The non-significant result of the main effect of product

97

involvement can be explained by the loss of variance of product involvement that occurs

in this dichotomization process.

Table 6.15. Interactions Effects: Product Involvement x Program Involvement

Dependent Variable: Amount of Interactions with Background Items

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement High Inv. with .568 .385 149
Program
Products Low Inv. with .401 .322 140
Program
Low Involvement High Inv. With .550 .538 149
Program
Products Low Inv. With .329 .471 140
Program
Mean
Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 .30 2.22 .137
Program Involvement (Between-Subjects) 1 5.46 21.77 .000
Product x Program Involvement l .11 .81 .368
Error 287 .13

 

6.16. The Interaction Effects of Product Involvement and Attitude toward Program on
Interactions and Attitude toward Brands (Study 2)

The interaction effects of product involvement x attitude toward the program on
the amount of interactions and attitude toward brands were tested. Earlier results of
hypothesis testing showed that product involvement was positively related with the
amount of interactions (Hypothesis 1), attitude toward the program was positively related
with the amount of interactions (Hypothesis 6) and with the attitude toward brands

(Hypothesis 7). To test the interaction effect, only background brands were examined.

98

Furniture (3 brands) and home accessories (2 brands) were used for the high involvement
products, and tableware was used for the low involvement product. The mean scores of
attitude toward brand in each product involvement condition were calculated for analysis.

Next, attitude toward program was trichotomized for the analysis as the
participants demonstrated generally positive attitude toward the program. (See Table
6.16.1 for descriptive statistics.) With the middle group eliminated, there were 202

responses available for this analysis.

Table 6.16.1. Attitude toward Program: Creating Conditions

 

 

Attitude Scores . . . . . "
Mean Median Condrtromng Criteria N
Attitude toward 5.8683 6.20 Highly Positive“ M = 7.0 101
Program Med-Positive 5.7 < M < 7.0 87
Less Positive" M < 5.7 101

 

 

 

 

* used for conditioning
** Total N = 289

With these newly created attitude toward program conditions (Highly Positive vs.
Less Positive) and product involvement categories, a mixed ANOVA was conducted
again as the product involvement was a within-subject variable. Table 6.16.2 summarizes
the results of the product involvement x attitude toward the program interaction effect on

the amount of interactions.

99

Table 6.16.2. Interactions Effects: Product Involvement x Program Attitude

Dependent Variable: Amount of Interactions with Background Items

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive A3303 .514 .377 101
Products Less Positive A3303 .406 .305 101
Low Involvement Highly Positive A3303 .475 .540 101
Products Less Positive A3300 .307 .464 101
Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 .48 3.34 .069
Program Involvement (Between-Subjects) 1 1.93 8.48 .004
Product x Program Involvement 1 .09 .64 .425
Error 200 .14

 

The interaction effects on the amount of interactions were not found between
product involvement and attitude toward the program. Next, the same interaction effect
(i.e., product involvement x attitude toward the program) was examined on the changes
in the attitude toward the brands in iPPLs (A32 - A31: the difference between the second
measure of A3 and first measure of A3). Product involvement was not hypothesized to
affect the valence of attitude toward the program. However, the affect transfer hypothesis
predicts that the positive or negative attitude toward the program would be reflected in
the attitude toward the brands placed in the program. The purpose of examining this
interaction effect was to see in which product involvement condition (i.e., high vs. low)
the affect transfer hypothesis would be more strongly identified. Another mixed ANOVA
was conducted.

As in testing Hypothesis 7 (i.e., effect of attitude toward program on attitude

toward brands), the participants who made at least one interaction were taken into the

100

analysis because those who did not interact with iPPLs were not assumed to show any
significant changes in attitude toward brands. Test of Hypothesis 7 employed all 24
brands in this analysis. However, the changes in attitude toward background brands were
examined in testing the Product Involvement x Program Attitude interaction effect. It was
to control for the possible effects from attitude toward each character. Table 6.16.3 shows

the results of the interaction effect analysis.

Table 6.16.3. Interactions Effects: Product Involvement x Program Attitude

Dependent Variable: Changes in Attitude toward the Background Brands

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive A3300 — .015 .639 89
Products Less Positive A3303 .165 .642 75
Low Involvement Highly Positive A3303 .127 .703 89
Products Less Positive A3303 .267 1.047 75
Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) l 1.21 2.29 .132
Program Attitude (Between-Subjects) 1 2.08 3.21 .075
Product x Program Attitude 1 .03 .07 .799
Error 162 .53

 

No interaction effect was found. However, the main effect of the attitude toward
the program (A3303) suggests an interesting finding. Although not statistically significant
Q) = .075), the mean scores of changes in attitude toward brands (A3) show that the
changes were greater in less positive A3303, suggesting a negative relationship between
A3303 and A3. This is opposite to the findings in Hypothesis 7, where A3303 was found

to have a positive relationship with A3. Also, the participants with highly positive A3303

101

showed a decrease in A3 of high involvement products (M = -.015). It was not

statistically significant, but it suggests another interesting implication.

6.17. The Interaction Effects of Program Involvement and Attitude toward Program on
Interactions and Attitude toward Brands (Study 3)

The interaction effects of program involvement x attitude toward the program on
the amount of interactions and attitude toward brands were tested by two-way ANOVA.
Results from Hypothesis 3 and 6 showed that the amount of interactions was positively
related with program involvement and attitude toward the program, respectively. The
between-subjects attitude toward the program conditions (i.e., highly positive vs. less
positive) and program involvement conditions (i.e., high vs. low) created previously were
used. Also, only background brands were examined. First, the interaction effect was
examined on the participants’ amount of interactions.

The results in Table 6.17.1 show that no interaction effect was found. Likewise,
the results indicate that the main effect ofAPROG was not significant. In fact, attitude
toward the program (A3300) was found to have a positive relationship with the amount of
interactions (Hypothesis 6). This conflict of results might be explained by the loss of
variance ofAPROG that occurred in the trichotomizing process. That is, the effect OprRoo
was tested with a simple regression analysis in testing Hypothesis 6 and yielded a
significant but weak relationship (R2=.031). The trichotomization process ignores the
variance oprRoo in the same condition and consequently the main effect could have

produced non-Significant results. The above results also indicate that A3303 has a

102

negative relationship with the amount of interactions in low program involvement

condition. However, this relationship was not found to be statistically significant.

Table 6.17.1. Interactions Effects: Program Involvement x Program Attitude

Dependent Variable: Amount of Interactions

 

 

 

 

Std.

Independent Variables Mean Deviation N
High Program Highly Positive A3300 3.20 2.109 91
Involvement Less Positive A3303 2.67 1.506 6
Low Program Highly Positive A3303 2.00 .943 10
Involvement Less Positive A3303 2.34 1.609 95

Mean
Independent Variables df Square F p
Program Involvement 1 25.654 6.916 .009
Program Attitude 2 6.265 1.689 .187
Program Involvement x Program Attitude 2 2.586 .697 .499
Error 283 3.709

 

The interaction effect of program involvement x attitude toward the program on
the changes in attitude toward brands was also tested. Results from Hypothesis 7 showed
that attitude toward brands was positively related with attitude toward the program. On
the other hand, program involvement is not hypothesized to be related with the valence of
attitude toward brands. The purpose of examining this program involvement x attitude
toward the program interaction effect was to see in which program involvement condition
(i.e., high vs. low) the affect transfer hypothesis would be more strongly identified.
Another two-way ANOVA was conducted, but no interaction effect was found. Table

6.17.2 illustrates the results.

103

Table 6.17.2. Interactions Effects: Program Involvement x Program Attitude

Dependent Variable: Changes in Attitude toward the Background Brands

 

 

 

 

Std.

Independent Variables Mean Deviation N
High Program Highly Positive A3303 .0542 .5636 91
Involvement Less Positive A3303 .4630 .6292 6
Low Program Highly Positive A3303 -.4722 .5588 10
Involvement Less Positive A3303 .1941 .6250 95

Mean
Independent Variables df Square F p
Program Involvement 1 2.897 8.521 .004
Program Attitude 2 2.335 6.870 .001
Program Involvement x Program Attitude 2 .355 1.044 .353
Error 283 .340

 

Another problem from the above results is that the program involvement and the
program attitude were too closely related. Note that this relationship was not
hypothesized and not tested because the program involvement would not affect the
valence of the attitude. The relationship is found from the sample size in each cell. (see
Tables 6.17.1 and 6.17.2.) Both program involvement and attitude toward the program
were trichotomized, and the participants in the two extreme ends were taken into the
analyses. Consequently, participants were similarly distributed in high (N = 149) and low
(N = 140) involvement conditions. Likewise, participants were similarly distributed in
positive (N = 87) and less positive (N = 101) A3303 conditions. Nevertheless, most
participants ended up in either high-involvement / highly-positive A3303 or low-
involvement/less-positive A3303 condition because the involvement and attitude were

closely related.

104

6.18. The Interaction Effects of Attitude toward Program and Attitude toward Characters
on Interaction and Attitude toward Brands (Study 4)

The interaction effects of attitude toward the program (A3300) x attitude toward
the character (ACHAR) on the amount of interactions and attitude toward brands were
tested. Three conditions for attitude toward character (i.e., highly positive vs. moderately
positive vs. less positive) were created for each of the six characters by trichotomizing
participants’ responses on A3333 (Table 6.18.1). To test the interaction effect, paired
brands with particular characters were examined. Three brands were paired with each
character, and the aggregated amount of interactions was used for analysis.

First, a series of two-way ANOVA was performed on ACHAR of each character to
examine the A3300 x ACHAR interaction effect on the amount of interactions. Earlier
results of hypotheses showed that attitude toward the program and toward five of six
characters were positively related with the amount of interactions (Hypotheses 6 and 8).
Table 6.18.2 shows the interactions effect for Chandler, and a significant result was found.

The results show that the amount of interactions increases as the attitude toward
the character improves. Generally, the amount of interactions is greater when A3300 is
less positive. However, the result shows that when the level ofACHAR is moderate,
participants with more positive A3303 demonstrated greater amount of interactions. A
post hoc test revealed that in both A3303 conditions, participants with highly positive
attitude toward Chandler (ACHAR) showed more amount of interactions than those with

moderately positive and less positive ACHAR.

105

Table 6.18.1. Attitude toward the Character: Creating Conditions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Scores
Character Mean Median Conditioning Criteria N
Chandler 5.7780 6.00 Highly Positive M > 6.4 97
Moderately Positive 99
Less Positive M < 5.4 93
Joey 5.9856 6.40 Highly Positive M > 6.6 104
Moderately Positive 79
Less Positive M < 5.8 106
Monica 5.8047 6.00 Highly Positive M > 6.4 99
Moderately Positive 85
Less Positive M < 5.4 104
Phoebe 5.5356 5.80 Highly Positive M > 6.2 97
Moderately Positive 90
Less Positive M < 5.2 102
Rachel 6.3453 6.60 Highly Positive M > 6.8 105
Moderately Positive 78
Less Positive M < 6.2 106
Ross 5.6287 6.00 Highly Positive M > 6.2 106
Moderately Positive 80
Less Positive M < 5.2 103
All 5.8464 6.00 Highly Positive M > 6.30 90
Characters Moderately Positive 107
Less Positive M < 5.54 92
Table 6.18.2. Interactions Effects: A3300 x ACHAR: Chandler
Dependent Variable: Amount of Interactions
Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR 4.3077 3.6382 52
A3300 Moderately Positive ACHAR 4.2286 3.5817 35
Less Positive ACHAR 2.2857 2.2336 14
Less Positive Highly Positive ACHAR 4.9130 2.3532 23
A3303 Moderately Positive ACHAR 2.5217 2.4096 23
Less Positive ACHAR 2.3273 2.8025 55
Mean
Independent Variables df Square F p
Attitude toward the Program 2 12.971 1.332 .266
Attitude toward the Character 2 155.745 15.989 .000
Program Attitude x Character Attitude 4 28.505 2.926 .021
Error 280 .340

 

106

Figure 6.18.1. Interactions Effects: A3303 x Acrmr: Chandler

Less Positive
' A
,' PROG

   

Highly Positive
APROG .

 

 

Less Positive Moderately Highly Positive
ACH AR POSIIIVC ACHAR ACH AR

Attitudes toward Rachel and Ross were also found to have a significant
interaction effect. Specifically, the pattern of interaction effect of Rachel was identical to
that of Chandler. However, the pattern for Ross was found to be different. The amount of
interactions was found to be decreasing as ACHAR improves from less positive to
moderately positive. However, the main effects oprROG and ACHAR were found to be
non-significant in the case of Ross.

A post hoc test for Rachel’s case indicated that in both A3303 conditions,
participants with highly positive attitude toward Rachel (Acrwr) showed more
interactions than those with moderately positive and less positive ACHAR. For Ross, the
difference in the amount of interactions between highly positive ACHAR and less positive
ACHAR was found to be significant in a post hoc test. Tables 6.18.3 and 6.18.4 and Figures

6.18.2 and 6.18.3 show the corresponding results.

107

Table 6.18.3. Interactions Effects: A3303 x A3333: Rachel

Dependent Variable: Amount of Interactions

 

 

 

 

Std.

Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR 4.5690 3.5100 58
A3303 Moderately Positive ACHAR 4.2963 3.5281 27

Less Positive ACHAR 1.4375 2.1899 16
Less Positive Highly Positive ACHAR 5.6111 2.7470 18
A3303 Moderately Positive ACHAR 2.6250 2.6344 24

Less Positive ACHAR 2.2881 2.4357 59

Mean

Independent Variables df Square F P
Attitude toward the Program 2 2.250 .244 .784
Attitude toward the Character 2 236.809 25.684 .000
Program Attitude x Character Attitude 4 25.545 2.771 .028
Error 280 9.220

 

Figure 6.18.2. Interactions Effects: A3300 x AC3“: Rachel

 

Less Positive
" A
,' PROG

   

Highly Positive
APROG

 

Less Positive Moderately

ACHAR Positive ACHAR

108

Highly Positive
ACHAR

Table 6.18.4. Interactions Effects: A3303 x A0133: Ross

Dependent Variable: Amount of Interactions

 

 

 

 

 

Std.

Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR 4.0000 3.8957 52
A3303 Moderately Positive ACHAR 3.7333 2.9587 30

Less Positive ACHAR 4.4211 3.2543 19
Less Positive Highly Positive ACHAR 3.8261 2.6398 23
A3303 Moderately Positive ACHAR 2.4783 2.8102 23

Less Positive ACHAR 2.8000 2.8441 55

Mean
Independent Variables df Square F p
Attitude toward the Program 2 23.490 2.229 .110
Attitude toward the Character 2 18.833 1.787 .169
Program Attitude x Character Attitude 4 27.009 2.563 .039
Error 280 10.537
Figure 6.18.3. Interactions Effects: A3300 x AC3“: Ross
Highly Positive

APROG

 

 

Less Positive

ACHAR POSIIIVC ACHAR

Vess Positive

Moderately Highly Positive
ACHAR

Attitudes toward the remaining three characters (i.e., Joey, Monica, and Phoebe)

were also examined. However, no significant interaction effects were found in relation to

the attitude toward the program. Tables 6.18.5, 6.18.6, and 6.18.7 show the results.

109

Table 6.18.5. Interactions Effects: A3300 x AC3“: Joey

Dependent Variable: Amount of Interactions

 

 

 

 

 

 

 

 

 

Std.

IndQendent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR 4.0615 3.5350 65
A3303 Moderately Positive ACHAR 3.8929 3.3592 28

Less Positive ACHAR 3.8750 4.1209 8
Less Positive Highly Positive ACHAR 3.0000 2.5820 7
A3303 Moderately Positive ACHAR 2.6667 2.4976 15

Less Positive ACHAR 3.0127 2.9066 79

Mean
Independent Variables df Square F p
Attitude toward the Program 2 13.694 1.247 .289
Attitude toward the Character 2 3.693 .336 .715
Program Attitude x Character Attitude 4 2.348 .214 .931
Error 280 .340
Table 6.18.6. Interactions Effects: A3303 x A3333: Monica
Dependent Variable: Amount of Interactions
Std.

Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR 4.2453 3.7360 53
A3303 Moderately Positive ACHAR 4.0556 3.3205 36

Less Positive ACHAR 2.7500 2.8644 12
Less Positive Highly Positive ACHAR 4.0000 2.8868 19
A3303 Moderately Positive ACHAR 3.7059 3.2358 17

Less Positive ACHAR 2.4615 2.5744 65

Mean

Independent Variables df Square F p
Attitude toward the Program 2 2.497 .238 .788
Attitude toward the Character 2 71.220 6.788 .001
Program Attitude x Character Attitude 4 3.311 .316 .867
Error 280 10.491

 

110

Table 6.18.7. Interactions Effects: A3303 x Aw“: Phoebe

Dependent Variable: Amount of Interactions

 

 

 

 

Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACT-TAR 3.8246 3.2685 57
A3303 Moderately Positive ACHAR 4.6452 3.9034 31
Less Positive ACHAR 3.2308 3.2918 13
Less Positive Highly Positive ACHAR 3.6364 28026 11
A3303 Moderately Positive ACHAR 3.4000 2.8357 30
Less Positive ACHAR 2.6167 2.7867 60
Mean
Independent Variables df Square F p
Attitude toward the Program 2 8.597 .791 .455
Attitude toward the Character 2 16.765 1.542 .216
Program Attitude x Character Attitude 4 5.093 .468 .759
Error 280 10.874

 

Similarly, the interaction effects oprROG x ACHAR on the changes in attitude

toward paired brands were tested on each character. Earlier results from testing

hypotheses showed that attitude toward the program and toward five of six characters

were positively related with the changes in attitude toward paired brands (Hypotheses 7

and 10). To test the interaction effects, the mean scores of attitude toward the paired

brands were used for analysis.

Another series ofANOVA was performed. However, significant interaction effects

were not found from any of the six characters. Tables 6.18.8 to 6.18.13 show the results

from the tests.

111

Table 6.18.8. Interactions Effects: A3303 x A3333: Chandler

Dependent Variable: Changes in Attitude toward Brands Paired with Chandler

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1872 .4385 52
A3300 Moderately Positive ACHAR .3513 .3904 35
Less Positive ACHAR -.1323 .4676 14
Less Positive Highly Positive ACHAR .3427 .6055 23
A3303 Moderately Positive ACHAR .3070 .5283 23
Less Positive ACHAR .0347 .5067 55
Mean
Independent Variables df Square F p
Attitude toward the Program 2 .192 .950 .388
Attitude toward the Character 2 2.016 9.976 .000
Program Attitude x Character Attitude 4 .212 1.048 .383
Error 280 .202
Table 6.18.9. Interactions Effects: A3303 x Am“: Joey
Dependent Variable: Changes in Attitude toward Brands Paired with Joey
Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1936 .4586 65
A3300 Moderately Positive ACHAR .2288 .4630 28
Less Positive ACHAR .1481 .3564 8
Less Positive Highly Positive ACHAR .4921 .7631 7
A3303 Moderately Positive ACHAR .1510 .6084 15
Less Positive ACHAR .1411 .5154 79
Mean
Independent Variables df Square F p
Attitude toward the Program 2 .060 .280 .756
Attitude toward the Character 2 .529 2.471 .086
Program Attitude x Character Attitude 4 .185 .864 .486
Error 280 .214

 

112

Table 6.18.10. Interactions Effects: A3303 x A33“: Monica

Dependent Variable: Changes in Attitude toward Brands Paired with Monica

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1704 .4864 53
A3300 Moderately Positive ACHAR .3080 .3695 36
Less Positive ACHAR .0046 .4450 12
Less Positive Highly Positive ACHAR .1795 .6683 19
A3303 Moderately Positive ACHAR .4541 .5002 17
Less Positive ACHAR .0881 .5053 65
Mean
Independent Variables df Square F p
Attitude toward the Program 2 .120 .576 .563
Attitude toward the Character 2 1.404 6.735 .001
Program Attitude x Character Attitude 4 .103 .493 .741
Error 280 .208
Table 6.18.11. Interactions Effects: A3300 x Aw“: Phoebe
Dependent Variable: Changes in Attitude toward Brands Paired with Phoebe
Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1850 .4635 57
APROG Moderately POSIIIVC ACH AR .3286 .391 9 31
Less Positive ACHAR .0429 .4336 13
Less Positive Highly Positive ACHAR .0168 .5528 11
A3303 Moderately Positive ACHAR .2555 .7301 30
Less Positive ACHAR .1501 .4353 60
Mean
Independent Variables df Srprare F p
Attitude toward the Program 2 .137 .644 .526
Attitude toward the Character 2 .900 4.242 .015
Program Attitude x Character Attitude 4 .197 .926 .449

Error 280 .212

113

Table 6.18.12. Interactions Effects: APROG x Aqua: Rachel

Dependent Variable: Changes in Attitude toward Brands Paired with Rachel

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1702 .4481 58
Amos Moderately Positive ACHAR .3384 .4438 27
Less Positive ACHAR .0730 .4314 16
Less Positive Highly Positive ACHAR .3090 .5330 18
A9300 Moderately Positive ACHAR .4158 .5286 24
Less Positive ACHAR .0223 .5228 59
Mean
Independent Variables df Square F p
Attitude toward the Program 2 .059 .293 .746
Attitude toward the Character 2 1.598 7.879 .000
Program Attitude x Character Attitude 4 .194 .955 .433
Error 280 .203
Table 6.18.13. Interactions Effects: Apnoc x Acmut: Ross
Dependent Variable: Changes in Attitude toward Brands Paired with Ross
Std.
Independent Variables Mean Deviation N
Highly Positive Highly Positive ACHAR .1850 .4919 52
Apgog Moderately Positive ACHAR .2938 .3923 30
Less Positive ACHAR .0916 .4001 19
Less Positive Highly Positive ACHAR .1655 .6390 23
Apxoo Moderately Positive ACHAR .3820 .3758 23
Less Positive ACHAR .0775 .5544 55
T Mean
independent Variables df Square F p
Attitude toward the Program 2 .017 .082 .922
Attitude toward the Character 2 .865 4.113 .017
Program Attitude x Character Attitude 4 .170 .810 .520
_Error 280 .210

 

114

6.19. The Interaction Effects of Product Involvement and Attitude toward Characters on
Interactions and Attitude toward Brands (Study 5)

In testing the product involvement x attitude toward characters interaction effect,
paired brands were used in mixed analysis of variance (ANOVA). To do this, the amount
of interactions and attitude toward particular product categories had to be calculated
based on the product involvement. Six product categories were used for paired brands,
and they were clothes (12 brands), beverages (2 brands), inexpensive electronic goods
(telephone, 1 brand), tableware (1 brand), kitchenware (1 brand), and food (dessert, 1
brand).

Among these categories, clothes (M = 6.14) were used for high involvement
products, and kitchenware (M = 4.37), tableware (M = 4.18), and food (M = 3.85) were
used for low involvement products. Inexpensive electronic goods and beverages, which
were found to show moderate level of product involvement, were discarded from the
analysis. The high involvement product category had 12 brands, and the low involvement
category had three brands.

To ease the comparison, the amount of interactions per brand was calculated for
high and low involvement products. Finally, a series of mixed ANOVA was conducted to
test the product involvement x attitude toward characters interactions effect on the
amount of interactions. Tables 6.19.1 to 6.19.6 show the results for each character. No

interaction effects were found on the amount of interaction.

115

Table 6.19.1. Interactions Effects: Product Involvement x Aw“: Chandler

Dependent Variable: Amount of Interactions with Paired Items with Chandler

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .239 .201 97
Products Less Positive ACHAR .134 .169 93
Low Involvement Highly Positive ACHAR .247 .364 97
Products Less Positive Acmn .140 .227 93

Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 .01 .1 l .738
Attitude toward Character (Between-Subj ects) 1 1.08 13.33 .000
Product x Program Involvement 1 .00 .00 .958
Error 188 .05
Table 6.19.2. Interactions Effects: Product Involvement x Aann: Joey
Dependent Variable: Amount of Interactions with Paired Items with Joey

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .220 .205 104
Products Less Positive Acmuz .138 .171 106
Low Involvement Highly Positive ACHAR .221 .370 104
Products Less Positive ACHAR .223 .324 106

Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) l .20 3.59 .059
Attitude toward Character (Between-Subjects) 1 .16 1.61 .206
Product x Program Involvement 1 .18 3.33 .069
Error 208 .05

 

116

Table 6.19.3. Interactions Effects: Product Involvement x Am“: Monica

Dependent Variable: Amount of Interactions with Paired Items with Monica

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .231 .216 99
Products Less Positive ACHAR .119 .142 104
Low Involvement Highly Positive ACHAR .259 .377 99
Products Less Positive ACHAR .154 .245 104

Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 .10 2.06 .153
Attitude toward Character (Between-Subjects) 1 1.21 14.35 .000
Product x Program Involvement .00 .03 .865
Error 201 .05
Table 6.19.4. Interactions Effects: Product Involvement x Am“: Phoebe
Dependent Variable: Amount of Interactions with Paired Items with Phoebe

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .214 .187 97
Products Less Positive ACHAR .142 .180 102
Low Involvement Highly Positive ACHAR .216 .333 97
Products Less Positive ACHAR . 1 80 .3 10 102

Mean
Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 .04 .83 .362
_ Attitude toward Character (Between-Subjects) 1 .29 3.29 .071

Product x Program Involvement 1 .03 .63 .427
Error 197 .05

 

117

Table 6.19.5. Interactions Effects: Product Involvement x Aqua: Rachel

Dependent Variable: Amount of Interactions with Paired Items with Rachel

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .295 .201 105
Products Less Positive Acmm .097 .135 106
Low Involvement Highly Positive ACHAR .270 .379 105
Products Less Positive ACHAR .138 .260 106
Mean

Indegendent Variables df Square F p
Product Involvement (Within-Subjects) 1 .01 .14 .705
Attitude toward Character (Between-Subjects) 1 2.87 33.41 .000
Product x Program Involvement l .12 2.44 .120
Error 209 .05

 

Table 6.19.6. Interactions Effects: Product Involvement x Am“: Ross

Dependent Variable: Amount of Interactions with Paired Items with Ross

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .202 .184 106
Products Less Positive ACHAR .146 .184 103
Low Involvement Highly Positive ACHAR .217 .353 106
Products Less Positive ACHAR .191 .297 103
Mean

_Independent Variables df Square F p
Product Involvement (Within-Subjects) l .09 1.83 .178
Attitude toward Character (Between-Subjects) 1 .17 1.93 .166
Product x Program Involvement 1 .02 .45 .502

_Error 207 .05

118

Table 6.19.7. Interactions Effects: Product Involvement x Aw“: Chandler

Dependent Variable: Changes in Attitude toward Brands Paired with Chandler

 

 

 

 

 

 

 

 

 

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .294 .534 97
Products Less Positive ACHAR .040 .516 93
Low Involvement Highly Positive ACHAR -.073 .696 97
Products Less Positive ACHAR -.087 .917 93

Mean

Independent Variables df Square F J)
Product Involvement (Within-Subjects) 1 5.79 17.07 .000
Attitude toward Character (Between-Subjects) 1 1.70 2.85 .093
Product x Program Involvement 1 1.37 4.03 .046
Error 207 .34
Table 6.19.8. Interactions Effects: Product Involvement x Aqua: Ross
Dependent Variable: Changes in Attitude toward Brands Paired with Ross

Std.
Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .292 .553 106
Products Less Positive ACHAR .118 .559 103
Low Involvement Highly Positive ACHAR —.126 .634 106
Products Less Positive ACHAR -.038 .908 103

Mean

Independent Variables df Square F p
Product Involvement (Within-Subjects) 1 8.64 25.67 .000
Attitude toward Character (Between—Subj ects) 1 .19 .33 .565
Product x Program Involvement l 1.79 5.33 .022
Error 207 .34

 

119

Figure 6.19.1. Interactions Effects: Product Involvement x Am“: Chandler

 

 

 

High Inv.
Product
0
(No Changes in A3) 0 .................. ’ LOW Inv.
Product
Less Positive Highly Positive
ACHAR ACHAR

Figure 6.19.2. Interactions Effects: Product Involvement x Acnxn: Ross

 

 

 

High Inv.
/ PTOdUCt
0
(No Changes in A3) ’ ........
--------- Low Inv.
Product
Less Positive Highly Positive
ACHAR ACHAR

Next, another series of mixed ANOVA was conducted to test the product
involvement x attitude toward characters interactions effect on the changes in the attitude

toward paired brands. Attitudes toward only two characters (i.e., Chandler and Ross)

120

were found to have a significant interaction effect. Above Tables 6.19.7 and 6.19.8 and
Figures 6.19.1 and 6.19.2 show the results.

The above results indicate that the interaction effect was only marginally
significant with attitude toward Chandler (p = .046). Nevertheless, the main effect of
product involvement with both characters was found to be significant, and it showed that
the changes in attitude toward paired brands of low involvement products were negative
(which means the attitude declined by being exposed to the stimulus). The relationship
between product involvement and the changes in attitude toward brands was not
examined in hypotheses tests.

Furthermore, the main effects of attitude toward characters were not found to be
significant. Consequently, the above results would only be interpreted that for certain
characters the ACHAR had a stronger impact on the changes in A3 for high involvement
products than low involvement products. Supporting this interpretation, no significant
interaction effects were found in relation to the attitude toward the remaining four
characters (i.e., Joey, Monica, Phoebe, and Rachel). The results for the remaining four
characters are illustrated in Tables 6.19.9 to 6.19.12. Although significant interaction
effects were found from two characters (i.e., Chandler and Ross), it is difficult to
conclude that the significant results are truly meaningful because (1) the significant
relationships were found from only two characters among the six, (2) the significant
relationships were only marginally significant (p = .22; .46), and (3) the patterns of
significant interaction effects did not correspond to each other. Therefore, the significant
results from the Product Involvement x ACHAR interaction effects would be hard

generalize.

121

Table 6.19.9. Interactions Effects: Product Involvement x ACHAR: Joey

Dependent Variable: Changes in Attitude toward Brands Paired with Joey

 

 

 

 

 

 

 

 

 

Std.

Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .317 .538 104
Products Less Positive ACHAR .158 .538 106
Low Involvement Highly Positive ACHAR —.021 .748 106
Products Less Positive ACHAR —.010 .863 103

Mean
Independent Variables df Square F L
Product Involvement (Within-Subjects) 1 6.72 19.19 .000
Attitude toward Character (Between-Subjects) 1 .58 .99 .322
Product x Program Involvement 1 .76 2.16 .143
Error 208 .35
Table 6.19.10. Interactions Effects: Product Involvement x Am“: Monica
Dependent Variable: Changes in Attitude toward Brands Paired with Monica

Std.

Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .264 .539 99
Products Less Positive ACHAR .098 .524 104
Low Involvement Highly Positive ACHAR —.090 .808 99
Products Less Positive ACHAR —.066 .775 104

Mean
Independent Variables df Square F p
Product Involvement (Within-Subjects) l 6.79 20.59 .000
Attitude toward Character (Between-Subjects) 1 .51 .88 .350
Product x Program Involvement 1 .92 2.80 .096
Error 201 .33

 

122

Table 6.19.11. Interactions Effects: Product Involvement x Am“: Phoebe

Dependent Variable: Changes in Attitude toward Brands Paired with Phoebe

 

 

 

 

 

 

 

 

 

Std.

Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .259 .517 97
Products Less Positive ACHAR .154 .455 102
Low Involvement Highly Positive ACHAR —.l39 .642 97
Products Less Positive ACHAR —.098 .831 102

Mean
Independent Variables df Square F p
Product Involvement (Within-Subjects) l 10.49 33.41 .000
Attitude toward Character (Between-Subjects) l .10 .21 .644
Product x Program Involvement l 1.79 .53 .194
Error 1 97 .3 1
Table 6.19.12. Interactions Effects: Product Involvement x Amour: Rachel
Dependent Variable: Changes in Attitude toward Brands Paired with Rachel

Std.

Independent Variables Mean Deviation N
High Involvement Highly Positive ACHAR .310 .481 105
Products Less Positive ACHAR .061 .552 106
Low Involvement Highly Positive ACHAR -.072 .678 105
Products Less Positive ACHAR —.154 .755 106

Mean
Independent Variables df Square F P
Product Involvement (Within-Subjects) 1 9.36 30.75 .000
Attitude toward Character (Between-Subjects) l 2.88 6.00 .015
Product x Program Involvement .74 2.42 .121

Error 209 .30

 

123

So far, this chapter tested the hypotheses using different approaches. Some
relationships were examined on background products and some on paired brands. Some

excluded the responses from those who did not make any interactions and some included

all participants. Table 6.19.13 summarizes the approaches and methods used in the

analyses thus far, and presents the results.

Table 6.19.13. Summary of the Results

 

 

 

 

H I IV DV Items 2 Participants Method Result Table
1 Product Inv. Interactions Background All Regression Supported 6.4.2
2 Product Inv. Recall Background All Regression Supported 6.5.2
3 Program Inv. Interactions Background All Hotelling’s T2 Rejected 6.6
4 Program Inv. Recall Background All Hotelling’s T2 Rejected 6.6
5 APROG Recall Background All Regression Rejected 6.5.2
6 APROG Interactions Background All Regression Rejected 6.7
7 Amos Changes in AB All 3 All Regression Supported 6.8.3
8 ACHAR Interactions Paired Interacted Regression P. Supported 4 6.9.2
9 ACHAR Recall Paired 3 Interacted Regression Rejected 6.9.4
10 ACHAR Changes in AB Paired Interacted Regression P. Supported 4 6.10.2
l 1 Gender Interactions Background All T-test Rejected 6.1 I
12 Interactions Recall All All Regression Supported 6. 12
13 Interactions Product Inv. All All Regression P. Supported ‘ 6.13.2
14 Interactions Program Inv. All All Regression Supported 6.14.2
I Hypothesis number

2 Products or brands used for analysis
3 Only interacted brands were examined.
" Partially supported.

6.20. Test of the Hypothetical Model and Structural Relations

The overall relationships among the variables were tested by a structural equation

model (SEM). Bollen (1989) suggested that SEM provides a better description of the

structure of constructs because it simultaneously tests the relationships among the

constructs and it shows how existing relationships may change and additional

relationships may emerge. In testing the hypotheses, the manipulation of the participant’s

124

involvement with the program failed, and the variable was trichotomized. As a result,
only 289 out of 396 participants were available for subsequent analyses. However, for
SEM analysis those conditions were collapsed, and the level of involvement with the
program was used as a continuous variable so that all 396 participants would be available
for inclusion in the testing. The hypothesized relationships are presented in a model in

Figure 6.20.1.

Figure 6.20.1. Hypothesized Model

   
 

  

Product
Involvement 2

 
 
 
 
  

   
 

       
 
  

 
 

 

     
 
        
 
   

Interaction

] Plroduct t 1 Pro
nvo vemen Involvement 2

Program ................ 43. Brand
Involvement 1 Recall

Attitude:

Character Attitude:

Brand

 

——> Positive Relationship

  
 

Attitude:
Program 1

.................... ’ Negative Relationship

 

 

  

 

It should be noted that while testing hypotheses different criteria were employed

in choosing brands and products for dependent variables. That is, only background brands

125

were analyzed in testing certain hypotheses and paired brands were used for the others.
Sometimes, participants who did not make any interactions were removed from the
analysis, and sometimes only interacted brands were used. It is not feasible to conduct an
SEM analysis and examine the overall relationship of the variables when the variables
have different characteristics. Therefore, in an attempt to integrate these different aspects
of variable relationships, the SEM analysis focused on only one paired brand from each
character.

Each character was paired with three brands — two articles of clothing and one
other product. A brand (or an item) for analysis was selected so that the product category
of that item would not be paired with any other characters. In other words, the brand to
be used for SEM should be the only item in the product category. If another character is
paired with an item of the same category, it means the “Product Involvement Measure 1
4) Amount of Interactions —) Product Involvement Measure 2” relationship might be
breached. For example, if a participant did not interact with the item paired with the
character in question but instead interacted with the other item (under the same category)
paired with another character, (s)he might still demonstrate increased level of interactions
or product involvement, while this increase would not be explained by the model because
the participant’s amount of interaction would not represent the interaction made with the
other item.

Only three brands met the criterion described above (i.e., single item in a product
category), and they were kitchenware (paired with Monica), low-priced appliance (with
Joey), and dessert (with Phoebe). Consequently, three separate analyses were conducted.

The relationships that were rejected in the hypotheses test were not examined in the SEM

126

analysis. Multiple-group analysis could not be conducted because the participants saw all
three characters and brands (and thus cannot be divided into different groups). Figure

6.20.2 illustrates the tested model.

Figure 6.20.2. Tested Model

   
    

   

     
   
   
   
   
   
   
   

Product
Involvement 2

Attitude:
Character

 
 
    
 

     

 

  

Product Interaction
l I t 1 Program
nvo vemen ’ Involvement 2
Program ’ Brand
Involvement l , Recall

  

Attitude:
Character

Several differences are identified. First, all relationships are expected to be
positive based on the initial findings. Second, gender was removed from the model. Such
relationships as “Attitude toward the Program -) Recalls” and “Attitude toward the
Character -9 Recall” were also removed as they were rejected in hypothesis tests. On the
other hand, relationships between “Product Involvement —> Interactions and Recall,”
“Attitude toward the Character —) Interactions,” and “Interactions —> Product

Involvement” were examined in the model because they were only partially

supported/rejected from the previous tests. Finally, Attitude toward the Brand was

127

removed from the model. Testing the impacts of ACHAR and Apnoo on the changes in AB
employed responses only from interacted participants because it was not reasonable to
expect affect transfer to occur from those who could not see any brand information. The
proposed SEM analysis would employ both interacted and non-interacted participants,
and any changes in attitude toward the brand among those non-clicked participants would
not be explained by the model. Since it would pose a threat to the model’s overall validity,
the relationships regarding Attitude toward the Brand variable were not examined. As a
result, the tested model focused on the antecedents and consequences of audience

interactions with the ad.

6.21. SEM: Results

AMOS 4.0 was used to examine the model (Figure 6.20.2). This model did not
include the relationship between attitude toward the character and the amount of
interactions because the relationship was rejected from the earlier hypothesis test (see
Table 6.9.2).

Tables 6.21.1, 6.21.2, and 6.21.3 show the initial results from analyzing the model
for Joey’s low-priced appliance, Monica’s kitchenware, and Phoebe’s dessert,
respectively. Only a few number of coefficients were found to be statistically significant.
In particular, the model for Joey’s low-priced appliance showed no significant
relationships (Table 6.21.1), the model for Monica’s kitchenware showed only three
significant relationships (Table 6.21.2), and the model for Phoebe’s dessert showed only

two significant relationships (Table 6.21.3).

128

Furthermore, the goodness-of-fit indices indicated poor fit of the overall models
with the GFIs ranging from .735 to .809. The modification indices in AMOS were
examined to seek further improvements of the models. It was found that the suggested
changes fiom the modification indices were not consistent for the models for the three
characters. Therefore, it was concluded that there were no results that could be

generalized.

Table 6.21.1. Initial Result: InexpensiveAppliance / Joey

 

 

 

 

 

 

 

 

 

 

 

IVs DVs Unstd. S.E. Std. p
Coeff. Coeff.

Program Involvement 1 Interactions .007 .014 .026 .608
Product Involvement 1 Interactions .030 .017 .093 .079
Attitude-Program Interactions -.017 .020 -.044 .392
Interactions Program Involvement 2 .106 .080 .030 .185
Interactions Product Involvement 2 .016 .118 .005 .893
Good-of-fit Indices

x2 1511.303 (N=396, df=288)

p .000

Joreskog-Sorbom Goodness of Fit Index (GFI) .809

Adjusted Goodness of Fit Index (AGFI) .767

Bentler-Bonett Normed Fit Index (NF 1) .858

Tucker-Lewis Index (TLI) .867

Incremental Fit Index (IF I) .882

Comparative Fit Index (CF I) .882

Root Mean Square Error of Approximation (MSEA) .104

 

129

Table 6.21.2. Initial Result: Kitchenware / Monica

 

 

 

 

IVs DVs Unstd. S.E. Std. p
Coeff. Coeff.

Program Involvement 1 Interactions -.008 .015 -.026 .608
Product Involvement 1 Interactions .065 .020 .173 .002
Attitude-Program Interactions .041 .021 .098 .048
Attitude-Character Interactions .080 .024 .173 .001
Interactions Program Involvement 2 -. 122 .072 -.038 .092
Interactions Product Involvement 2 .211 .114 .075 .063
Good-of-fit Indices

x2 2123.949 (N=396, df=427)

p .000

Joreskog-Sorbom Goodness of Fit Index (GFI) .735

Adjusted Goodness of Fit Index (AGFI) .692

Bentler-Bonett Normed Fit Index (N F1) .823

Tucker-Lewis Index (TLI) .840

Incremental Fit Index (IF I) .854

Comparative Fit Index (CF I) .853

Root Mean Square Error of Approximation SRMSEA) .100

Table 6.21.3. Initial Result: Dessert / Phoebe

 

 

 

 

IVs DVs Unstd. S.E. Std. p
Coeff. Coeff.

Program Involvement 1 Interactions -.017 .018 -.047 .352
Product Involvement 1 Interactions .065 .019 .177 .001
Attitude-Program Interactions .017 .025 .034 .501
Attitude-Character Interactions .001 .027 .002 .964
Interactions Recalls .308 .019 .632 .000
Interactions Program Involvement 2 .021 .062 .008 .729
Interactions Product Involvement 2 .146 .090 .049 .106
Good-of-fit Indices

x2 2111.692 (N=396, df=457)

p .000

Joreskog—Sorbom Goodness of Fit Index (GFI) .742

Adjusted Goodness of Fit Index (AGFI) .702

Bentler-Bonett Normed Fit Index (N F1) .845

Tucker-Lewis Index (TLI) .863

Incremental Fit Index (IF I) .874

Comparative Fit Index (CPI) .874

Root Mean Square Error of Approximation SRMSEA) .092

130

Chapter 7. Conclusion & Discussion

7.1. The Role of Involvement in Advertising Exposure

The primary purpose of this research was to investigate the effect of involvement
on advertising effectiveness. In particular, consumers’ levels of involvement with the
program and the products in ads were examined in relation to their interactions with the
ads. The role of consumer involvement has been studied in a few studies on interactive
advertising (e.g., Cho & Leckenby, 1998; Cho & Leckenby, 1999; Ilfeld & Winer, 2002),
and they have found that consumers are more likely to interact with an ad that features a
product of high involvement. The current research has found a consistent result. Three
product categories (i.e., furniture, tableware, and home accessories) were examined, and
the participants who felt higher levels of involvement with a product were found to
demonstrate more interactions with the product’s ad.

The role of a different involvement type—involvement with the program—was
also examined. Previous studies have yielded conflicting results regarding the effect of
program involvement on advertising effectiveness. Some studies contended that the
cognitive capacity of audiences with higher levels of involvement with the program
would have fewer resources available to process the ad, and thus the involvement would
have a negative impact on advertising effectiveness (e.g., Pavelchak et al., 1988; Soldow
& Principe, 1981). On the other hand, others argued that higher levels of involvement
engender arousal, which would generate more acute processing of surrounding

information (e.g., Singh & Churchill, 1987; Srull, 1983).

131

The current research posited that, because the iPPLs in the interactive television
(ITV) interface appear during the program, higher levels of program involvement would
inhibit the program’s audience from interacting with the embedded ads. However, the
result (Table 6.6) showed that the amount of interactions were greater for those with
higher program involvement. Therefore, the aforementioned arousal approach (e. g.,
Singh & Churchill, 1987; Srull, 1983) was found to be more valid in the case of iPPL in
ITV. That is, when a consumer feels more involved with the program, (s)he would .
concentrate more on the program and experience higher levels of arousal. Explaining the
role of flow in online audience behavior, Novak et al. (2000) showed that increased
arousal generated heightened attention, which eventually led to the increased exploratory
behavior of the web users. In other words, the audience’s high program involvement
could produce more attention to the program and to its surrounding information, and it
might have generated more interactions.

This does not necessarily mean that the hypothesized negative relationship
(between program involvement and interactions) would always be invalid. The basics of
the “limited cognitive capacity” rationale employed in the Hypothesis 3 and Hypothesis 4
are believed to be valid. For example, when a consumer is extremely involved with the
program (i.e., at certain points during the program), the consumer would not be able to
divert his/her attention to surrounding information. But when the program does not
require the complete attention of the audience, the higher levels of the involvement with
the program could increase the audience’s levels of interest in the program’s adjacent
information as well, which will make the audience pay more attention to the information.

It should be noted that the current research used a situational comedy (or a sitcom) for the

132

experiment. Unlike some movies, television programs are not asking their audience to
pay full attention from the beginning to the end. There are commercial breaks, and
television programs are usually produced in a series.

This suggests several interesting implications. First, particular points or plots
during the program might show a positive or negative relationship between program
involvement and the amount of interactions. As described earlier, the audience might not
have enough cognitive capacity to pay attention to the program’s atmospheric stimuli.
The particular points of the program might be different times during the program (e. g.,
beginning, middle, end) or different points of the audience’s interest during the program
(e. g., scenes of vapid moment vs. tense moments). Second, it can be speculated that the
audience did not have to divide processing capacity in this research because the program
was paused while participants interacted with iPPLs. In other words, participants could
interact with the ads without missing much of the storyline. Finally, considering that a
movie is produced for a captive audience who is expected to stay in a theater and pay full
attention throughout its duration, the negative relationship between program involvement
and the interactions might be found better in movie reruns or movies made for television

compared to ordinary television programs.

7.2. The Role of Attitude in Advertising Effectiveness

This research examined two different types of attitudes—attitude toward the
program (ApRog) and attitude toward the character (ACHAR)—in relation to the amount of
interaction and brands recalled. Consumers’ positive attitude toward the program was

expected to decrease the amount of interactions because the iPPL would compete directly

133

with the program’s content for the audience’s attention. The rationale of this expectation
is as follows. Exposures to commercials in conventional television are rather mandatory
unlike the ads in print media. Previous studies (e.g., De Pelsmacker et al., 2002; Norris &
Colman, 1992) recognized that this skippable nature of advertising exposure (i.e., ad
exposure is not mandatory) reduces advertising effectiveness. Because the consumers
were exposed to the iPPLs placed in ITV on a voluntary basis, it was hypothesized that
the positive attitude toward the program would demonstrate similar effects as high
program involvement, and thus the positive attitude would negatively affect the amount
of interactions. The results found in the current research showed that the relationship was
in fact positive. That is, when the audience had a positive attitude toward the program,
the audience paid more attention to the program, which could have resulted in increased
interactions. Although the result was statistically significant, it should be noted that the
effect size was relatively small (Adjusted R2 = .027, Table 6.7), meaning that the
relationship is very weak at best.

On the contrary, the attitude toward the program was found to have a positive
relationship with the changes in attitude toward the brand, supporting the affect transfer
hypothesis (e.g., Mitchell & Olson, 1981; Shimp, 1981). The impact of the attitude
toward the program was found to be much stronger on the changes in attitude toward the
brand (R2 = .23 8) than on the amount of interactions.

Consumers’ attitude toward the character in the program was also hypothesized to
have a positive influence on the amount of interactions. This relationship was separately
examined on the six characters in the program, and significant relationships were found

in five characters (excluding Joey). However, it should be noted that even the remaining

134

one character showed marginally non-significant relationship (p = .052). The strength of
the significant relationship in the five characters was relatively weak, with the adjusted
R2 ranging from .014 to .218. However, when comparing the effect size of the attitude
toward the character on the amount of interactions to that of attitude toward the program
on the amount of interactions (Adjusted R2 = .027), it is found that the attitude toward the
character is more effective in increasing the audience’s interactions with the iPPLs.
However, it should be noted that the strength of both attitudes (toward the character and
the program) was weak.

Attitude toward the character was also found to have a positive relationship with
the changes in attitude toward the brand in most cases. Again, significant relationships
were found in only five characters (excluding Phoebe). The strength of the significant
relationship in the five characters varied but they were weak. The adjusted R2 ranged
from .023 to .084; see Table 6.10.2. From this, it can be found that the attitude toward the
program has a greater impact on the changes in the attitude toward the advertised brand
than does the attitude toward the character.

In interpreting the above results, a couple of issues have to be discussed. First, it
is not certain what characteristics in a character affect the relationship between the
attitude toward a character and the amount of interactions and the changes in attitude
toward a brand. As described earlier, the strength of the relationships varied and
significant associations were found from only five of the six characters. Because this
research employed an existing television program and well-known characters, individual
characteristics were not controlled for. This will be discussed again in the following

sections.

135

7.3. The Relationship between Attitude and Involvement

Basically, the level of involvement should not influence the valence of attitudes
(Andrews etal., 1990; Chattopadhyay & Nedungadi, 1990). However, involvement is
believed to affect or moderate the process by which an attitude is formed or changed
(Laczniak & Muehling, 1993). The current research found that these two seemingly
different variables, particularly involvement with the program and attitude toward the
program, had a similar positive impact on consumers’ interactions with embedded iPPLs.
Because product involvement also had a positive effect on the consumer’s interactions
with iPPLs, interaction effects among the three variables were suspected. However, no

significant relationships were found (see Table 6.15, 6.16.2, and 6.17.1).

7.4. The Role of Interactions in Program Consumption

A number of studies on advertising effectiveness in online media (e. g., Internet)
have focused on the importance of interactivity, which is characterized as an attribute of a
medium or consumers. They examined interactivity based on an assumption that the
audience favors interactivity and thus the interactive features in a medium are also
preferred by the audience. Accordingly, those previous studies have found that
interactivity generated positive outcomes in terms of advertising effectiveness. However,
the weakness in concentrating on interactivity is that interactive features or perceptions
can always change with the introduction of new technologies.

The “positive interactivity” assumption benefited from the advent of the Internet,

which is believed to be the first mass medium that embodied fiill interactivity. Although

136

the existence of interactive features would be helpful to the audience, it is questionable
whether the audience would always find those interactive features desirable. A few
studies have pointed out this problem (e.g., Lee & Lee, 1995). Currently, the online
audience is taking advantage of interactivity largely because the interactivity helps in
achieving their goals (e.g., finding information, etc.). But when the interactivity becomes
more common and typical features in other mass media, it is questionable whether the
audience would still prefer interactivity. For example, the novelty effect might wear off
and the audience’s perceptions of interactivity might change. When considering that
consumers’ perceptions of interactivity might become less favorable and that the
“positive interactivity” assumption might be violated, the positive role of interactivity
should be reconsidered.

In this regard, this dissertation proposed to use the audience’s interactions (as
opposed to the features or perceptions of interactivity) as a unit of analysis in examining
advertising effectiveness. In particular, the audience’s interactions with the embedded ads
represented a measure of advertising effectiveness in this dissertation. Moreover, the
interactions were used to describe a factor that might affect the audience’s media
consumption experience. Media consumption experience would include the level of
involvement the audience feels with the program or the product featured in the ads.
Conventional media consumption experience can be characterized as a passive
experience by a passive audience. Therefore, advertising effectiveness in mass media has
been commonly measured by memory, attitude, and purchase intention. However, actual

interactions made by the audience would be a more accurate measure in ITV environment.

137

Interactions would measure reach and frequency of the ad with greater accuracy; and, in
addition, they would represent voluntary exposure to the advertiser’s message.

This voluntary nature of interactions was expected to affect the audience’s
situational level of involvement with the program. The audience interaction has been
studied as a consequence. However, not many studies have examined the variable as an
antecedent, and it would be due to the fact that no mass media have possessed the
capability to allow the audience to interact with the ads and consume the content of the
medium simultaneously. In this sense, investigating the influence of interactions on
program-related variables would be unique and critical in the ITV environment because,
unlike conventional mass media, ITV is one of the very few media where audience can
interact with ads and consume the media content at the same time.

In testing hypotheses, it was found that the audience’s involvement could be
changed as a result of interactions. In particular, more interactions were found to increase
the audience’s situational involvement with the program (Table 6.14.2). It means that the
interactions made by the audience could actually increase his/her level of involvement
with the program, rather than interfering with the audience’s understanding of the
program and thus decreasing the involvement. However, three issues call for further
discussion. First, the nature of the program involvement affected by audience interactions
is only situational. The current study proposed that interactions would increase the
program involvement because the interactions would represent increased attention to the
communication process due to practicing active control. However, this increased
involvement level (with the program) is situational because the interactions with the

embedded ads would not be intrinsic to the program itself.

138

Second, although it is believed that the interaction would generally increase the
audience’s involvement, the impact might vary for different types of programs. A sitcom
was examined in this research. But it can be argued that certain programs (e.g., game
shows, educational programs) might ask the audience to invest more cognitive resources
than other programs (e. g., sitcoms, music videos) would do.

Finally, the cyclical process might be also considered. That is, this research has
found that the audience’s (intrinsic) program involvement increased the amount of
interactions, and that the interactions increased the audience’s (situational) program
involvement. Although the two types of involvement are different from each other (i.e.,
intrinsic vs. situational), an upward spiral process can be suspected within this
involvement-interactions-involvement relationship. If such a spiral process exists, it
would resemble online flow experience because flow experience is determined by the
audience’s level of control (Hoffman & Novak, 1996), and experienced by people who
are deeply involved (Lutz & Guiry, 1994) and focuses entirely on the interaction (Novak
et al., 2000). Flow in an online navigation has also been described to make the audience
intensively concentrated on the navigation activity, which would make the audience
highly involved with the navigation activity (Novak et al., 2000).

Unlike program involvement, consumers’ involvement with the product was not
found to be affected by their interactions with the ads. All eight product categories were
examined and significant relationships were found in only three categories (i.e., clothes,
furniture, and tableware). Although the relationships were positive, product involvement
was found to have decreased in most product categories, and the effect sizes were

relatively small. The adjusted R2 ranged from .011 to .023; see Table 6.13.2. In this

139

regard, it can be argued that product involvement is not affected by the amount of
audience interactions. This may be due to the fact that, compared to involvement with the
program, product involvement is closer to the intrinsic involvement that is not easily
changed by situational factors.

Finally, no significant difference in the amount of interactions was found between
female and male consumers in the current study. It should be noted that the large
standard deviation could be responsible in this non-significant result (see Table 6.11), and
that female participants actually demonstrated more interactions with the iPPLs than
males. This is contrary to what was expected in the Hypothesis 11. Based on previous
studies on gender differences in television watching behaviors, it was expected that males
would show more interactions than females. Those previous studies examined people’s
television watching behaviors in general, whereas a particular program was used in this
dissertation. Therefore, the findings in the current research can be interpreted that female
consumers can be reached as well as male consumers by using iPPLs. However, it should
be noted that certain factors regarding program types (e.g., sitcoms, soap operas, game
shows, etc.) or product categories employed for the iPPLs could have affected female

participants’ decision to interact with the ads.

7.5. Limitations and Suggestions for Future Studies

The current research has several limitations. First of all, it should be noted that the
research was conducted using a single laboratory experiment on college students.
Although there were initially 396 participants, only 289 were examined for most of the

analyses in this research and the number of participants was smaller in some analyses. In

140

order to provide more generalizable results to the advertisers and marketers, this research
needs to be replicated on a larger scale with a non-student sample.

Second, the interface of interactive television was presented to the participants on
a computer screen instead of a television. This constraint was unavoidable due to the
technological limitations. However, considering that operating a computer mouse while
watching a computer screen and operating a remote control while watching a television
may provide different experience to the audience, the external validity of this research
might be somewhat limited.

Third, the stimulus material in this research employed an existing television
program, which was already well known to the participants. This was due to the financial
constraints in producing a new TV show. Using an existing program caused several
problems. For example, its storyline could not be controlled, and thus the appearance
order of the iPPLs could not be controlled. The audience’s tendency to interact with a
certain iPPL might have been affected by the characteristics of the footage in which the
iPPL is embedded (e.g., IPPLs might draw less attention when placed in a tense moment).
This possible impact of could not be examined because of the lack of control over the
storyline.

Fourth, the overall involvement and attitude toward the program were very high
and favorable (Skewness: Program involvement = -.812; Attitude toward the program = -
1.559). This is also due to the use of an existing program that was already very popular.
In addition, participants’ attitude toward characters was also very favorable (Table 6.9.1).
However, the favorable attitude toward characters was predicted because the current

research had to select a program that featured well-known characters. Another problem

141

regarding the characters in the program was that there were too many characters
appearing in the program. This might have scattered the audience’s attention, which could
have reduced the number of brands recalled and the interactions.

Finally, the overall number of the recalled brands was rather small (M = 1.69).
This means that, when a single brand or a single product category was examined, most
participants could not recall any brands at all. The low amount of recalled brands may be
partly due to the fact that there were too many brands (i.e., n = 24) appearing in the
stimulus material. Moreover, the amount of interactions was relatively low. The mean
score for the amount of interaction was 4.04, which indicates that the participants viewed
an average of four brands during the program. The participants were specifically
instructed during the experiment to interact with the iPPLs only when they thought it was
necessary. Although this instruction was designed to maximize the reality and minimize
the novelty effect, it might have decreased the amount of interactions. On the other hand,
the low level of interactions might predict what would happen when ITV becomes a
reality. No matter what the cause of low level of interactions was, this low level of
interactions made it difficult to examine its relationship with other variables because most
participants reported no interactions when a single brand or product category was
examined. For example, when the amount of interactions with a kitchenware item was
examined (Table 6.22.1), 324 participants (81.8%) reported no interaction. The fact that
all structural equation models have failed to provide significant relationships among
variables might be explained by the low amounts of interactions and recalled brands.

Future studies on the advertising in interactive television should consider the

above issues. To lessen the above problems that arose from using an existing program

142

(i.e., failed manipulation on program involvement, positive attitudes, and the large
number of brands and characters in the program), future studies might use a short
segment of a program for an experiment, as opposed to using a whole episode.

F urtherrnore, other variables that were not employed in this research should be
considered. By employing various involvement and attitude variables, this research
attempted to predict the advertising effectiveness in the ITV environment. However,
some relationships tested in this research called for firrther investigation. For example,
the audience’s attitude toward a character was found to affect the amount of interactions
with the iPPLs paired with the character. Also, the audience’s attitude toward a character
was found to be related with their changes in the attitude toward brands featured in iPPLs.
But the results showed that the strength of the relationship varied in different characters.
Certain characteristics in a character (e. g., credibility, attractiveness, self-identification,
etc.) might be responsible, so further investigation is needed.

Also, the future studies on interactive advertising should focus more on the actual
interactions than the features or perceptions of interactivity. Unlike previous studies on
interactive advertising, this dissertation recognizes that interactivity might not always be
perceived as positive. Interactive features are described only as a device that would
increase the chances of interaction, not representing the true causes of interactions. In this
dissertation, people’s purposes in using a particular medium or content were posited as a
fundamental indicator of interactions, and (unlike interactivity) interactions were
portrayed as a relatively neutral construct. Therefore, future studies need to examine this

relationship in more details, and further indicators of interaction need to be explored.

I43

The difference between a character and an actor (or an actress) also needs to be
investigated. Many studies have examined the relationship between the source
characteristics and the advertising effectiveness (e.g., Goldsmith, Lafferty, & Newell,
2000; Gotlieb & Sarel, 1991; Ohanian, 1991). However, most of them focused on
consumers’ perceptions of the actor while not considering the story of the program or the
media vehicle in which the actor appears. Popular actors and actresses appear in a
number of different programs playing different characters. For example, Courteney Cox
(who stars in Friends as Monica) appeared in 30 movies and TV shows. Kevin Bacon
starred in 48 movies and TV programs (excluding talk shows). When celebrities endorse
a product fiom outside the program, consumers’ perceptions of the actor alone, not the
character, could explain the advertising effectiveness. But when the advertising occurs in
the program, as in the case of iPPL, the perceptions of the actor might be insufficient
because the perceptions of the actor would be affected by the characteristics of the
program and of the character the actor is playing. Therefore, in addition to the
perceptions of actors and actresses, characters need to be studies further.

Next, the variables used in this research are not exhaustive, and there might be
other variables that could complement the proposed model. For example, different types
of programs might affect the advertising effectiveness. As Lee and Lee (1995) projected,
the audience may not prefer having to interact with a television program. This tendency
to interact might vary in different programs. Therefore, this research should be extended
to test other program types than sitcoms. Considering that ITV will bring enormous
changes in the advertising environment, the importance of identifying and investigating

various factors that might affect advertising effectiveness in ITV cannot be overrated.

144

Finally, legal perspectives of iPPL must be investigated. Unlike product
placements (PPL) in movies, product placement in television is more severely regulated
by the Federal Communications Commission (F CC)’s sponsorship identification rules.
However, it should be noted that the nature of conventional PPL is forced exposure,
whereas the exposures to an iPPL is always voluntary. In this regard, application of
existing rules should be reevaluated, and studies on this regulatory aspect will greatly

contribute to the growth of iPPL practices.

7.6. Conclusions and Implications

A conventional television audience is passively exposed to commercials even
though the audience did not choose to be exposed. Therefore, possibilities always exist
that the audience is watching a commercial they do not like. This means the simple
exposure-based measures like reach and fi'equency are not correctly reflecting whether
consumers liked or disliked the ad. On the contrary, the advertising in ITV, especially
iPPL, delivers the advertising message only when the audience requests. Therefore, the
advertiser would become able to deliver more information without being intrusive or
annoying. Also, the voluntary nature of the advertising exposure is expected to increase
the advertising effectiveness among those exposed.

The current research examined factors that would affect the effectiveness of iPPLs
before ITV becomes widely diffused. For this study, the audience’s involvement with
product categories, involvement with the program, and attitudes toward the program and
its characters were examined. Consistent with previous studies on advertising and product

involvement, the current research has found that higher levels of product involvement

145

generated more interactions with the ad. Although product involvement was found to
increase the audience’s interactions with the embedded iPPLs, it does not mean that the
actual purchases will also rise. The findings from this study are strictly confined to the
audience’s attention and interactions with the brand. Higher levels of program
involvement were also found to increase the amount of interactions with the ad. It can be
interpreted that the arousal triggered by high levels of program involvement caused
heightened levels of the audience’s processing of the information in the program.
Consumers’ positive attitudes toward the program and a character were found to
be transferred to the attitude toward the brand. Particularly, the attitude toward the
program was found to have a stronger impact on the attitude toward the brand. It implies
that certain programs and characters will be preferred by advertisers. For advertisers,
these results provide an answer to the question as to which character and program they
should select for their brands. In particular, the results from testing Hypotheses 6 and 8
indicate that consumers’ attitudes toward the character were more effective in generating
their interactions with the paired iPPLs. On the other hand, the results from the
Hypotheses 7 and 10 show that consumers’ attitudes toward the program were more
effective in improving their attitudes toward the brands featured in the iPPLs. These
results provide valuable tips to advertisers. That is, when an advertiser’s primary purpose
is to generate consumers’ interactions with the ad or to increase their awareness with the
brand, the advertiser should pair the ad with favorable characters. On the contrary, when
the advertiser’s primary goal is to improve consumers’ attitudes toward the brand,
consumers’ attitudes toward the particular program should be considered rather than the

attitude toward particular characters in the program.

146

Although this result might be helpful for advertisers in selecting particular media
vehicles for their ads, it might pose a threat to the producers and stations of certain types
of programs. For example, advertisers’ preference for certain programs or characters
over others might expand the gap between popular programs and unpopular programs,
especially considering that the iPPL has a capability to generate direct sales.

Finally, audience members who interacted with iPPLs more actively showed a
bigger increase in their levels of involvement with the program than those who
demonstrated lesser amounts of interactions. Considering that program involvement
represents more intensive consumption of the program, this result implies that the
interactions within an ITV program can positively affect the audience’s evaluations and
consumption patterns of the program. Interactive television is expected to bring about a
number of changes to the current media industry including the relationship between
advertisers and producers. For instance, advertisers’ influence on the program might
increase because advertisers will not only prefer particular programs for their ads, but
also they will want to place their ads in their preferred places paired with preferred
characters. Consequently, the program formats are anticipated to endure dramatic changes
because the stations and the producers might want to secure as many opportunities for
iPPLs to host as many advertisers as possible.

This dissertation provides an idea as to how the iPPL in ITV, unlike the passive
commercials in conventional television environment, can change the audience’s
consumption patterns of programs and ads. When color televisions were first introduced

to the public, the industry had to experience a huge revolution in terms of its program

147

planning, make-up, lighting, and so on. When the iPPL becomes widely accepted, not
only the program content but also the diversity of programs could be affected.
Specifically, this dissertation describes that the iPPL can deliver the advertising
information only when it is wanted by the audience. Therefore, ITV and the iPPL provide
benefits to both audiences and advertisers. Audiences will benefit because they will not
be interrupted by unwanted and possibly intrusive commercial messages while they
watch a program, and yet they will be able to examine the information of the products of
their interest. Advertisers will benefit because the audience’s voluntary exposures to the
ads will promise a means to deliver their ads more efficiently and generate direct sales.
IPPLs will also provide media planners with more opportunities to improve their media
mix. For example, media planners would be able to choose particular programs or
characters based on whether the purpose of the advertising is to generate sales or to
increase interactions. In summary, advertisers, producers, and stations all need to prepare

from the various perspectives for the changes ITV and iPPL might bring about.

148

APPENDIX A. QUESTIONNAIRE FOR THE EXPERIMENT

Pretest for Attitude toward Brands

Following questions ask your feeling about certain brands.

 

 

 

 

 

 

 

 

 

 

 

1.]. Crew: J.CREW
bad good
pleasant unpleasant
unfavorable favorable
I.2. Gap: “1'
bad good
pleasant unpleasant
unfavorable favorable
1.3. Eddie Bauer: agrees/away
Bad good
Pleasant unpleasant
Unfavorable favorable
1.4. Calvin Klein: .5 ”A, y m 1
bad good
pleasant unpleasant
unfavorable favorable

1.5. Abercrombie & Fitch:

 

 

 

 

 

 

 

\1..-a....~..1.ie .\ I ”.11
bad good
pleasant unpleasant
unfavorable favorable
1.6. Polo Ralph Lauren: .3“,
Bad . good ...........
Pleasant unpleasant
Unfavorable favorable

 

1.7. Banana Republic:

 

 

 

 

 

   

{swam RIM 11.31-“
bad good
pleasant unpleasant
unfavorable favorable
1.8. American Eagle Outfitters:
bad ' ° ' ' good
pleasant unpleasant
unfavorable favorable

 

1.9. Arizona Jeans:
bad

 

pleasant

 

unfavorable

 

1.10. Tommy Bahama (Men’s Clothing):

Bad :
pleasant

 

unfavorable

 

1.11. Limited:
bad

 

pleasant

 

unfavorable

 

1.12. DKNY (Women’s Clothing):
bad ° :

 

pleasant

 

unfavorable

 

1.13. Sara Lee (Frozen Dessert):
bad ° °

 

pleasant

 

unfavorable

 

150

good

unpleasant
favorable

 

unpleasant
favorable

Limited
good
unpleasant

favorable

DKNY
good

unpleasant
favorable

good
unpleasant
favorable

1 .14. Impulse (Beverage):

bad
pleasant
unfavorable

1.15. Starbucks:
bad
pleasant
unfavorable

1.16. IKEA (Furniture):

bad
pleasant
unfavorable

1.17. Art Van Furniture:

bad
pleasant
unfavorable

1 .18. La-Z—Boy (Recliners):

bad
pleasant
unfavorable

1.19. Target (for Kitchenware):

bad
pleasant
unfavorable

1.20. WalMart (for Home Acce

Bad
Pleasant
Unfavorable

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ssories. e.g., message boards, CD

 

 

 

151

W
good
unpleasant
favorable
good
unpleasant
favorable
@
good
unpleasant
favorable
M
good
unpleasant
favorable

LA‘QBOY

good
unpleasant
favorable

Gunner
good

unpleasant
favorable

rmks): WW
good

unpleasant
favorable

1.21. OfficeMax (for Home Accessories. e.g., message boards, CD racks):

bad
pleasant
unfavorable

1.22. Pottery Barn (Kitchenware, Tableware, etc.):

Bad
Pleasant
Unfavorable

1.23. Crate & Barrel (Kitchenware, Tableware, etc.):

Bad
pleasant
unfavorable

1.24. Panasonic:
bad
pleasant
unfavorable

 

 

 

 

 

 

 

 

 

 

 

 

good
unpleasant
favorable

good

unpleasant
favorable

d
goo

unpleasant
favorable

W
good

unpleasant
favorable

 

Thank you. You will have to fill out another set of questionnaire when you

come to the main experiment. Please mark your calendar, and arrive

152

five minutes prior to the experiment.

Posttest for Other Measures

Thank you for participating in the experiment. The purpose of this study is to
' examine audiences’ responses to “Interactive TV.” Please read carefully before you
start the questionnaire.

Based on your own perception, make each item a separate and independent
judgment.

Work at fairly Q'gh speed through this questionnaire. Do not worry or puzzle over
individual items. It is your first impressions, the immediate feelings about the items
that we want.

On the other hand, please do not be careless, because we want your true impressions.

Here is how you are to use these scales.

 

Place your check mark or circle according to how closely your perception is related
to one or the other end of the scale.

 

CD Question Example 1

How important are your parents in choosing a computer?

Unimportant : ‘/ : : : : : Important

 

(1) Question Example 2

 

Strongly Strongly
Disagree Agree
I am familiar with the Intel commercial. I 2 3 4 5 6 0

 

 

Important

 

Please do not skip questions. Also, never put more than one check mark or circle
on a single scale

153

2.1. To me Tableware is:

important
boring
relevant
exciting
means nothing

2.2. To me Clothing is:

important
boring
relevant
exciting
means nothing

2.3. To me Beverages are:

important
boring
relevant
exciting
means nothing

2.4. To me Home Accessories are:

important
boring
relevant
exciting
means nothing

(e.g., plates, bowl, etc.)

 

 

 

 

 

(e.g., jackets, pants, etc.)

 

 

 

 

 

(e.g., soda, juice, etc.)

 

 

 

 

 

 

 

 

 

 

2.5. To me Telephones (not mobile phones) are:

important
boring
relevant
exciting
Means nothing

 

 

 

 

 

154

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

(e.g., CD racks, message boards, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

2.6. To me Frozen Dessert Products are:

important

 

boring

 

relevant

 

exciting

 

means nothing

 

2.7. To me Furniture is:
important

(e.g., dining table, sofa, etc.)

 

boring

 

relevant

 

exciting

 

means nothing

 

2.8. To me Kitchen Gadget Products are:

important

 

boring

 

relevant

 

exciting

 

means nothing

 

2.9. To me “Friends” (NEC sitcom) is:

important

 

boring

 

relevant

 

exciting

 

means nothing

 

bad

 

 

pleasant
unfavorable

 

appealing

 

unattractive

 

(e.g., frozen pies, cakes, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

(e.g., coffee mugs, plates, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

good
unpleasant
favorable
unappealing
attractive

 

 
 

This is the end of the Ste .
Do NOT turn the page over.

Now, please follow the instructions on your screen.

 

155

Please fill out the rest of the questionnaire

 

ONLY WHEN the program is over.

The following items are about your feeling while watching the show.
Please describe your feeling by placing check marks on the scales given below.

3. Thinking when I was watching the show, Friends was:

 

 

 

 

 

 

 

 

 

important : : : : : : unimportant
boring : : : : : : interesting
relevant : : : : : : irrelevant
exciting : : : : : : unexciting
means nothing : : : : : : means a lot to me
bad : : : : : : good
pleasant : : : : : : unpleasant
unfavorable : : : : : : favorable
appealing : : : : : : unappealing
unattractive : : : : : : attractive

 

156

F ollowin uestions ask about our feelin about each character in Friends.

 

7.]. Chandler is:

 

Matt LeBlanc

7.3. Monica is :

 

 

 

Courteney Cox Arquettc

7.4. Phoebe is:

 

 

 

 

 

Lisa Kudrow

7.5. Rachel is:

 

 

Jennifer Aniston

bad
pleasant
unfavorable
appealing
unattractive

bad
pleasant
unfavorable
appealing
unattractive

bad
pleasant
unfavorable
appealing
unattractive

bad
pleasant
unfavorable
appealing
unattractive

bad
pleasant
unfavorable
appealing
unattractive

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

157

good
unpleasant
favorable
unappealing
attractive

good
unpleasant
favorable
unappealing
attractive

good
unpleasant
favorable
unappealing
attractive

good
unpleasant
favorable
unappealing
attractive

good
unpleasant
favorable
unappealing
attractive

7.6. Ross is:

 

 

  

 

 

 

bad : : : : : : good
pleasant : : : : : : unpleasant
unfavorable : : : : : : favorable
‘ - appealing : : : : : : unappealing
13"“ semimm" unattractive : : : : : : attractive

 

8. Please list all brands you remember seeing during the show (Friends).

(Note: Please write down brands. not products. For example, Chevrolet, IBM, and Timex
are brands. Cars, computers, and watches are NOT.)

 

 

 

 

 

 

 

 

 

 

 

 

This is the end of the Step 2.

Before proceeding to the next page,

3T0 P please take off the first 6 pages of
this questionnaire and submit them
to the researcher.

 

158

Now, think when you were watching the program in the experiment.
Please indicate the degree you agree or disagree to the following statements.

10.1. To me Tableware is:

important
boring
relevant
exciting
means nothing

10.2. To me Clothing rs:

important
boring
relevant
exciting
means nothing

10.3. To me Beverages are:

important
boring
relevant
exciting
means nothing

10.4. To me Home Accessories are:

important
boring
relevant
exciting
means nothing

10.5. To me Telephones (not mobile phones) are:

important
boring
relevant
exciting
Means nothing

(e.g., plates, bowl, etc.)

 

 

 

 

 

(e.g., jackets, pants, etc.)

 

 

 

 

 

(e.g., soda, juice, etc.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

159

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

(e. g., CD racks, message boards, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

unimportant
interesting
irrelevant
unexciting
means a lot to me

10.6. To me Frozen Dessert Products are:

important
boring
relevant
exciting
means nothing

10.7. To me Furniture is:

important
boring
relevant
exciting
means nothing

10.8. To me Kitchen Gadget Products are:

important
boring
relevant
exciting
means nothing

 

 

 

 

 

(e.g., dining table, sofa, etc.

 

 

 

 

 

 

 

 

 

 

(e.g., fi'ozen pies, cakes, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

)

unimportant
interesting
irrelevant
unexciting
means a lot to me

(e.g., coffee mugs, plates, etc.)

unimportant
interesting
irrelevant
unexciting
means a lot to me

 

Following questions ask your feeling about certain brands.

I l .1 . JCrew:
bad
pleasant
unfavorable

11.2. Gap:
bad
pleasant
unfavorable

 

 

 

 

 

 

160

jCREW’
good
unpleasant
favorable

good
unpleasant
favorable

11.3. Eddie Bauer:
Bad

 

Pleasant

 

Unfavorable

 

11.4. Calvin Klein:
bad

 

pleasant

 

unfavorable

 

11.5. Abercrombie & Fitch:
bad '

 

pleasant

 

unfavorable

 

11.6. Polo Ralph Lauren:
Bad '

 

Pleasant

 

Unfavorable

 

11.7. Banana Republic:
bad °

 

pleasant

 

unfavorable

 

11.8. American Eagle Outfitters:

bad

\\11‘1\'1‘ \\1“\(.11(11 11‘11111{.\

 

pleasant

 

unfavorable

 

l 1.9. Arizona Jeans:
bad

 

pleasant

 

unfavorable

 

161

, War-tr
good

unpleasant
favorable

“'1! 1’ ---.,
good

unpleasant
favorable

\11111 mmlm 111-. 11

good

 

unpleasant
favorable

OOOOOOOOOOO

good
unpleasant
favorable

 

15,\\1\..\=\ m H 131,19;
good
unpleasant

favorable

 

good
unpleasant
favorable

ARIZONA

unpleasant

good

favorable

11.10. Tommy Bahama (Men’s Clothing):

Bad :

 

 

pleasant
unfavorable

 

11.11. Limited:
bad

 

pleasant

 

unfavorable

 

11.12. DKNY (Women’s Clothing):
bad : : :

 

pleasant

 

unfavorable

 

11.13. Sara Lee (Frozen Dessert):
bad : '

 

pleasant

 

unfavorable

 

1 1.14. Impulse (Beverage):
bad :

 

pleasant

 

unfavorable

 

1 1.15. Starbucks:
bad

 

pleasant

 

unfavorable

 

11.16. IKEA (Furniture):
bad '

 

pleasant

 

unfavorable

 

162

[19001019 ago/tam I
good
unpleasant

 

favorable

Limited
good
unpleasant
favorable

DKNY
good

unpleasant
favorable

Jame»
good

unpleasant
favorable

Inc-0v M

good
unpleasant
favorable

good
unpleasant
favorable

good
unpleasant
favorable

11.17. Art Van Furniture:
bad

 

pleasant

 

unfavorable

 

1 1.18. La-Z-Boy (Recliners):
bad °

 

pleasant

 

unfavorable

 

11.19. Target (for Kitchenware):
bad °

 

pleasant

 

unfavorable

 

[H
Van
good

unpleasant

favorable

LAaaov
good

unpleasant
favorable

Gunner
good

unpleasant
favorable

11..20 WalMart (for Home Accessories. e. g, message boards, CD racks): mam-r

Bad

 

Pleasant

 

Unfavorable

 

good
unpleasant
favorable

11 .21. OfficeMax (for Home Accessories. e. g., message boards, CD racks): @

bad

 

pleasant

 

unfavorable

 

11.22. Pottery Barn (Kitchenware, Tableware, etc.):
Bad . . . . .

 

Pleasant

 

Unfavorable

 

11.23. Crate & Barrel (Kitchenware, Tableware, etc.):

Bad

 

pleasant

 

unfavorable

 

163

good
unpleasant
favorable

1'1111111311111\

good
unpleasant
favorable

good

unpleasant
favorable

1 1.24. Panasonic: M

 

 

bad : : : : : : good
pleasant : : : : : : unpleasant
unfavorable : : : : : : favorable

 

 

Finally, we would like to get some information about you. Please mark the
appropriate answer.

1. What is your gender? Male Female

2. What is your age? (Write)

 

3. What is your grade level?

Freshman Sophomore Junior

Senior Graduate

4. What is your ethnic background?
__ American Indian or Alaskan Native
_ Asian, Asian American, or Pacific Islander
_ Black or African American
_ Mexican, Mexican American or Chicano
__ White (non-Hispanic)
__ Other, please specify:

 

This is the end of this questionnaire.
Thank you for your time and cooperation!
Please leave this questionnaire in front of the monitor.

164

APPENDIX B. INSTRCTION FOR THE EXPERIMENT

Instruction for Participants in All Program Involvement Conditions (High and Low)

 

Instructions for Practice Stage: Page 1 of 6

 

The following paragraphs contain very important instructions
regarding this experiment. Please read this information carefully.

 

Shortly, you will be watching a 3-minute TV program, which has been prepared to
make you familiar with the interactive television (ITV) interface and to let you practice
its features using a computer.

This ITV interface is not yet commercially available in the U.S., although it is
available in many European countries. The main idea of the ITV is to allow the
television audience to interact with the program contents.

In the following practice stage of the experiment, you will be watching a 3-minute
segment of a popular sitcom - Friends. There is some product information
embedded in the program. The idea is that audiences will be able to get additional
information about the products appearing in the program, or even purchase them,
through the simple operation of their remote control.

This function has been activated in the ITV interface in this experiment. Therefore,
during both the practice session and the main experiment, you will be able to access
product information by clicking your mouse instead of using a remote control.

The next pages will show you the interface in more details.

GO TO THE NEXT PAGE

   

165

Instructions for Practice Stage: Page 2 of 6

Although you will be watching a program on this computer screen instead of a
television set, the program will look exactly the same as it would in an ordinary
situation, except for one thing.

In the bottom-right corner of the screen, you will sometimes find a rotating blue icon
“®_"

 

<Figure 1> Icon in the screen

at. \
Icon is rotating

When the icon is on the screen, it means that the icon is clickable and that there is
product information available for you to see by clicking the icon.

GO TO THE NEXT PAGE

166

Instructions for Practice Stage: Page 3 of 6

When you watch the show, assume that you have noticed a product that interests
you (e.g., clothes worn by a character), and that there is the icon on the screen. If
you click the icon, the program will pause, and the list of available products will pop
up (Figure 2). If the product that interested you is not included in the list, you can
close the list simply by clicking the Close (Down-Arrow) button. The program will

resume.

—>
H 1 ~ £41....

'“
1e; I
fail
r. r~
11“ ‘-

4—j‘ . .........
‘ ‘ i n 'w. um

y \
\
I '1‘ NBC-(5'7-

<Figure 2> Product List

GO TO THE NEXT PAGE

hoods

 

167

Instructions for Practice Stage: Page 4 of 6

If the product that interested you is included in the list, click on the product name. A
small box that contains detailed product information will pop up. This is called a
“product information box." (Figure 3)

- ~91 -
: , it ‘3
mil
.1 ‘ ,1
- r

1‘

More Information Button .
a that!“ M1. CNI' Mal

_ In Ion-ls 2 no“ rut-tur-

Product
Details
turn "-1 portion

Jx '
. - .- :' ' — . cur-m - y
.4 .,4- :4, .. .4- . ’4 if. J44 71:24 ":4 ,
4 {Ire—nee Closrng Button
Prevrous Menu Button I ‘

q
a NBCIOH

   

<Figure 3> Product Information Box

You will notice that the product information box has several buttons such as [Save
for Later], [More Product Information], [Buy], [Previous Menu], and [X] (Close).
Because this is an experiment, the [Save for Later] and [More Product Information]
buttons have been deactivated. That is, these two buttons CANNOT be clicked.

GO TO THE NEXT PAGE

168

Instructions for Practice Stage: Page 5 of 6

Unlike the [Save for Later] and [More Product Information] buttons, the [Buy] button
is clickable. However, the [Buy] button is only partially working. Because this is a
test, clicking this [Buy] button does not mean that you are actually purchasing the
product or paying for the product. However, we want you to click the [Buy] button fl

ou feel that ou mi ht want to urchase the roduct if the interface were real.
(Figure 4)

 

The [Previous Menu] button will close the product information box and re-open the
list of the available products. And finally, the [X] (Close) button will close the product
information box and resume the program so you can continue watching it.

\

 

,1 .4
'GbiuNBCxort:

<Figure 4> Product lnforrnation Box: Disabled Button(s)

GO TO THE NEXT PAGE

169

Instructions for Practice Stage: Page 6 of 6

Now, you are ready to proceed to the 3-minute practice session. Please put your
headset on if you haven’t already done so. You don’t have to use the keyboard. The
mouse will be your input device.

During the practice session, please interact with the products as much as you want
because the purpose of this session is to let you get used to the interface.

If you have any questions about the interface, please raise your hand now and let
the research administrator know. OthenNise, you may click the button below and

start the practice.
BEGIN THE PRACTICE

170

Instruction for Participants in the High Program Involvement Condition
(After Practice)

 

Instructions for Main Expgriment: Page 1 of 4

We hope you enjoyed the practice session.

Now, you are about to begin the main experiment. Before we let you start, we’d like
to tell you a few more things. First of all, this main experiment is different from the
practice session in several ways.

(1) The length of the program is approximately 25 minutes.

(2) Product information will not be available all of the time. You will find
information only when the icon “CD" is rotating on the screen.

(3) _Y_gr_1r movements on the computer screen willpe recordeg by computer
software.

You are not being recorded by a camcorder. Only your on-screen activities
will be recorded; for example, clicking, opening/closing windows, and moving
the mouse pointer.

It is extremely important that you act as if you were watching a real ITV
program in the real world. In other words, we ask you to not can every

single product embedded in the prggram. Please open the product
information only when you think the product interests you.

 

GO TO THE NEXT PAGE

171

Instructions for Main Exmriment: Page 2 of 4

The program you will be watching for the main experiment is a different episode of
Friends. As mentioned earlier, you will watch the whole 25 minutes of the show.
The episode is #408, “The One With Chandler in a Box.” In case you haven’t
watched the show for a while, or for those who do not know about this program very
well, here is the synopsis of this episode.

 

 

In previous episodes, Chandler became attracted to Kathy (guest star Paget
Brewster), who was then dating Joey. Despite the guilt, Chandler’s feelings
got deeper, and so did Kathy's feelings. One day, Chandler kissed her, and
she dumped Joey. Chandler confessed, but Joey felt betrayed.

In this episode, Joey is still upset at Chandler about Kathy. Joey tries to get
over it but can't, and decides to move out... until he and Chandler stumble
upon a way for Chandler to make it up to him—spending most of
Thanksgiving day in a box.

Monica injures her eye but doesn't want to have to see Richard (former
boyfriend) again; she arranges to see the on-call doctor, who turns out to be
very cute... and Richard's son. She invites him over to their Thanksgiving
dinner.

The gang decides to do secret Santa for each other; Ross torments Rachel
about always exchanging gifts, until she can't stand it and shows him all the
stuff she saved from their relationship.

 

GO TO THE NEXT PAGE

172

 

Instruction for Main Exgriment: Page 3 of 4

Beginning in March 2003, AT&T Broadband and Cox Communications will implement
this Interactive Television interface in a nationwide test market using their digital
cable service. Several cities in California, Georgia, Michigan, Florida, and
Connecticut have been chosen for the testing, and the Greater Lansing area is
included. A set-top box for this application will be rented at a low price.

The purpose of this experiment is to see how audiences react to the interface.
Therefore, when the show ends, you will be asked to evaluate the overall episode.
You will also be asked about some of the details of the episode. Answers to the
questions are all included in the show. If you pay enough attention to the program,
you will be able to answer the questions.

We encourage you to pay attention to the program. Two participants who provide
the highest number of correct answers will be rewarded with a cash prize.

GO TO THE NEXT PAGE

 

173

Instructions for Main Expgriment: Page 4 of 4

Now, you are about to begin the main experiment.

Again, we ask you not to view every product embedded in the program during the
experiment. Just act like it is a real situation - open the product information only

when you find the product interesting, and click the [BUY] button only when you think

you might actually buy the product if it were a real-world situation.

Once the experiment begins, you are not allowed to ask any questions or talk to anybody.
So, if you have any questions concerning this experiment, please raise your hand now
and let the research administrator know.

Otherwise, click the below button, start the experiment, and have fun.

BEGIN THE EXPERIMENT

174

Instruction for Participants in the Low Program Involvement Condition
(After Practice)

 

Instructions for Main Exppriment: Page 1 of 2

We hope you enjoyed the practice session.

Before we let you start the main experiment, we’d like to tell you a few more things.
First of all, this main experiment is different from the practice session in several ways.

(1) The length of the program is approximately 25 minutes.
(2) Product information will not be available all of the time. You will find
information only when the icon “( i )” is rotating on the screen.

(3) Your movements on the computer screen will be recorded by computer
software.

You are not being recorded by a camcorder. Only your on-screen activities
will be recorded; for example, clicking, opening/closing windows, and moving
the mouse pointer.

It is extremely important that you act as if you were watching a real ITV
program in the real world. In other words, we ask you to not ogn even

single product embedded in the prpgram. Please open the product
information only when you think the product interests you.

GO TO THE NEXT PAGE

 

175

Instructions for Main Experiment: Page 2 of 2

Now, you are about to begin the main experiment.

The program you will be watching for the main experiment is a different episode of
Friends. As mentioned earlier, you will watch the whole 25 minutes of the show.
The episode is #408, “The One With Chandler in a Box.”

Again, we ask you not to view every product embedded in the program during the
experiment. Just act like it is a real situation - open the product information only
when you find the product interesting, and click the [BUY] button only when you think
you might actually buy the product if it were a real-world situation.

Once the experiment begins, you are not allowed to ask any questions or talk to anybody.
So, if you have any questions concerning this experiment, please raise your hand now
and let the research administrator know.

Otherwise, click the below button, start the experiment, and have fun.

BEGIN THE EXPERIMENT

 

176

APPENDIX C. DESCRIPTIVE STATISTICS

Composition of the Samples as a Whole

 

 

 

 

 

N 396
Age Mean 20.61
Std. Deviation 2.27
Median 20.00
Gender Male 134 (33.8%)
(%) Female 262 (66.2%)
Ethnic Background (%)
American Indian or Alaskan Native 2 (.5%)
Asian, Asian American, or Pacific Islander 54 (13.6%)
Black or African American 28 (7.1%)
Mexican, Mexican American or Chicano 14 (3.5%)
White (non-Hispanic) 288 (2.8%)
Other 10 (2.5%)
Class Level (%)
Freshman 66 (1 6.7%)
Sophomore 89 (22.5%)
Junior 147 (37.1%)
Senior 93 (23.5%)

 

177

Sample Composition by Conditions: High Program Involvement Condition

 

 

 

Conditions: Program Involvement High Low

N (%) 289 149 140

Age Mean 19.99 21.14
Std. Deviation 1.46 2.64

 

Gender Male
(%) Female

22 (14.8%)
127 (85.2%)

75 (53.6%)
65 (46.4%)

 

Ethnic Background (%)
American Indian or Alaskan Native
Asian, Asian American, or Pacific
Islander
Black or African American
Mexican, Mexican American or Chicano
White (non-Hispanic)
Other

1 ( .7%)
14 (9.5%)

5 (3.4%)
6 (4.1%)
121 (81.8%)
1 (.7%)

None
26 (18.6%)

16 (11.4%)
6 (4.3%)
89 (62.9%)
4 (2.9%)

 

Class Level (%)
Freshman
Sophomore
Junior
Senior

 

34 (23.0%)
33 (22.3%)
60 (40.5%)
21 (14.2%)

 

16 (11.4%)
28 (20.0%)
55 (39.3%)
41 (29.3%)

 

178

Summary of Audience’s Interactions with iPPLs: By Product Category

 

 

 

 

 

 

 

 

Interactions
Product Category Itemsl Mean Std. Median
(Ratez) Deviation
Clothes 12 2.197 (18.3%) 2.251 2
Furniture 3 1.502 (50.1%) 1.140 0
Inexpensive Electronics 1 .134 (13.4%) .438 0
Beverages 2 .692 (34.6%) .989 0
Home Accessories 2 .948 (47.4%) 1.948 0
Kitchenware 1 .210 (21.0%) .471 0
Tableware 2 .374 (18.7%) .744 0
Desserts 1 .248 (24.8%) .560 0
N = 396
l The number of items in the category
2 Mean divided by the number of overall items in the category
Summary of Audience’s Interactions with iPPLs: By Paired Characters
Interactions
Product Category Itemsl Mean Std. Median
(Ratez) Deviation
Chandler 3 .606 (20.2%) .939 0
Joey 3 .215 (7.2%) .638 0
Monica 3 .467 (15.6%) .837 0
Phoebe 3 .664 (22.1%) 1.014 0
Rachel 3 1.192 (39.7%) 1.366 0
Ross 3 .508 (16.9%) .974 0
N = 396

I The number of items paired with the corresponding character
2 Mean divided by the number of overall items in the category

179

Summary of Audience’s Interactions with iPPLs: By Brands

 

 

 

 

Interactions
Brand Name No. of Mean Std. Median
Items Deviation
Abercrombie & Fitch 1 .114 .341 0
American Eagle 1 .139 .431 0
Arizona Jeans 1 .046 .232 0
Art Van 1 .447 .660 0
Banana Republic 1 .144 .411 0
Calvin Klein 1 .058 .265 0
Crate & Barrel 1 .210 .471 0
DKN Y 1 .308 .588 0
Eddie Bauer 1 .450 .719 0
Gap * 1 .035 .222 0
IKEA 1 .217 .491 0
Impulse 1 .283 .552 0
J.Crew *"‘ 1 .571 .758 0
La-Z-Boy 1 .230 .560 0
Limited, The 1 .109 .384 0
Office Max 1 .182 .435 0
Panasonic 1 .134 .438 0
Polo Ralph Lauren 1 .159 .485 0
Pottery Barn 1 .202 .455 0
Sara Lee 1 .248 .560 0
Starbucks l .409 .736 0
Target 1 .172 .473 0
Tommy Bahama 1 .066 .303 0
WalMart 1 .404 .710 0
N = 396

"‘ Item with the lowest amount of interactions
" Item with the highest amount of interactions

I80

Frequencies of Audience’s Interactions with iPPLs

 

 

 

N 396
Mean 4.040
Median 3
Mode 3
Std. Deviation 3.701
Minimum 0
Maximum 22
Percentiles The Number of Interactions

10 0

20 1

30 2

40 2

50 3

60 4

70 5

80 7

 

181

BIBLIOGRAPHY

Alba, J ., Lynch, J ., Weitz, B., Janiszewski, C., Lutz, R., et al. (1997). Interactive home
shopping: Consumer, retailer, and manufacturer incentives to participate in
electronic marketplaces. Journal of Marketing, 61 (July), 38-53.

Aldersey-Williams, H. (1996). Interactivity with a human face. Technology Review, 99(2),
34-36.

Andrews, J. C., Durvasula, S., & Akhter, S. H. (1990). A framework for conceptualizing
and measuring the involvement construct in advertising research. Journal of
Advertising, 19(4), 27-40.

Antil, J. H. (1984). Conceptualization and operationalization of involvement. In T. C.
Kinnear (Ed.), Advances in Consumer Research: Vol. 11 (pp. 203-209). Ann Arbor,
MI: Association for Consumer Research.

Avery, R., & Ferraro, R. (2000). Verisimilitude or advertising? Brand appearances on
prime-time television. Journal of Consumer Aflairs, 34(2), 217-244.

Babin, L. A., & Carder, S. T. (1996). Viewers’ recognition of brands placed within a film.
International Journal of Advertising, 15(2), 140.

Bailey, M. (1992). The eflects of progressive levels of interactivity in an interactive video
lesson on achievement: Attitude and peer interaction. Unpublished doctoral
dissertation, Kansas State University.

Balasubramanian, S. K. (1994). Beyond advertising and publicity: Hybrid messages and
public policy issues. Journal of Advertising, 23(4), 29-46.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, 191-21 5.

Bandura, A. (1982). Self-efficacy: mechanism in human agency. American Psychologist,
37(2), 122-147.

Bandura, A., & Adams, N. E. (1977). Analysis of self-efficacy theory of behavioral
change. Cognitive Therapy and Research, [(4), 287-310.

Barlett, M. (2002, January 21). Interactive TV finally ready to take off. BizReport,
Retrieved May 22, 2002, from
http://www.bizreport.com/article.php?art_id=2801 &width= 1 024

Batra, R., & Ray, M. L. (1983). Operationalizing involvement as depth and quality of
cognitive response. In R. P. Bagozzi & A. M. Tybout (Eds), Advances in

182

Consumer Research: Vol. 10 (pp. 309-313). Ann Arbor, MI: Association for
Consumer Research.

Berthon, D. N. R. V. N. J.-P., & Davies, T. (1999). Going with the flow: Web sites and
customer involvement. Internet Research, 9(2), 109-116.

Bettman, J. R. (1979). An Information Processing Theory of Consumer Choice. Reading,
MA: Addison-Wesley.

Bloch, P. H. (1981). An exploration into the scaling of consumers’ involvement with a
product class. Advances in Consumer Research, 8, 61 -65.

Bloch, P. H. (1982). Involvement beyond the purchase process: Conceptual issues and
empirical investigation. Advances in Consumer Research, 9, 413-417.

Bowen, L., & Chaffee, S. H. (1974). Product involvement and pertinent advertising
appeals. Journalism Quarterly, 5 1 (Winter), 613-621, 644.

Brennan, 1., & Dubas, K. M. (1999). The influence of product-placement type and
exposure time on product-placement recognition. International Journal of
Advertising, 18(3), 323-338.

Brody, E. W. (1990). Communication Tomorrow: New Audiences, New Technologies, New
Media, New York: Praeger Press.

Bumkrant, R. E., & Sawyer, A. G (1983). Effects of involvement and message content on
information-processing intensity. In R. J. Harris (Ed.), Information Processing
Research in Advertising (pp. 43-64). Hillsdale, NJ: Lawrence Erlbaum Associates.

Center for Media Research (2002, August 29). Research brief — Ad clutter diminishes
performance. Center for Media Research, Retrieved August 29, 2002, from
http://www.mediapost.com/research/cfmr_briefArchive.cfin?s=1 80540

Buss, D. D. (1998). Making your mark in movies and TV. Nation Is Business, 86(12), 28-
32.

Cacioppo, J. T., & Petty, R. E. (1979). Effects of message repetition and position on
cognitive responses, recall, and persuasion. Journal of Personality and Social
Psychology, 37(January), 97-109.

Cannon, H. M. (1982). A new method for estimating the effect of media context — Using
value profiles of ads. Journal of Advertising Research, 22(5), 41-47.

Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and
comprehension processes. Journal of Consumer Research, 1 5 (September), 21 O-
224.

183

Chaiken, s. (1980). Heuristic versus systematic information processing and the use of
source versus message cues in persuasion. Journal of Personality and Social
Psychology, 39, 752-766.

Chatterjee, P., Hoffman, D. L., & Novak, T. P. (1998). Modeling the clickstream:
Implications for Web-based advertising eflorts. Unpublished manuscript.
Retrieved June 3, 2002, from
http://ecommerce.vanderbi1t.edu/research/papers/pdf/manuscripts/Clickstream-

pdf.pdf

Chattopadhyay, A., & Nedungadi, P. (1990). Ad affect, brand attitude, and choice: The
moderating roles of delay and involvement. In E. Goldberg, G. Gorn, & R. W.
Pollay (Eds), Advances in Consumer Research: Vol. 1 7 (pp. 619-620). Provo, UT:
Association for Consumer Research.

Chen, M. (1984). Computers in the lives of our children: Looking back on a generation of
television research. In R. Rice et al. (Eds), The New Media: Communication,
Research, and Technology (pp.269-286). Beverly Hills, CA: Sage.

Cho, C.-H. (1999). How advertising works on the WWW: Modified elaboration
likelihood model. Journal of Current Issues and Research in Advertising, 21(1),
33-50.

Cho, C.-H., & Leckenby, J. D. (1998). Copytesting of advertising on the WWW: Clicking
motivation profile. Proceedings of the Conference of the American Academy of
Advertising, 26-36.

Cho, C.-H., & Leckenby, J. D. (1999). Interactivity as a measure of advertising
effectiveness: Antecedents and consequences of interactivity in Web advertising.
Proceedings of the Conference of the American Academy of A dvertising, 162-179.

Cialdini, R. B., & Petty, R. E. (1981). Anticipatory opinion effects. In R. E. Petty, T. M.
Omstrom, & T. C. Brock (Eds), Cognitive responses in persuasion (pp. 217-235).
Hillsdale, NJ: Lawrence Erlbaum Associates.

Clancy, K. J ., & Ostlund, L. E. (1976). Commercial effectiveness measures. Journal of
Advertising Research, 16(1), 29-34.

Colford, S. W., & Magiera, M. (1991). Products in movies: How big a deal. Advertising
Age, 62(June), 12.

Cooper, A. (1999), The Inmates are Running the Asylum: Why High- Tech Products Drive
Us Crazy and How to Restore the Sanity. Indianapolis, IN: SAMS.

Copeland, G, & Schweitzer, K. (1993). Domination of the remote control during family

viewing. In J. R. Walker & R. V. Bellamy (Eds), Remote Control in the New Age
of Television (pp.155-168). Westport, CT: Praeger Publishers.

184

Comwell, N. C., Everett, S.-L., Everett, S. E., Moriarty, S., Russomanno, J. A., Tracey,
M., & Trager, R. (1993). Measuring RCD use: Method matters. In J. R. Walker &
R. V. Bellamy (Eds), Remote Control in the New Age of Television (pp.43-55).
Westport, CT: Praeger Publishers.

Curtis, J. (1996). Taking pride of place. Marketing, November, 28, 55.

d’Astous, A., & Seguin, N. (1999). Consumer reactions to product placement strategies in
television sponsorship. European Journal of Marketing, 33(9/10), 896-910.

DeLorme, D. E., & Reid, L. N. (1999). Moviegoers’ experiences and interpretations of
brands in films revisited. Journal of Advertising, 28(2), 71-95.

DeLorme, D. E., Reid, L. N., & Zimmer, M. R. (1994). Brands in films: Young
moviegoers ’experiences and interpretations. Paper presented at the Conference of
the American Academy of Advertising.

De Pelsmacker, P., Geuens, M., & Anckaert, P. (2002). Media context and advertising
effectiveness: The role of context appreciation and context/ad similarity. Journal
of Advertising, 31(2), 49-61.

Derks, P., & Arora, S. (1993). Sex and salience in the appreciation of cartoon humor.
Humor-International Journal of Humor Research, 6(1), 57-69.

DiClemente, C. C. (1981). Self-efficacy and smoking cessation maintenance: A
preliminary report. Cognitive Therapy and Research, 5, 175-187.

Eastman, S. T., & Newton, G D. (1995). Delineating grazing: Observations of remote
control use. Journal of Communication, 45(1), 78-96.

Evrard, Y., & Aurier, P. (1996). Identification and validation of the components of the
person-obj ect relationship. Journal of Business Research, 3 7(2), 127-134.

Fischer, B. R. (1996, January). Making your product the star attraction. Promo, 42-47.

Fortin, D. R. (1997). The impact of interactivity on advertising eflectiveness in the new
media. Unpublished doctoral dissertation, University of Rhode Island.

Frazer, C., & McMillan, S. (1996, May), Sophistication on the World Wide Web:
Evaluating structure, function and commercial goals of Web sites. Paper presented
at the Advertising and Consumer Psychology Conference, Bloomfield Hills, MI.

F risby, C. M. (2001). Building theoretical insights to explain differences in remote
control use between males and females: A meta-analysis. Journal of Current
Issues and Research in Advertising, 21(2), 59-76.

FTC Sponsorship Identification Rule, 47 CPR. §73.1212 (1996).

185

Gardner, M. P. (1985). Does attitude toward the ad affect brand attitude under a brand
evaluation set? Journal of Marketing Research, 22(May), 192-198.

Gardner, M. R, Mitchell, A. A., & Russo, J. E. (1978). Chronometric analysis:
Introduction and an application to low involvement perception of advertisements.
In H. K. Hunt (Ed.), Advances in Consumer Research: Vol. 5 (pp. 581-589). Ann
Arbor, MI: Association for Consumer Research.

Gibson, L. D. (1983). If the question is copy testing, the answer is not recall. Journal of
Advertising Research, 23(1), 39-55.

Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical analysis of its
determinants and malleability. Academy of Management Review, I 7(April), 183-
211.

Goldsmith, R. E., Lafferty. B. A., & Newell, S. J. (2000). The impact of corporate
credibility and celebrity credibility on consumer reaction to advertisements and
brands. Journal of Advertising, 29(3), 43-54.

Gotlieb, J. B., & Sarel, D. (1991). Comparative advertising effectiveness: The role of
involvement and source credibility. Journal of Advertising, 20(1), 38-45.

Gould, S. J ., Gupta, P. B., & Grabner-Krauter, S. (2000). Product placements in movies:
A cross-cultural analysis of Austrian, French, and American consumers’ attitudes
toward this emerging, international promotional medium. Journal of Advertising,
29(4), 41 -58.

Greene, W. F. (1988). Maybe the valley of the shadow isn’t so dark after all. Journal of
Advertising Research, 28(5), 11-15.

Greenspan, R. (2002, September 11). Interactive TV gaining acceptance. CyberAtlas,
Retrieved September 12, 2002, from http://cyberatlasinternet.com/big_picture/
applications/article/0,,1301_1461471,00.html

Greenwald, A. G (1968). Cognitive learning, cognitive response to persuasion and
attitude change. In A. G Greenwald, T. C. Brock, & T. M. Ostrom (Eds),
Psychological Foundations of Attitudes (pp. 147-170). New York: Academic Press.

Greenwald, A. G, & Leavitt, C. (1984). Audience involvement in advertising: Four levels.
Journal of Consumer Research, I 1 (June), 581-592.

Grover, R. (2002, June 17). Hollywood Product Placement, Circa 2054. Business Week.
Retrieved June 20, 2002, from http://www.businessweek.com/bwdaily/dnflash/
jun2002/nf20020617_8135.htm

Gunter, B., Fumham, A., & Beeson, C. (1997). Recall of television advertisements as a
function of program evaluation. Journal of Psychology, 131(5), 541 -553.

186

Gupta, P. B., & Gould, S. J. (1997). Consumers’ perceptions of ethics and acceptability of
product placements in movies: Product category and individual differences.
Journal of Current Issues and Research in Advertising, 19(1), 37-50.

Gupta, P. B., & Lord, K. R. (1998). Product placement in movies: The effect of
prominence and mode on audience recall. Journal of Current Issues and Research
in Advertising, 20(1), 47-59.

Ha, L., & James, E. L. (1998). Interactivity reexamined: A baseline analysis of early
business Web sites. Journal of Broadcasting & Electronic Media, 42(F all), 457-
474.

Hale, K. (1999a, Sep. 6). Enhanced TV: Wink Communications. Broadcasting & Cable,
129(37), 36.

Hale, K. (1999b, Sep. 6). Personal TV: Replay TV. Broadcasting & Cable, 129(37), 26.

Heeter, C. (1986). Perspectives for the development of research on media systems,
Unpublished doctoral dissertation, Michigan State University.

Heeter, C. (1988). Gender differences in viewing styles. In C. Heeter & B. Greenberg
(Eds), C ableviewing. Norwood, NJ: Ablex Publishing.

Heeter, C. (1989). Implications of new interactive technologies for conceptualizing
communication. In J. L. Salvaggio & J. Bryant (Eds), Media use in the
information age: Emerging patterns of adoption and consumer use flap. 217-235).
NJ: Lawrence Erlbaum Associates.

Heeter, C. (2000). Interactivity in the context of designed experiences. Journal of
Interactive Advertising, 1(1).

Heeter, C., & Greenberg, B. S. (1985). Profiling the zappers. Journal of aDvertising
Research, 25(2), 15-19.

Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychology, 21, 136-
141.

Hoffman, D. L., & Novak, T. P. (1996). Marketing in hypermedia computer-mediated
environments: Conceptual foundations. Journal of Marketing, 60(July), 50-68.

Hoffman, D. L., Novak, T. P., & Chatterjee, P. (1995). Commercial scenario for the Web:
Opportunities and challenges. Journal of C omputer-Mediated Communication,
1(3).

Ilfeld, J. S., & Winer, R. S. (2002). Generating website traffic. Journal of Advertising
Research, 42(5), 49-61

187

 

lnsko, C. A., Lind, E. A., & LaTour, S. (1976). Persuasion, recall, and thoughts.
Representative Research in Social Psychology, 7, 66-78.

ITV Marketer (n.d.). Prime Time Nears for Interactive TV in Europe. Retrieved May 22,
2002, from http://www.itvmarketer.com/mktres/europe/mreur02.htrn

Jacobs, G, & Dransfield, H. (1998). Scenarios for interactive TV — Europe’s uncertain
future. Long Range Planning, 31(3), 396-405.

Jaffe, M. (1996). Media interactivity, cognitive flexibility, and self-eflicacy. Unpublished
doctoral dissertation, University of Michigan.

Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-
analysis. Psychological Bulletin, 106(2), 290-314.

Kanner, B. (1993, November 29). All the right movies. New York, 20-21.

Kapferer, J .-N., & Laurent, G (1993). Further evidence on the consumer involvement
profile: Five antecedents of involvement. Psychology & Marketing, 10(4), 347-
355.

Karrh, J. A. (1994). Eflects of brand placements in motion pictures. Paper presented at the
Conference of the American Academy of Advertising. Athens, GA.

Karrh, J. A. (1998). Brand placement: A review. Journal of Current Issues and Research
in Advertising, 20(2), 31-49.

Kettanurak, V. (1996). An empirical investigation of the degree of interactivity in an
interactive multimedia instructional program. Unpublished doctoral dissertation,
University of Wisconsin - Milwaukee.

Krugman, D. M., Cameron, G T., White-McKearney, C. (1995). Visual attention to
programming and commercials: The use of in-home observations. Journal of
Advertising, 24(1), 1-12.

Krugman, H. E. (1965). The impact of television advertising: Learning without
involvement. Public Opinion Quarterly, 29, 349-356.

Krugman, H. E. (1967). The measurement of advertising involvement. Public Opinion
Quarterly, 30(Winter), 583-596.

Ku, L.(1992). Impacts of interactivity from computer-mediated communication in an
organizational setting: A study of electronic mail. Unpublished doctoral
dissertation, Michigan State University.

Laczniak, R. N., & Muehling, D. D. (1993). Toward a better understanding of the role of
advertising message involvement in ad processing. Psychology & Marketing,
10(4), 310-319.

188

Laurent, G, & Kapferer, J .-N. (1985). Measuring consumer involvement profiles. Journal
of Marketing Research, 22(February), 41-53.

Law, S. & Braun, K. A. (2000). I’ll have what she’s having: Gauging the impact of
product placements on viewers. Psychology & Marketing, 1 7(12), 1059-1075.

Lee, J .-H., & Edwards, S. M. (2002). The blame game: The moderating role of attribution
between self-efficacy and performance on attitude toward Internet sites [Abstract].
Proceedings of the Conference of the American Academy of Advertising, 98.

Lee, B., & Lee, R. S. (1995). How and why people watch TV: Implications for the future
of interactive television. Journal of Advertising Research, 35(6), 9-18.

Leigh, J. H., & Menon, A. (1987). Audience involvement effects on the information
processing of umbrella print advertisements. Journal of Advertising, 16(3), 3-12.

Longfellow, T. A., & Celuch, K. G (1993). Segmenting customers by their degree of
service involvement. Proceedings of the Summer Educators Conference,
American Marketing Association, 390-396.

Lord, K. R., & Bumkrant, R. E. (1993). Attention versus distraction: The interactive
effect of program involvement and attention devices on commercial processing.
Journal of Advertising, 22(1), 47-60.

Lord, K. R., Lee, M.-S., & Sauer, P. L. (1994). Program context antecedents of attitude
toward radio commercials. Journal of the Academy of Marketing Science, 22(1),
3-15.

Lutz, R. J., & Guiry, M. (1994). Intense consumption experiences: Peaks, performances,
and flows. Paper presented at the Winter Marketing Educators’ Conference, St.
Petersburg, FL, February.

Maclnnes, I. (1994). A model for standard setting: High definition television.
Contemporary Economic Policy, 12(4), 67-78.

Maclnnins, D. J ., & Jaworski, B. J. (1989). Information processing from advertisements:
Toward an integrative framework. Journal of Marketing, 53(October), 1-23.

Maclnnis, D. J., Moorman, C., & Jaworski, B. J. (1991). Enhancing and measuring
consumers’ motivation, opportunity, and ability to process brand information from
ads. Journal of Marketing, 55(October), 32-53.

Maclnnis, D. J ., & Park, C. W. (1991). The differential role of characteristics of music on
high- and low-involvement consumers’ processing of ads. Journal of Consumer
Research, 18(2), 161-173.

189

MaeKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural

antecedents of attitude toward the ad in an advertising pretesting context. Journal
of Marketing, 53(April), 48-65.

Massey, B. L., & Levy, M. R. (1999). Interactivity, online journalism, and English-
language Web newspapers in Asia. Journalism & Mass Communication Quarterly,
76, 138-151.

McClellan, S. (1997, Apr. 9). Microsoft buying Web-TV. Broadcasting & Cable, 12 7(1 5),
11.

McDonald, S. (1996, December). The most-wanted list. TV Guide, 28, 8.

McGuire (1976). Some internal psychological factors influencing consumer choice.
Journal of Consumer Research, 2(March), 302-319.

McMillan, S. J. (1998). Who pays for content? Funding in interactive media. Journal of
C amputer-Mediated Communication, 4(1).

McMillan, S. J. (2000a).Interactivity is in the eye of the beholder: Function, perception,
involvement, and attitude toward the Web site. Proceedings of the Conference of
the American Academy of Advertising, 71-78.

McMillan, S. J. (2000b). What is interactivity and what does it do? Paper presented at the

Conference of Association for Education in Journalism and Mass Communication,
Phoenix, AZ.

Merriam-Webster Collegiate Dictionary (2002). Retrieved August 1, 2002, from
http://www.m-w.com

Miller, M. (2001). Jon Boltax: At TV's inflection point. Television Broadcast Is Digital T V,
24(4), 33-34.

Miller, M. C. (1990). End of story. In M. C. Miller (Ed), Seeing through movies (pp. 186-
246). New York: Pantheon Books.

Mitchell, A. A. (1981). Dimensions of advertising involvement. In K. B. Monroe (Ed.),
Advances in Consumer Research: Vol. 8 (pp. 25-30). Ann Arbor, MI: Association
for Consumer Research.

Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediators of
advertising effects on brand attitudes? Journal of Marketing Research, 18(August),
318-332.

Morris, M., & Ogan, C. (1996). The Internet as mass medium. Journal of Communication,
46(1), 39-50.

190

Nahl, D. (1996). Affective monitoring of Internet learners: Perceived self-efficacy and
success. Journal of American Society for Information Sciences, 33, 100-109.

Nebenzahl, I. D., & Secunda, E. (1993). Consumers’ attitudes toward product placement
in movies. International Journal of Advertising, 12(1), 1-11.

Nelson, M. R. (2002). Recall of brand placements in computer/video games. Journal of
Advertising Research, 42(2), 80-92.

Neuman, W. R. (1991). The Future of the Mass Audience. Cambridge, MA: Cambridge
University Press.

Newhagen, J. E. (1998, July). Hitting the agenda reset button for the Internet: The
problem of matching research with development. Paper presented at the
International Communication Association Conference, Jerusalem, Israel.

Newhagen, J. E., Cordes, J. W., & Levy, M. R. (1995). Night1y@nbc.com: Audience
scope and the perception of interactivity in viewer mail on the Internet. Journal of
Communication, 45(3), 164-175.

Newhagen, J ., & Rafaeli, S. (1996), Why communication researchers should study the
Internet: A dialogue. Journal of Communication, 46(1), 4-13.

Norris, C. & Colman, A. M. (1992). Context effects on recall and recognition of
magazine advertisements. Journal of Advertising, 21(3), 37-46.

Novak, T. P., & Hoffman, D. L. (1997). New metrics for new media: Toward the
development of Web measurement standards. World Wide Web Journal, 2(1), 213-
246.

Novak, T. P., Hoffman, D. L., & Yung, Y.-F. (2000). Measuring the customer experience
in online environments: A structural modeling approach. Marketing Science, 19(1),
22-44.

Ohanian, R. (1991). The Impact of celebrity spokespersons’ perceived image on
consumers' intention to purchase. Journal of Advertising Research, 31(1), 46-54.

Osgood, C. E., & Tannenbaum, P. H. (1955). The principle of congruity in the prediction
of attitude change. Psychological Review, 62, 42-55.

Paivio, A. (1986). Mental Representativeness: A Dual Coding Approach. New York:
Oxford University Press.

Pastore, M. (2000, December 22). Digital TV, t-commerce capturing consumer interest.
CyberAtlas, Retrieved May 14, 2002, from
http://cyberat1asinternet.com/big_picture/
hardware/article/0,1323,5921_543651,00.htrn1

191

Pastore, M. (2002, January 23). ITV awaits prime time. CyberAtlas, Retrieved May 14,

2002, from http://cyberatlasinternet.com/big _picture/app1ications/article/
0,,1301_959941,00.html

Pavelchak, M. A., Antil, J. H., & Munch, J. M. (1988). The super bowl: An investigation
into the relationship among program context, emotional experience, and ad recall.
Journal of Consumer Research 15(3), 360-367.

Pavlik, J. V. (1996). New Media Technology. Needham Heights. MA: Allyn and Bacon.

Perry, S. D., Jenzowsky, S. A., King, C. M., & Yi, H. (1997). Using humorous programs
as a vehicle for humorous commercials. Journal of Communication, 47(1), 20-39.

Perse, E. M., & Ferguson, D. A. (1993). Gender differences in remote control use. In J. R.
Walker & R. V. Bellamy (Eds), Remote Control in the New Age of Television
(pp. 169-1 86). Westport, CT: Praeger Publishers.

Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease
persuasion by enhancing message-relevant cognitive responses. Journal of
Personality and Social Psychology, 3 7, 1915-1926.

Petty, R. E., & Cacioppo, J. T. (1980). Effects of issue involvement on attitudes in an
advertising context. In G G Gorn & M. E. Goldberg (Eds), Proceedings of the
division 23 program (pp. 75-79). Montreal, Canada: American Psychological
Association.

Petty, R. E., & Cacioppo, J. T. (1981). Issue involvement as a moderator of the effects on
attitude of advertising content and context. In K. B. Monroe (Ed.), Advances in
Consumer Research: Vol. 8 (pp. 20-24). Ann Arbor, MI: Association for

Consumer Research.

Petty, R. E., & Cacioppo, J. T. (1983). Central and peripheral routes to advertising
effectiveness: The moderating role of involvement. Journal of Consumer
Research, 10(September), 135-146.

Petty, R. E., & Cacioppo, J. T. (1986). Communication and Persuasion: Central and
Peripheral Routes to Attitude Change. New York: Springer-Verlag.

Petty, R. E., Cacioppo, J . T., & Schumann, D. (1983). Central and peripheral routes to
advertising effectiveness: The moderating role of involvement. Journal of
Consumer Research, 10(September), 135-146.

Phillips, J. M., & Gully, S. M. (1997). Role of goal orientation, ability, need for
achievement, and locus of control in the self-efficacy and goal-setting process.
Journal of Applied Psychology, 82(October), 792-802.

192

 

Pracejus, J. W. (1995). Is more exposure always better? Effects of incidental exposure to
a brand name on subsequent processing of advertising. Advances in Consumer
Research, 22, 319-22.

Rafaeli, S. (1988). Interactivity from new media to communication. In R. P. Hawkins, J.
M. Wiemann, & S. Pingree (Eds), Advancing Communication Science: Merging
Mass and Interpersonal Processes (pp. 110-134). Beverly Hills, CA: Sage.

Rafaeli, S. (1990). Interacting with media: Para-social interaction and real interaction. In
B. Ruben & L. Lievrouw (Eds), Mediation, Information and Communication
(pp. 125-1 81). New Brunswick: Transaction Publishers.

Rafaeli, S., & Sudweeks, F. (1997). Networked interactivity. Journal of Computer
Mediated Communication, 2(4).

Ratchford, B. T. (1987). New insights about the FCB grid. Journal of Advertising
Research, 27(4), 24-38.

Rice, R. (1984). New media technology: Growth and integration. In R. Rice & Associates
(Eds), The New Media: Communication, Research, and Technology (pp. 33-54).
Beverly Hills, CA: Sage.

Richins, M. L., & Bloch, P. H. (1986). After the new wears off: The temporal context of
product involvement. Journal of Consumer Research, 13(September), 280-285.

Rogers, E. (1986). Communication Technology: The New Media in Society, New York:
Free Press.

Russell, C. A. (1998). Toward a fi'amework of product placement: Theoretical
propositions. Advances in Consumer Research, 25, 357-362.

Sawyer, Alan (1981), Repetition, Cognitive Responses and Persuasion, In R. E. Petty, T.
M. Ostrom, & T. C. Brock (Eds), Cognitive responses in persuasion (pp. 237-
261). Hillsdale, NJ: Lawrence Erlbaum Associates.

Shaw, T., Amason, K., & Belardo, S. (1993). The effects of computer mediated
interactivity on idea generation: An experimental investigation. IEEE
Transactions on Systems, Man, and Cybernetics, 23(May/June), 737-746.

Shimp, T. A. (1981). Attitude toward the ad as a mediator of consumer brand choice.
Journal of Advertising, 10(2), 9-15, 48.

Singh, S. N., & Churchill, G A. (1987). Arousal and advertising effectiveness. Journal of
Advertising 16(1), 4-11.

Soldow, G F., & Principe, V. (1981). Responses to commercials as a function of program
context. Journal of Advertising Research, 21(2), 59-65.

193

Spalter, M. (1996). Maintaining a customer focus in an interactive age. In E. Forrest & R.
Mizerski (Eds), Interactive Marketing: The Future Present (pp. 163-188).
Lincolnwood, IL: NTC Business Books.

Srull, T. K. (1983). The impact of affective reactions in advertising on the representation
of product information in memory. In R. P. Bagozzi & A. M. Tybout (Eds),
Advances in Consumer Research: Vol. 10. Ann Arbor, MI: Association for
Consumer Research.

Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal
of Communication, 42(4), 73-93.

Stewart, D. W., & Ward, S. (1994). Media effects on advertising. In J. Bryant 85 D.
Zillman (Eds), Media Eflects, Advances in Theory and Research. Hillsdale, NJ:
Lawrence Erlbaum Associates.

Stroud, J. (2002). Interactive advertising on the Sky digital platform? Retrieved Aug. 15,
2002, from http://www.1ondon.edu/marketing/Future/Future_Media_Events/FM_
Presentations /Adrian_Stroud_pres_-_with_slides_de1eted.ppt

Sun Microsystems, Inc. (2001). Javam technologies for interactive television. Retrieved
Jun. 10, 2002, from http://java.sun.com/products/javatv/whitepapers/
TechInterTV052 1 01 .pdf

Swann, P. (2000). TV Dot Com: How Television is Shaping Our Future. New York: TV
Books.

Swedlow, T. (2000). Interactive enhanced television: A historical and critical perspective.
InteractiveTV Today, Retrieved June 15, 2002, from
http://www.itvt.com/etva.pdf

Swedlow, T. (2002, August 19). InteractiveTV today — Newsletter (Issue 4.66).
InteractiveTV Today, Retrieved August 19, 2002, from http://www.itvt.com

Tabachnick, B. G, & Fidell, L. S. (2000). Using multivariate statistics (4th ed.).
Needham Heights, MA; Allyn and Bacon.

Tafarodi, R. W., Milne, A. B., & Smith, A. J. (1999). The confidence of choice: Evidence
for an augmentation effect on self-perceived performance. Personality & Social
Psychology Bulletin, 25(November), 1405-1416.

Tavassoli, N. T., Schultz, C. J ., II, & F itzsimons, GJ. (1995). Program involvement: Are
moderate levels best for ad memory and attitude toward the ad? Journal of
Advertising Research, 35(5), 61-71.

TechTrends, Inc. (2000, December 15). T V-based e-commerce: An investigation of

consumer interest, pricing and payment preferences. Retrieved May 14, 2002,
from http://www.mindbranch.co.kr/report/print.asp?serviceid=r325-0004

194

Tedesco, R. (1996, Oct. 28). WavePhore, PBS using vertical blanking. Broadcasting &
Cable, 126(45), 90.

Turcotte, S. (1995). Gimme a Bud! T he feature film product placement industry.
Unpublished master Is thesis, University of Texas at Austin.

Vollmers, S., & Mizerski, R. (1994). A review and investigation into the effectiveness of
product placements in films. Proceedings of the Conference of the American
Academy of Advertising, 97-102.

Weaver, D. T., & Oliver, M. B. (2000). Television programs and advertising: Measuring
the eflectiveness of product placement within Seinfeld. Paper presented at the
Annual Conference of the International Communication Association, Acapulco,
Mexico.

Weinberger, M. G, & Spotts, H. E. (1989). A situational view of information content in
TV advertising. Journal of Marketing, 53(1), 89-646.

Weitz, B. A. (1978). Relationship between salesperson and performance and
understanding of customer decision making. Journal of Marketing Research,
15(4), 501-516.

Williams, R, Rice, R. E., & Rogers, E. M. (1988). Research Methods and the New Media.
New York: The Free Press.

Williams, R, Stover, S., & Grant, A. E. (1994). Social aspects of new media technologies.
In J. Bryant & D. Zillman (Eds), Media Eflects: Advances in Theory and
Research (pp. 463-482). Hillsdale, NJ: Lawrence Erlbaum Associates.

Wright, P. L. (1973). The cognitive process mediating acceptance of advertising. Journal
of Marketing Research, I 0(February), 53-62.

Wright (1974). Analyzing media effects on advertising responses. Public Opinion
Quarterly, 38(Summer), 192-205.

Wu, G (1999). Perceived interactivity and attitude toward Web sites. Proceedings of the
Conference of the American Academy of A dvertising, 254-262.

Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of Consumer
Research, 12(December), 341-352.

Zaichkowsky, J. L. (1994). The personal involvement inventory: Reduction, revision, and
application to advertising. Journal of Advertising, 23(4), 59-70.

195