PLACE II RETURN BOX to roman this Moat from your record.
TO AVOID FINES Mum on or baton duo duo.

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmatlvo ActloNEquul Opponunlty Institution
Wanna-m

THE EFFECTS OF ATTITUDE TOWARD THE AD
ON ATTITUDE TOWARD THE BRAND:
THE MODERATING ROLE OF DELAY AND REPETITION

BY

Kartik Pashupati

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Mass Media Ph.D. Program
College of Communication Arts and Sciences

1996

ABSTRACT

THE EFFECTS OF ATTITUDE TOWARD THE AD
ON ATTITUDE TOWARD THE BRAND:
THE MODERATING ROLE OF DELAY AND REPETITION

By
Kartik Pashupati

A number of studies in the last two decades have established that
consumers’ attitude toward the ad (Aad) has a direct effect on their attitude
toward the advertised brand (Ab). Most of these studies have measured Ab
immediately after exposure to the stimulus ad, and suggest a linear relationship
between Aad and Ab. However, the familiarity-based sleeper effect (Moore and
Hutchinson 1983) suggests that over time, Ab will be a U-shaped function of Aad,
with negatively and positively evaluated ads producing higher levels of Ab
compared to neutral ads.

This dissertation tested hypotheses based on the familiarity-based sleeper
. effect. It also tested the hypothesis that ads producing negative and positive Aad

would be better recalled than neutral ads. The study also examined the effect of
advertising repetition on the relationship between Aad and Ab. It was I
hypothesized that liking for all types of ads (including negatively evaluated
ones) would increase as a result of moderate levels of repetition.

The hypotheses were tested using a 12 cell factorial design (3 ad types, 2
repetition levels, and 2 delay levels). Experimental subjects were exposed to
commercials for real (but unfamiliar) products embedded in a thirty minute
television program, along with filler commercials. Subjects were exposed to the
target—ad either once or thrice. Their Ab was measured either immediately after

ad exposure or seven days later.

The results of the study showed that the negative ad produced the greatest
recall, and the positive ad the least. The data did not support the existence of a
familiarity-based sleeper effect, as Ab was found to be a linear function of Aad for
both no-delay and delay groups. Repetition was found to increase liking of all
three ad types, although the increment in liking was statistically significant only
for the negative ad.

This study also examined the factor structure of Aad- The results indicate
that Aad is determined by four factors: entertainment value, irritation, (3)

utilitarianism, and distinctiveness.

Capyrisht by
Kartik Pashupati
1996

To my mother, and to the memory of my father,
for all their sacrifices toward my education.

ACKNOWLEDGEMENTS

This dissertation would not have been possible without the guidance,
support and encouragement of many mentors, colleagues, friends and loved
ones. First, I would like to thank the Chair of my dissertation committee, Dr.
Bruce Vanden Bergh, for his patience and perseverence in overseeing this
document from proposal to completion. It is indeed a privilege to have worked
with him on this and other projects.

I am grateful to the other members of my dissertation committee — Dr.
Thomas Baldwin, Dr. Bonnie B. Reece, and Dr. R. Dale Wilson — for their
guidance. I could not have asked for a better group of guides. None of them
were easy taskmasters, but they were always fair.

I am grateful to all the professors at Michigan State who taught me so
much, especially Dr. Gordon E. Miracle and Dr. E. Lincoln James for mentoring
me during my early days in the PhD. program. I would also like to thank my
peers in the Mass Media Ph.D. program, and all my friends at Michigan State
University. Their company made my years in the doctoral program extremely
enjoyable, besides teaching me a lot.

I would also like to thank my colleagues in the Department of
Communication Arts at the University of West Florida, for their patience and
encouragement during the final phases of this dissertation. I am especially
grateful to Peter Gershon for allowing me to use his documentary, Sou them

Voices, Southern Words, as part of the stimuli for my experiment, and to Cindy

Hill, who selflessly spent endless hours in the editing room to insert commercials
seamlessly into the documentary.

Various family members deserve a note of thanks. My father who, until
his dying day, wanted to know when I was going to ”complete the dissertation.”
My mother, whose persistent reminders are largely responsible for the eventual
completion of this document. My brother and his wife, who have provided
financial, moral and emotional support throughout my academic career. And of
course, my marvelous, fabulous wife, Pushkala, without whom none of this
would have been either possible or worthwhile.

Finally, I would like to thank the American Academy of Advertising for
supporting this research project with a Doctoral Dissertation Competition grant.
It was an honor to be awarded the grant. I hOpe this dissertation does justice to

the award.

TABLE OF CONTENTS

List of Tables ................................................................................................................ viii
List of Figures ............................................................................................................... x
Chapter 1: INTRODUCTION ..................................................................................... 1
Chapter 2: LITERATURE REVIEW ........................................................................... 7
2.1. THEORIES OF ADVERTISING EFFECTIVENESS ............................. 7

2.1.1. Hierarchy-of—Effect Models ..................................................... 7

2.1.2. Message Learning Approaches ............................................... 8

2.1.3. Elaboration Likelihood Model ................................................ 9

2.1.4. The ELM and Conditioning Approaches .............................. 10

2.2. ATTITUDE TOWARD THE AD ............................................................ 11

2.2.1 Antecedents of Aad ..................................................................... 13

2.2.2 Effects of Aad ................................................................................ 16

2.2.3 Moderators of Aad ....................................................................... 17

2.3. MEMORY AND ATTITUDE TOWARD THE AD .............................. 19

2.3.1. Immediate and Delayed Effects .............................................. 19

2.3.2. Emotional Response and Ad Recall ........................................ 21

2.3.3. Details of three memory-related studies ................................ 24

2.4. THE SLEEPER EFFECT ........................................................................... 31

2.5. EFFECTS OF REPETITION ..................................................................... 33
Chapter 3: HYPOTHESES ........................................................................................... 41
3.1. IMMEDIATE VERSUS DELAYED EFFECTS ........................................ 41

3.2. REPETITION EFFECTS ........................................................................... 42
Chapter 4: METHOD ................................................................................................... 45
4.1. VARIABLES AND SCALES .................................................................... 45

4.1.1. Independent Variables ............................................................. 45

4.1.2. Dependent Variables ................................................................ 46

4.1.3. Moderating Variables ............................................................... 46

4.2 SCALE RELIABILITIES ............................................................................. 49

4.3. SUBJECTS .................................................................................................. 50

viii

4.4. STIMULUS MATERIALS ........................................................................ 50

4.4.1. Pretesting .................................................................................... 50
4.4.2. Stimulus program ..................................................................... 53
4.5. PROCEDURE ............................................................................................ 56
Chapter 5: RESULTS .................................................................................................... 58
5.1 MANIPULATION CHECK ...................................................................... 58
5.2 GLOBAL CORRELATIONS ..................................................................... 59
5.2.1. Tercile Split ................................................................................. 60
5.3 HYPOTHESIS TESTING ........................................................................... 61
5.3.1. Unaided recall across ad types ................................................ 61
5.3.2. Aided recall across ad types .................................................... 64
5.3.3. Overall effect of Aad on Ab ..................................................... 65
5.3.4 Immediate and Delayed Effects of Aad on Ab ...................... 66
5.3.5. Effect of delay on ad recall ....................................................... 69
5.3.6. Effects of repetition on Aad and Ab ....................................... 75
5.3.7. Effect of repetition on overall ad recall .................................. 78

5.3.8. Effect of repetition on immediate versus delayed
recall ....................................................................................................... 80
5.3.9. Effect of repetition on recall across ad types ......................... 82
5.4 EFFECT OF INVOLVEMENT ................................................................. 83
5.5 SUMMARY ................................................................................................ 84
Chapter 6: DISCUSSION AND CONCLUSIONS .................................................... 85
6.1 EFFECT OF Aad ON Ab ........................................................................... 85
6.2 DIMENSIONS OF Aad .............................................................................. 87
6.2.1. Factor analysis ........................................................................... 88
6.2.2. Regression analysis ................................................................... 92
6.3 RECALL OF DIFFERENT AD TYPES .................................................... 94
6.4 FAMILIARITY BASED SLEEPER EFFECT ........................................... 95
6.5 REPETITION: WEARIN AND WEAROUT .......................................... 96

ix

6.6 LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH ...... 97

6.6.1. Stimulus materials ..................................................................... 97

6.6.2. Subjects ....................................................................................... 98

6.6.3. Brand familiarity effects ........................................................... 98

6.6.4. Variation in delay ...................................................................... 100

6.6.5. Ad and Brand Cognitions ........................................................ 100

6.7 CONCLUSION .......................................................................................... 101
APPENDIX 1: Questionnaire ...................................................................................... 102
APPENDIX 2: Scripts of the Stimulus Commercials ............................................... 119
LIST OF REFERENCES ............................................................................................... 123

LIST OF TABLES

Table 1 - Personal/ individual antecedents of attitude toward the ad ............... 37
Table 2 - Ad-related factors/ antecedents of Aad .................................................... 38
Table 3 - Other factors influencing Aad ..................................................................... 39
Table 4 - Consequences/ effects of Aad .................................................................... 39
Table 5 - Moderators of Aad ........................................................................................ 40
Table 6 - Scales used to measure key constructs ...................................................... 47
Table 7 - Reliability coefficients for key measurement scales ................................ 49
Table 8 - Pretest results ................................................................................................ 52

Table 9 - Mean scores for attitude toward the ad, attitude toward the
brand and purchase intentions across ad types .............................................. 58

Table 10 - Correlations between attitude toward the ad, attitude toward
the brand and purchase intentions ................................................................... 60

Table 11 - Mean scores for attitude toward the ad, attitude toward the

brand and purchase intentions across ”ad groups” ....................................... 61
Table 12 - Unaided recall of different ad types ........................................................ 62
Table 13 - Unaided recall of different ads classified by ad group ......................... 63
Table 14 - Aided recall of different ad types ............................................................ 64
Table 15 - Aided recall of different ads classified by ad group ............................. 65
Table 16 - Regression Analysis results: Standardized Regression

Coefficients ........................................................................................................... 65
Table 17 - Mean Ab and PI Scores by Ad Type and Delay Condition .................. 66
Table 18 - Mean Ab and PI Scores by Ad Group and Delay Condition ............... 67
Table 19 - Effects of Aad and Delay on Ab ................................................................ 68
Table 20 - Unaided Recall of Ads by Delay Condition ........................................... 69

xi

Table 21 - Unaided Recall of Ad Types by Delay Condition ................................. 70

Table 22 - Unaided Recall of Ad Groups by Delay Condition .............................. 72
Table 23 - Aided Recall of Ad Types by Delay Condition ..................................... 73
Table 24 - Aided Recall of Ad Groups by Delay Condition ................................... 74
Table 25 - Mean Aad and Ab Scores by Repetition Condition ................................ 76
Table 26 - Effects of Aad and Repetition on Ab ........................................................ 78
Table 27 - Unaided Recall of Target Ad by Repetition ........................................... 79
Table 28 - Aided Recall of Target Ad by Repetition ................................................ 80
Table 29 - Unaided Recall of Ad Types by Repetition Condition ......................... 80
Table 30 - Unaided Recall of Ads by Delay and Repetition .................................. 81
Table 31 - Regression results for the effect of involvement and Aad on Ab ........ 84
Table 32 - Loadings on varimax rotated principal components ............................ 89

Table 33 - Standardized regression coefficients: Effect of Factors on Global
Aad ......................................................................................................................... 93

xii

LIST OF FIGURES

Figure 1 - A typology of ad based persuasion mechanisms .................................. 12
Figure 2 - Relationships predicted by the Generalization Hypothesis ................. 25
Figure 3 - Relationships predicted by the Distraction Hypothesis ....................... 26
Figure 4 - Relationships predicted by the Distinctiveness Hypothesis ................ 27
Figure 5 - Relationships predicted by the Familiarity-Based Sleeper

Hypothesis ........................................................................................................... 27
Figure 6 - Relationships predicted by the Affect-Based Sleeper Hypothesis ...... 28
Figure 7 - Expected interaction between repetition and delay .............................. 44
Figure 8 - Expected interaction between repetition and ad affect ........................ 44
Figure 9 - Sequence of commercial breaks ............................................................... 55
Figure 10 - Comparison of unaided recall by ad type and ad group ................... 64
Figure 11 - Observed relationship between Aad (ad group) and Ab for no-

delay and delay conditions ................................................................................ 67
Figure 12 - Unaided recall of different ad types by no-delay and 7-day

delay groups ........................................................................................................ 71
Figure 13 - Unaided recall of different ad groups by no-delay and 7-day

delay groups ........................................................................................................ 72
Figure 14 - Aided recall of different ad types by no-delay and 7-day delay

groups ................................................................................................................... 74
Figure 15 - Aided recall of different ad groups by no-delay and 7-day

delay groups ........................................................................................................ 75
Figure 16 - Effect of repetition on Aad for each ad type ......................................... 76
Figure 17 - Effect of repetition on Ab scores for each ad type ................................ 77
Figure 18 - Unaided recall of ads by delay and repetition ..................................... 81
Figure 19 - Observed improvement in recall across ad types ................................ 82

xiii

Chapter 1
INTRODUCTION

Researchers in the field of marketing and advertising have long been
concerned with understanding and influencing consumer behavior. One of the
most frequently used models of consumer behavior, in both academic research
and in marketing decision making, is the hierarchy of effects model of
advertising. Hierarchical models of consumer behavior, such as the one
proposed by Lavidge and Steiner (1961) stipulated that changes in purchase
behavior had to be preceded by changes in consumers' attitudes toward the
advertised brand. These attitude changes were in turn held to be preceded by
changes in (brand related) cognitions. The Lavidge and Steiner model and other
hierarchical models owe their genesis directly to message-learning theories of
attitude change, originally pioneered by Hovland and his colleagues in the 19503
(Hovland, Janis and Kelley 1953).

The role of attitudes in explaining behavior has attracted the attention of
researchers in soda] psychology, communication and marketing. McGuire (1976)
has noted that the area of attitude change constitutes the largest single body of
literature in social psychology. The work of researchers such as Fishbein and
Azjen (1974) has provided fresh impetus for researchers investigating the impact
of attitudes on behavior (Petty and Cacioppo 1981).

According to Fishbein's Theory of Reasoned Action, consumers' attitudes
toward an advertised brand should be mediated exclusively by their brand
beliefs, and their evaluations of those beliefs. Thus, advertising can influence
consumer behavior primarily through changing beliefs or evaluations about the
brand. However, research by Mitchell and Olson (1981) suggested that changes
in brand beliefs do not account for all the variations in consumers' attitudes

toward an advertised brand. Mitchell and Olson (1981) and Shimp (1981)

2
hypothesized that consumers' attitudes toward the advertisement has a direct

impact on their attitudes toward the brand, which is not completely captured by
measures of the changes in their brand beliefs. C

Following the work of Mitchell and Olson (1981) and Shimp (1981), the
topic of attitude toward the ad (Aad) and its impact on attitude toward the
advertised brand (Ab) has received much attention from researchers in the last
two decades. A recent review of the research in this area by Muehling and
McCann (1993) lists over 95 published works dealing directly or indirectly with
the impact of Aad- Yet, Muehling and McCann (1993) feel that several research
issues in this area need to be given further attention. One of the topics that has
been mentioned as worthy of consideration in future studies is the role that
memory plays in moderating the longer term effects of Aad on Ab. This
dissertation proposes to extend the research in this specific domain.

There has been surprisingly little research investigating the impact of
delay (between ad exposure and Ab measurement) on the relationship between
Aad and A1,. Moore and Hutchinson (1983, 1985) are among the few researchers
who have investigated this relationship. In their initial research, using real print
ads as stimuli, they found that in the short run (2 days after ad exposure), A}, is a
linear function of Aad, with positively evaluated ads producing more positive
brand attitudes. After a 7-day delay, however, Abwas shown to be a U-shaped
function of Aad. This suggests that (1) pe0ple tend to remember affectively
stronger ads (regardless of valence) more than neutral ads, and (2) over time,
brand familiarity (working through ad recall) may be playing a greater role in
determining Ab than the direct transfer of affect through Aad (Moore and
Hutchinson 1983).

In their follow-up research, Moore and Hutchinson (1985) found results
similar to their 1983 study. However, the linear/ curvilinear interaction

3

observed in their data was not as consistent as they had expected. They offered
two possible explanations for this: (1) the print ads which were used as stimuli
did not produce genuinely extreme affective reactions; (2) an experimental
manipulation had made their subjects overly sensitive to brand names in ads
(Moore and Hutchinson 1985, p. 76). Since researchers as well as practitioners
are of the opinion that television commercials are more effective in producing
emotional responses than print ads, one of the objectives of this study is to
extend the research of Moore and Hutchinson (1985) by using television
commercials as stimuli, instead of print advertising.

The immediate and delayed effects of emotional feelings produced by
television commercials were explored by Thorson and Friestad (1989). They
found that emotional commercials are more likely to be recalled than
commercials unaccompanied by emotion. These findings are consistent with
those of Moore and Hutchinson (1983, 1985) reported above. However, Thorson
and Friestad (1989) measured ad recall as the dependent variable, while Moore
and Hutchinson (1983, 1985) used persuasion measures such as brand affect,
change in liking for the brand, and changes in purchase intention (willingness to
consider the brand). It would be easy to conclude from this data that there is
some collinearity between ad recall and persuasion. However, the use of recall
measures, and the relationship between ad recall and persuasion, has been
fraught with controversy in the literature on advertising copytesting (Stewart, et
al. 1985). The present study will extend the existing research in the domain by
using both recall and persuasion measures, and examining the correlation
between them.

The effect of delay on Aad and Ab has also been examined in a study by
Chattopadhyay and N edungadi (1992)., As stimuli, they used television

commercials for an unfamiliar brand of pen, embedded in program material.

4
The independent variables manipulated were ad type (likable and neutral

versions of a single ad for the same product), attention paid to the ad (high and
low), and delay (no delay and 7-day delay). The dependent variables were Aad,
Ab, number of ad-directed cognitive responses, and number of brand-directed
cognitive responses. The results of the research included the findings that (1) the
number of ad-directed cognitions declined over time, but the number of brand-
directed cognitions was more resilient to decay; (2) while there was a decline
over time in Aad for subjects exposed to the likable ad, there was actually an
increase over time in Aad for subjects exposed to the neutral ad; (3)10w-attention
subjects in the 7-day delay group who were exposed to the likable ad rated the
brand lower than those exposed to the neutral ad (italics added). Chattopadhyay
and Nedungadi (1992) suggest that in the absence of strong evidence, the data
seem to suggest the existence of a "sleeper effect" arising from a dissociation
between the neutral ad and the advertised brand. A similar theory is also offered
by Moore and Hutchinson (1983). (A brief discussion of the literature on the
sleeper effect is offered in the next chapter.)

While the study by ChattOpadhyay and N edungadi (1992) does not
directly address the issue of the impact of delay on the Aad —> Ab relationship, it
does offer support for some of the findings of Moore and Hutchinson (1983,
1985). ChattOpadhyay and N edungadi (1992) used two versions of a single
("target") television commercial as their stimulus, along with several filler ads
embedded in a program environment. This is a laudable effort to achieve a more
realistic exposure environment within the confines of an experimental setting.
However, Chattopadhyay and Nedungadi (1992) acknowledge that using only
one exposure to the target commercial might reduce the accessibility of ad
cognitions and ad attitude in consumer memory. They suggest that future

research should study the impact of ad repetition on decreasing, or at least

5
postponing, the negative effects of delay. Recent research by Haugtvedt and his

associates (Haugtvedt, Leavitt and Scheiner 1993, Haugtvedt, et al. 1994,
Haugtvedt and Wegener 1994) has also underlined the need to study the effects
of repetition, not just in terms of changing attitudes but also in terms of varying
attitude strength. Accordingly, another objective of this study is to augment the
knowledge in this area by studying the impact of repetition on immediate and
delayed measures of Aad, Ab and recall.

In sum, this study will examine relationships in the following domains,

using television commercials as stimuli:

1. The immediate and delayed impact of affective responses to the ad (Aad)

on affective responses to the brand (Ab ), ad recall and purchase intention.

2. The impact of ad repetition on immediate and delayed measures of Aad,
Ab, ad recall and purchase intention.

This study will make contributions to the existing body of knowledge by
(1) attempting to replicate and extend the results found by previous researchers,
and (2) studying the effects of interactions not studied by previous researchers.

The proposed area of research should be of interest to both academicians
and practitioners. Much of the research on the effects of advertising has used
measures immediately following ad exposure. However, advertising is expected
to work on at least two levels: first, in terms of producing short term sales, and
second, in terms of creating and retaining brand image over the longer term. The
longer term effects of advertising are especially important in the light of the
recent interest in the subject of brand equity (e.g., Aaker and Biel 1993).
Therefore, any study of the delayed effects of advertising can help advertisers

understand better how advertising works at both these levels.

6
This dissertation is divided into six chapters. A detailed literature review

is presented in Chapter 2. The hypotheses flowing from the literature are
presented in Chapter 3. The methodology and experimental design used to test
the hypotheses are presented in Chapter 4. The results of the study are reported
in Chapter 5. Finally, Chapter 6 contains a discussion of the results reported in
Chapter 5. This chapter also contains an extended discussion of some of the
issues in the literature concerning the multi-dirnensional nature of attitude
toward the ad. The academic contributions of this study and its managerial
implications are also discussed in Chapter 6, along with limitations and

suggestions for future research.

Chapter 2
LITERATURE REVIEW

This chapter provides a review of the literature on advertising
effectiveness, with particular emphasis on the effects pertaining to attitude
toward the ad (Aad), memory, ad repetition and prior brand familiarity. First,
various cognitive theories of advertising effectiveness are reviewed, with
emphasis on the hierarchy-of-effects models that have guided advertising
planning in recent decades. Second, the literature on attitude toward the ad is
reviewed. Emphasis is placed on literature pertaining to the delayed impact of
Aad on Ab, and on findings about the connection between Aad and recall. A
general discussion is followed by a detailed review of three studies. The final
section of this chapter reviews the literature pertaining to the effects of
advertising repetition.

2.1. THEORIES OF ADVERTISING EFFECTIVENESS
2.1.1. Hierarchy-of-Effect Models

Ask most people why they think businesses should advertise, and their
response is likely to be, "Why, to sell things, of course!" Indeed, a lot of marketers
do set advertising objectives in terms of increases in sales or market share.
However, both marketing practitioners and academic researchers have
recognized at least since the early 19605, that it is often difficult to establish a
strict correlation between advertising and sales for two reasons: (1) advertising is
only one of many factors influencing sales, and (2) the contributory role of
advertising often occurs primarily over the long run (Aaker and Myers 1987, p.
86). This realization has led advertisers to specify advertising objectives in terms
of (1) the ultimate behavior (e. g., trial, brand switching, reinforcement of loyalty,
etc.) that advertising is attempting to precipitate in the target audience, and (2)

changes in the psychological variables preceding the performance of that

8
behavior. The notion of defining advertising objectives in terms of

communication tasks (such as creating awareness, changing or reinforcing
existing attitudes, creating knowledge of brand attributes, etc.) rather than in
terms of marketing tasks (such as increasing sales or market share), received
much impetus in the early 19605 due to the work of Lavidge and Steiner (1961).
In the same year, Colley (1961) wrote a book under the sponsorship of the
Association of National Advertisers, which pOpularized these ideas among
practitioners (Aaker and Myers 1987).

Colley (1961) stipulated that the process of setting advertising objectives
should be guided by some kind of a hierarchy-of-effects model, such as that
formulated by Lavidge and Steiner (1961). According to the Lavidge and Steiner
hierarchical model, consumer response to advertising begins with awareness and
knowledge, followed by liking and preference for the advertised brand, followed
by conviction about the brand, and finally by purchase of the brand. Palda (1966)
summarized these steps into the now familiar three stage ”Cognition—> Affect
—> Conation” model. As noted by Smith and Swinyard (1988), the introduction
of the Lavidge and Steiner (1961) model caused marketing researchers to begin a
closer examination of the cognitive dimensions of consumer responses to
advertising.

2.1.2. Message Learning Approaches

Hierarchical models of advertising effects, such as the one proposed by
Lavidge and Steiner (1961), were in many ways directly derived from what Petty
and Cacioppo (1981) refer to as "message-learning approaches" to understanding
attitude change. The message learning approach, based on the pioneering
research of Hovland and his associates (Hovland, Janis and Kelley 1953),

emphasized that the learning and memory of persuasive arguments were critical

9

for the formation of initial attitudes, as well as for the change or reinforcement of
existing attitudes.

The message learning approach to understanding attitude was challenged
by researchers such as Greenwald (1968) when empirical studies failed to
demonstrate that a significant correlation existed between learning and
persuasion in all instances. The cognitive response approach, which emphasized
the importance of an individual's idiosyncratic cognitive reactions to a
persuasive message, was offered as an alternative to the message learning
approach (Haugtvedt, Leavitt and Scheiner 1993). An example of this approach
is found in the work of Greenwald and Leavitt (1985), who suggest that
individuals have control over the amount of attention devoted to the
communication and adjust attention levels depending on their involvement with
the communication.

2.1.3. Elaboration Likelihood Model

As noted by Alwitt and Mitchell (1985b), the cognitive response approach
of Greenwald and Leavitt (1985) has a direct parallel in the elaboration likelihood
model (ELM) (Cacioppo and Petty 1985, Petty and Cacioppo 1981). The ELM
focuses on the amount of cognitive resources devoted to processing the content
of the message, and on the type of verbal responses that an individual makes to a
persuasive message (Alwitt and Mitchell 1985b). If individuals put considerable
effort into processing the information from the message, and allocate most of
their cognitive resources to this task, then persuasion is said to occur by the
central route. On the other hand, if individuals put little effort into processing the
content of the message, and instead use other cues in the message (such as
number of message arguments, spokesperson characteristics, visual appeal of the
ad, etc.), then persuasion is said to occur by the peripheral route. In their initial

analysis, Petty and Cacioppo (1981) treated central processing as pertaining to

10

message content, and peripheral processing as related to the context of the
message (Lutz 1985).
2.1.4. The ELM and Conditioning Approaches

One of the major contributions of the ELM is that it provides a means of
reconciling high involvement theories of persuasion (such as the message
learning approach, and cognitive response approach) with earlier, low
involvement theories, such as classical and Operant conditioning. This contribution
is especially important, given the recent spate of database driven customer
loyalty programs (such as frequent flyer programs, the Discover Card cash-back
program, etc.), which are essentially based on some form of operant conditioning
(Peter and Olson 1990, Chapter 9). While the information-processing (or
cognitive) approach has largely replaced behaviorism (or stimulus-response
reinforcement theory) as the dominant paradigm in persuasion research
(McGuire 1976), some researchers have expressed dissatisfaction with the
explanatory power of cognitive response models in the case of advertisements for
low involvement products (e.g., Krugman 1965). Some recent researchers (e.g.,
Gorn 1982) have suggested that advertising for low involvement products
probably persuades through a classical conditioning process rather than a
message learning process.

The failure of cognitive learning models to explain consumer behavior in
low involvement situations resulted in research interest in consumers' affective
responses to advertising. Under traditional high involvement models, the
executional elements of advertisements were viewed only as vehicles for
communicating the message. However, empirical research by Mitchell and
Olson (1981) and others suggested that individuals' affective reactions to the

executional elements in advertisements could also affect persuasion (Alwitt and

11

Mitchell 1985a). This led to a stream of research on the effects of consumers'
attitudes toward the ad (Aad)-
2.2. ATTITUDE TOWARD THE AD
The current stream of research into consumers' attitude toward the ad

W~w_._. —.

(Aad) and its impact on attitude toward the advertised brand_’(Ab)flrs~_u§”ually

 

“.-

(”-

traced to molarticles that appeared in1981: Shimp (1981) and Mitchell and
Olson (1981) both suggested that the beliefs of consumers about a brand are not
the only mediators of the impact that an ad has on their attitude toward the

brand, They found empirical support for the hypothesis that attitude toward the

 

ad mediates consumers' attitudes toward the brand. Mitchell and Olson (1981)
spgadétéa that the transference of affect from the ad to the brand could be
occurring due to some kind of classical conditioning process.

Using Petty and Cacioppo's (1981) ELM framework, Lutz (1985) has
offered a conceptual model of the affective and cognitive antecedents of Aad- He
proposed a typology of four alternative ad-based persuasion mechanisms: (1)
classic message-based persuasion, (2) dual mode persuasion, (3) pure affect
transfer, and (4) contextual evaluation transfer. These models are graphically
presented in Figure 1. Lutz (1985) notes that the first three mechanisms
(particularly the first two) are based upon the ELM. However, he states that
ELM does not deal with the fourth situation (contextual evaluation transfer),
where both the context and the content are processed peripherally.

Lutz (1985) has gone on to observe that the typical advertising pretesting
situation is best represented by the situation that he labels "Contextual
Evaluation Transfer (see Figure 1)." In such situations, consumers are typically
exposed to ads for unfamiliar brands, and asked to draw inferences and form
attitudes about the brand based exclusively on the sample ad. However,
empirical research by MacKenzie and Lutz (1989) and others (Brown and

12
Stayrnan 1992) suggests that "Dual Mode Persuasion" (also referred to the as

"Dual Mediation Hypothesis") is the most prevalent type of ad-based persuasion.

 

 

 

 

 

 

Figure 1.
A typology of ad-based persuasion mechanisms.a
FOCUS OF PROCESSING
Ad Content Ad Context
(Mi \ \ Atts \
\ < \ \ ‘
Central Cad ------ > Aad Cad €> d
Processan I
Mode l
V
Message-Based Persuasion Dual Mode Persuasion
Atts Atts \
\ \
Perlpheral Cad ______ > Aad; {Cad} " {>Aad
Processlng " " " - - ’
Mode
Cb —————— >Ab Cb —————— >Ab
Pure Affect Transfer Contextual Evaluation Transfer

 

 

 

 

Aad = Attitude toward the ad.

Atts = Attitude toward advertising in general, attitude toward the advertiser, moods.
Cad = Ad cognitions.

Cb = Brand cognitions.

Solid arrows indicate strong positive relationships.
Dashed arrows indicate relationships hypothesized to be zero or near-zero.

aAdapted from Lutz, 1985.

13
Since the early work of Mitchell and Olson (1981) and Shimp (1981), over

90 studies dealing with the effects of Aad have been published in the marketing
and advertising literature (Muehling and McCann 1993). These studies have
~ investigated the antecedents of Aad, and its cognitive, affective and behavioral
effects. Researchers have also investigated the moderating effects of various
factors on the Aad —> A}, relationship. The moderating factors which have been
investigated include involvement, processing goal, ad type, retrieval cues,
personal relevance, brand familiarity, prior brand attitudes, nature of claims and
time. These variables are summarized in Tables 1 through 5 (the tables are
grouped together at the end of this chapter for convenient reference). Some of
the major findings with respect to these variables are reported below.
2.2.1 Antecedents of Aad

In their review, Muehling and McCann (1993) have subdivided the
antecedents of Aad into three categories: (1) personal/ individual factors, (2) ad
related factors, and (3) ”other” factors. This typology will be retained in the
following discussion.
a. Personal] Individual Antecedents

Lutz and his colleagues (Lutz, MacKenzie and Belch 1983; MacKenzie and
Lutz 1989; MacKenzie, Lutz and Belch 1986) have suggested that a consumer’ s
Aad may be influenced by one or more factors inherent in the consumer. Some of
these individual factors — attitudes toward advertising in general, attitudes
toward the advertiser, and the individual’s moods — are hinted at in Figure 1,
which is adapted from Lutz (1985). MacKenzie and Lutz (1989) tested a more
elaborate structural model of Aad formation, which included five specific
antecedent variables: (1) ad credibility, defined as the extent to which the
consumer perceives claims made about the brand in the ad to be truthful and

believable; (2) ad perceptions, defined as a multidimensional array of consumer

14

perceptions of the advertising stimulus, including executional factors but
excluding perceptions of the advertised brand. The underlying determinants of
ad perceptions are said to be the actual executional characteristics of the ad
stimulus and the individual’s attitude toward advertising in general; (3) attitude
toward the advertiser, defined as a learned predisposition to respond in a
consistently favorable or unfavorable manner toward the organization
sponsoring the ad; (4) attitude toward advertising in general; (5) mood, defined
as the consumer’ s affective state at the time of exposure to the ad stimulus.
MacKenzie and Lutz (1989) do not define the impact of mood on Aad very
clearly, merely stating that ”the essential character of the mood determinant of
Aad is that it is an affective state that influences Aad-fl They go on to state that
”[due to] the inherently low levels of involvement associated with advertising
exposure, mood may become associated directly with a stimulus object [i.e., the
ad]” rather than having its impact mediated by cognitive activity wherein the
nature of information processing is influenced by mood (MacKenzie and Lutz
1989, p. 54).

In addition to Lutz and his colleagues, other researchers have established
that ad cognitions (cognitive responses toward particular aspects of an ad) may
directly influence Aad (Muehling and McCann 1993). In the literature, ad
cognitions have most commonly been measured through an open-ended thought
listing procedure; the number of negative thoughts are subtracted from the
number of positive thoughts to yield a net ”ad cognitions” score. This ad
cognition score was hypothesized to be positively correlated with Aad
(MacKenzie and Lutz 1989).

Several researchers have suggested that individual affective, emotional
and non-ad-related responses evoked at the time of ad exposure may influence

Aad (Muehling and McCann 1993). This conceptualization suggests that an

ind

COl

ant

CC

81'

15
individual’s emotional response(s) to an ad is a distinct construct from Aad, a

contention supported by Batra and Ray (1986a), Burke and Edell 1989, and Stout
and Leckenby 1986, among others.

Other individual factors that have been posited to affect Aad include
individuals’ prior brand attitudes and prior Aad (Edell and Burke 1987),
utilitarian and non-utilitarian brand beliefs (Mittal 1990), age of respondents
(Freiden 1984), and education of respondents (Macklin, Bruvold and Shea 1985).
b. Ad-related antecedents

A number of studies have examined the relationship between various
characteristics of the ad, and Aad. The most commonly researched ad
characteristics have been the use of humor, use of celebrities, use of music, the
context of viewing the ad (number of competitive ads, ad sequence, program
environment, program involvement, effect of previous ads), and ad content
(number of arguments, imagery, complexity, use of visuals, message quality,
message sidedness, claim strength, sexual appeals, number of exposures,
distinctiveness and likability).

Researchers have found that the use of humor can enhance consumers’
Aad, although the effects of humor may be moderated by prior evaluations of the
advertised brand and / or individuals’ processing goals. Likewise, the use of
celebrities may also have a positive effect on Aad. Researchers have also found
interactions between source credibility and message-sidedness (i.e., the use of
one-sided versus two-sided messages) (Muehling and McCann 1993, p. 46).
While Kamins (1989) reported that two-sided non-celebrity ads resulted in the
most favorable Aad, Hastak and Park (1990) found no direct message-sidedness
effect on individuals’ Aad-

In addition to the effects of humor and celebrity endorsers, researchers

have investigated the impact of many other ad-content related variables on Aad-

16
Batra and Ray (1986b) found that ads containing few message arguments yielded

more favorable Aad than ads containing many arguments. Cox and Cox (1988)
and Zinkhan and Martin (1983) found some support for the proposition that
complex ads may be more positively evaluated than simple ads. The effect of ad
complexity is moderated by the number of exposures and the complexity of the
receivers.

The context of viewing the ad has also been found to influence Aad- Ads
placed in television programs evoking positive or happy feelings resulted in
more favorable attitudes toward the ad (Kamins, Marks and Skinner, 1991;
Villareal 1985). Soldow and Principe (1981) found that Aad was enhanced when
ads were placed in low involvement programs. Machleit and Wilson (1988) have
emphasized that researchers investigating the effects of television commercials
should be conscious of the effects of the program context.

c. Other antecedents

Apart from individual and ad related elements, researchers have
investigated the effects of delay (which will be discussed in greater detail later in
the chapter), involvement and product novelty on Aad. Thorson and Page (1990)
found that ads for high involvement products produced more positive Aad than
ads for low involvement products. Cox and Locander (1987) found Aad to be
more positive for ads featuring novel products, compared to ads for familiar
products.

2.2.2 Effects of Aad

In keeping with Muehling and McCann’s (1993) typology, the following
section will briefly outline the findings in the literature on the cognitive, affective
and behavioral effects of Aad-

17
a. Cognitive Effects

There are several studies documenting the cognitive responses that are
influenced by Aad- Aad has a direct impact on brand attribute beliefs (Hastak
and Olson 1989), belief strength and belief confidence (Droge and Darrnon 1987).
Aad influences perceptions of ad credibility and persuasiveness (Gelb and Pickett
1983). It also has an impact on brand cognitions (Homer 1990; Muehling,
Laczniak and Stoltman 1991 ), brand recall, brand recognition and fact recognition
(Zinkhan, Locander and Leigh 1984).

b. Affective Effects

The most common criterion variable that has been studied in the Aad
literature is attitude toward the brand (Ab). Muehling and McCann (1993) report
that at least 37 studies reviewed by them supported the notion that Aad has a
direct influence on Ab under a variety of conditions. Other studies have
investigated the effects of Aad on attitude toward purchasing the brand. Leigh,
Rethans and Whitney (1987) and Muehling (1987) found that there is a positive
relationship between Aad and attitudes toward purchasing, but Madden,
Debevec and Twible (1985) found no such effect.

c. Behavioral Effects

In the literature on Aad, the most commonly studied behavioral criterion
variable has been purchase intention. (Muehling and McCann (1993) found at
least seven studies that reported that positiveAad tends to produce astronger
intention «to buy the advertised brandFIOther behavioral effects that have been
studied include brand interest, brand consideration, viewing time and repeat
purchase.

2.2.3 Moderators of Aad
Researchers on the effects of Aad on various criterion variables have

recognized that such effects are moderated by a variety of factors. The most

18

commonly recognized moderating variable is involvement. As seen in Figure 1,
Lutz’s (1985) model posits that the effect of Aad is dependent upon two kinds of
involvement: (1) the consumer’ 5 advertising message involvement, and (2)
advertising execution involvement. Among others, Petty, Cacioppo and
Schumann (1983) have suggested that the relationship between Aad and Ab is
likely to be strong under conditions of high involvement, and weak under
conditions of high involvement. This is consistent with the ELM
conceptualization that advertising stimuli are primarily aids to peripheral, rather
than central, processing of product-related information.

On the other hand, Muehling and Laczniak (1992) have suggested that the
effect of Aad on Ab is fairly robust across involvement levels. However, the
cognitive and affective responses preceding Aad are likely to have differential
effects on Aad, depending upon individuals’ level of involvement (Muehling and
McCann 1993).

The framework of ELM was also applied by Keller (1991) to investigate the
moderating effects of ad processing goals and retrieval cues. When consumers
are engaged in a brand-processing task, the effect Aad is highest when
accompanied by both ad execution and brand-related cues. When consumers
engage in non-brand processing, the effect of Aad is highest when accompanied
by ad execution cues.

Another moderating variable that has received some attention from
researchers is brand familiarity. Machleit and Wilson (1988) hypothesized that
Aad would not have a significant effect on Ab when the effect of prior brand
attitude was controlled. Their data supported this model. However, Edell and
Burke (1986) found that Aad had a significant effect on A1,, even for familiar
brands, although the effect was greater for unfamiliar brands. Phelps and
Thorson (1991) also reported that Aad has a statistically significant effect on A5,

19
even for familiar brands, although the effect of Aad is indeed attenuated by prior

brand attitude.
2.3. MEMORY AND ATTITUDE TOWARD THE AD

Of the 95 published studies listed in the literature review by Muehling and
McCann (1993), only three have dealt with the impact of delay (passage of time)
as a variable moderating the Aad —> Ab relationship. Toward the conclusion of
their review, Muehling and McCann (1993) have suggested that "memory based
explanations of Aad effects should be given further consideration in future Aad
studies (p. 53)." As has been discussed in Chapter 1, a study of the delayed
effects of advertising is especially important in the light of the recent interest in
brand equity. This also underlines the need to study the interaction of delay with
initial attitudes toward the ad, in producing brand attitudes. Accordingly, this
section deals with the studies relating memory and Aad- The important results
from the three relevant studies (Moore and Hutchinson 1983, 1985;
Chattopadhyay and Nedungadi 1992), and two other related studies (Thorson
and Friestad 1989; Thorson, Chi and Leavitt 1992) are first summarized below. A
detailed review of the three individual studies is then presented.
2.3.1. Immediate and Delayed Effects

Moore and Hutchinson (1983, 1985) tested the impact of delayed
measurement on the relationship between Aad and persuasion variables
(purchase intention, measured as "change in brand consideration"). Print ads
were used as stimuli. Empirical support was found for the following two

hypotheses:

1. Brand awareness and ratings of emotional reactions to ads (Aad) are
curvilinearly related such that both negative and positive ads produce
greater increments in brand awareness than neutral ads.

20

2. Attitude toward the brand (Ab) and rated purchase likelihood are linearly
related to ratings of emotional reactions to the ad (Aad) immediately
following exposure, but curvilinearly related for delayed measures.

Taken together, the two studies by Moore and Hutchinson (1983, 1985)
seem to provide limited support for a familiarity-based sleeper effect (see detailed
discussion below). One somewhat surprising finding was that, if measured after
a delay, negatively evaluated ads produced a more positive Ab than neutral ads.
This finding appears to fly in the face of a classical-conditioning-based
explanation of the effects of Aad- Moore and Hutchinson (1985) suggest that ads
producing greater affective responses (either positive or negative) are more likely
to be encoded in memory than neutral ads. Over time, the association between
the negative ad and the brand weakens, leaving a stronger memory trace (akin to
brand familiarity) for the brand with the negative ad, compared to the neutral ad.
Moore and Hutchinson (1983, 1985) did not use brand recall as an explicit
variable in their study. However, recall clearly plays a key role in their
hypotheses about the role of memory in explaining the impact of Aad on Ab and
persuasion.

A study by Chattopadhyay and Nedungadi (1992) provides an extension
of Moore and Hutchinson's (1983, 1985) work. Chattopadhyay and Nedungadi
(1992) contrasted the effect of two types of television commercials (likable and
neutral) for the same product, on the number of ad and brand cognitions
generated among subjects. Measures were taken immediately after exposure, as

well as seven days later. The relevant findings from this study were:
1. The more likable ad resulted in fewer brand cognitions than the neutral ad.

2. The number of brand-directed cognitions was more resilient to decay over
time than the number of ad-directed cognitions.

3. Ad type had a significant effect on Aad immediately following exposure, but
' not after a delay. Specifically, while Aad declined over time for those

21

exposed to the likable ad, there was an increase in Aad over time for those
exposed to the neutral ad.

4. Ad type had a significant effect on Ab immediately after ad exposure,
regardless of the level of attention paid to the ad.

5. In the case of the delay group, (a) ad type had no impact on A}, for the high-
attention group; (b) the likable ad resulted in a lower Ab than the neutral ad
for the low-attention group. (The researchers suggest that this is because in
the likable-ad, low-attention condition, neither ad attitude nor brand
cognitions are accessible.)

2.3.2. EgotionahRes onse and Ad Recall
In a related study, Thorson and Friestad (1989) tested the impact of

Emotional responses to advertisingflonad, message recall. It should be noted that

"""M --wrW~r.Am~¢"” "WA-rm

 

Thorson and Friestad make a clear distinction between Aad, and the valence and
intensity of emotions experienced by viewers while watching a television
commercial. They contend that Aad measures, as found in the literature, are
highly cognitive in nature and require the operation of semantic memory,
whereas memory for television commercials is encoded in episodic memory.
Episodic memory is defined as the mental storage of personal experiences, and
their spatial and temporal context. Semantic memory is the mental storage of
general knowledge/”Thorson and Friestad's (1989) study provided support for
the following hypdtheses:

1. Emotional commercials are more likely to be recalled than commercials
inacfompanied by emotion.
2. The stronger the emotion generated, the greater its effects on memory will

be.

3. Strong emotional commercials are more likely to be recalled before weaker
emotional commercials, or those failing to engender any emotional
response.

4. The kind of strength of an emotional response experienced during a
commercial is likely to serve as an organizer of recall, particularly in the

22

absence of other reasonable organizing principles (such as similarity of
products and their attributes).

The criterion variable tested by Thorson and Friestad (1989) was ad
message recall. Recall was tested immediately after ad exposure, using a free
recall procedure. Of course, this raises the question of whether recall is in fact a
valid predictor of any other criterion variable of interest to marketers (such as
attitude change or purchase intention). This issue has been widely discussed in
the literature on advertising c0pytesting. In their review of the theoretical

foundations of copytesting, Stewart, et al. (1985) make the following comment:

Recall of a unique aspect of a commercial may or may not indicate whether
the viewer will associate the product with the desired usage occasion or
emotion. That is, viewers may remember having seen an ad and yet not feel
positively about the advertised product, even though the ad itself
communicated a positive message. People may not believe the
advertisement's claim, or may simply ignore this claim. In fact, an ad may
be remembered because it was aversive or ridiculous. In other words, a high
recall score does not necessarily imply that the ad was persuasive or that it even
promoted a positive attitude toward the brand. (Stewart, et al., 1985, p. 20,
italics added).

In the View of the controversy surrounding recall measures, it is
interesting to see that the findings of Thorson and Friestad (1989), who used
recall measures, seem to closely parallel those of Moore and Hutchinson (1985),
who used persuasion measures. Further, Stewart (1986) has reported that ads
leaving stronger memory traces and having more brand differentiating impact
are more likely to produce greater persuasion. Thorson, Chi and Leavitt (1992)
have sought to explain Stewart's (1986) finding in terms of the memory "engram"
created by emotional ads. They posit that when an ad creates emotion in the
viewer, the memory engram for the experience is enhanced over non-emotional
conditions. In their empirical study, Thorson, Chi and Leavitt (1992) report that

there was a clear linkage between memory (recall) and attitudes for emotional

23
ads, but found no such linkage for unemotional ads. Page, Thorson and Heide

(1990) tested the hypothesis that emotional commercials were more likely to be
recalled than neutral commercials. The hypothesis was not supported by their
data: the differences in recall scores were in the expected direction, but were not
statistically significant.

Like Thorson and Friestad (1989), Edell and Moore (1993) view Aad and
ad-induced feelings as distinct constructs. They studied the immediate and
delayed effects of ad-induced feelings on Aad and Ab. Aad and Ab were
measured as criterion variables, which are determined by ad-induced feelings.
Of the four hypotheses tested by Edell and Moore (1993), the following two are
relevant to this study:

1. Ad-induced feelings and brand claims can be recalled equally well.

2. There will be no difference in the effects of ad-induced feelings on Aad and
Ab, regardless of whether the effects are measured immediately after
exposure to the ad, or following a May delay.

The first hypothesis was supported by the data. The second hypothesis
was only partially supported. Ads that produced upbeat, uneasy and negative
feelings had the same impact on Aad regardless of whether measures were taken
immediately following exposure, or after a three day delay. However, the effect
was found to decay in the case of ads producing warm feelings (Edell and Moore
1993, p. 205). (Similar results were reported when Ab was used as the criterion
variable.)

The foregoing discussion shows that, while there has been research in the
past investigating the impact of ad affect on recall and persuasion, there is need
for further research to extend and synthesize these findings. In the following
section, a detailed summary is provided of the two studies by Moore and
Hutchinson (1983, 1985), and the study by Chattopadhyay and Nedungadi (1992)

24
that provide the impetus for some of the hypotheses proposed for the present
study.
2.3.3. Details of three memory-related studies
This subsection provides a detailed review of three studies pertaining to

memory and its impact on the Aad -—> Ab relationship. The objective of

M .4...- “ “fin-“M.-

g—on-A'w

providing this detailed summary is twofold: (1) to illustrate some of the
methodological approaches that have been followed in studying the relevant
variables, and (2) to focus on some of the limitations of previous research, which
can be overcome by replication and extension.

a. Moore and Hutchinson (1983)

Moore and Hutchinson tested five hypotheses concerning the effect of Aad
on Ab. The first three hypotheses assume that the immediate and delayed effects
of Aad do not differ. The last two hypotheses are alternative formulations, and
make totally contradictory predictions. An empirical validation of the
hypotheses was sought. In the discussion below, the hypotheses are represented
graphically to aid comprehension.

1. Generalization Hypothesis. According to this hypothesis, affective reactions to
the ad are associated directly with the brand through a conditioning
process. Ab should increase linearly with Aad (see Figure 2).

2. Dis traction Hypothesis. The assumption here is that ads which elicit strong
emotional responses will attract so much attention that they will inhibit
brand-related cognitive processing (i.e., distract the consumer from the
brand). If distraction is the mediator of ad effects, then ads eliciting strong
affective reactions, regardless of valence, should impair brand memory and
attitude change. This hypothesis suggests that Ab is an inverted-U function
of Aad (see Figure 3).

3. Distinctiveness Hypothesis. If strong affective reactions to advertising increase
memory for advertising, then Ab may be more favorable for brands
associated with ads eliciting intense affective reactions, relative to ads
eliciting little or no affective reaction. An implicit assumption of this
hypothesis is that reactions to the ad and reactions to the brand are separate

25

in memory. This hypothesis suggests a J-shaped or U-shaped relationship
between Aad and Ab (Figure 4).1

Figure 2.
Relationships predicted by the Generalization Hypothesis.

Brand Affect (Ab)

r I I

 

 

I r T
Negative Neutral Positive
Ad Affect (Aad)

4. Familiarity-based "Sleeper" Hypothesis. This hypothesis predicts that
immediately after exposure to the ad, there will be a linear relationship
between Aad and Ab (i.e., the generalization hypothesis will hold), but after
a delay, the direct effect of Aad will decay. Over time, the "indirect"
influence of brand familiarity will play a greater role in determining Ab
(similar to the distinctiveness hypothesis). Thus, consumers will feel more
positively toward brands that they remember, regardless of whether they
initially liked the ads or not. An implicit assumption here is that ads
producing extreme affective reactions — whether positive or negative — are
likely to be better remembered than neutral ads (see Figure 5).

5. Afi'ect-Based ”Sleeper" Hypothesis. The predictions made by this theory are
exactly the reverse of those made by the Familiarity-Based Sleeper
Hypothesis. The rationale underlying this hypothesis is that Aad and Ab
can initially be separated in memory. Further, it is assumed that Aad has
little influence on Ab immediately following exposure. Instead, initial
brand evaluations are based upon brand attributes and brand familiarity.
Therefore, if ads eliciting a strong affective reaction are attended to more

 

1 Moore and Hutchinson do not quite explain why the curve should be I—shaped rather than U-
shaped. The implicit thinking seems to be that, while disliked ads will be remembered better
than neutral ads, and thus create a more positive brand attitude than neutral ads, ads which
are liked will also be remembered, and will create an even more positive brand attitude than
disliked ads which are remembered.

26

than neutral ads, one would expect a J-shaped relationship between Aad
and Ab immediately following exposure. After some delay, this theory
assumes that Aad and Ab may become confused in memory, and
consequently, Ab will be a linear function of Aad (see Figure 6).

Figure 3.
Relationships predicted by the Distraction Hypothesis.

 

 

3

S.

E

u

r:

E

m
l t l
Negative Neutral Positive

Ad Affect (Aad)

Moore and Hutchinson tested the alternative hypotheses by showing print
ads on slides to experimental subjects. Ad affect was measured rather than
manipulated. Using "change in brand consideration"2 as a measure of Ab, the
researchers found empirical support for the Familiarity Based Sleeper
Hypothesis. However, when they used a measure of brand knowledge ("change
in pr0portion of correct brand-product associations") as the dependent variable,
only the seven-day delay group showed the U-or J-shaped curve that was
expected.

 

2 The difference in pre-and post-exposure scores to a question regarding subjects' willingness to
consider a particular brand if they were in the market for the product category associated
with it.

27

Figure 4.

Relationships predicted by the Distinctiveness Hypothesis.

Brand Affect (Ab)

 

 

r I .
Negative Neutral Positive
Ad Affect (Aad)

Figure 5.

Relationships predicted by the Familiarity-Based "Sleeper" Hypothesis.

Brand Affect (Ab)

 

’0 Immediate

 

r I I
I I '

Negative Neutral Positive
Ad Affect (Aad)

28

Figure 6.
Relationships predicted by the Affect-Based "Sleeper" Hypothesis.

’0 Delayed

Immediate

Brand Affect (Ab)

 

I I I
I I '

Negative Neutral Positive
Ad Affect (Aad)

 

b. Moore and Hutchinson (1985)

Extending their earlier work, Moore and Hutchinson tested the following

hypotheses:

1. Brand awareness and ratings of emotional reactions to ads (Aad) are
curvilinearly related such that both negative and positive ads produce
greater increments in brand awareness than neutral ads.

2. Attitude toward the brand (Ab) and rated purchase likelihood are linearly
related to ratings of emotional reactions to the ad (Aad) immediately
following exposure, but curvilinearly related for delayed measures.

3. Prior familiarity or awareness of a brand attenuates the indirect effect of ad
affect (Aad) on brand attitude (Ab) by creating equal levels of brand
awareness for positive, neutral and negative ads. This hypothesis predicts a
main effect for prior exposure to brand names and an interaction between
prior exposure and ad affect for brand awareness measures.

29

4. The amount of variance in brand attitude ratings accounted for by brand
awareness increases with delay, while the variance accounted for by
emotional reaction to the ad (Aad) decreases with delay. (Variants of this
hypothesis were examined to test specific memory models of how the direct
and indirect effects of affective reactions change over time.)

Print ads projected on a screen were used as stimuli. The overall design of
the study included three independent variables: (1) Training (subjects were
trained to memorize twenty brand names, which were either the same as for the
ads that they would be exposed to, or difl'erent from the ad they were exposed to;
the objective of this "Training" was to induce differential levels of brand
familiarity among the two groups); (2) Delay (two- or seven-days); (3) Ad affect
(Aad), which was measured rather than manipulated. Dependent measures were
taken of (1) affect toward the brand (change in brand liking and change in
purchase likelihood), and (2) brand familiarity / knowledge (change in brand
familiarity, and change in brand knowledge).

With reference to the Aad —> Ab relationship, the findings of this study
were similar to those of Moore and Hutchinson (1983). The data seem to provide
limited support for a familiarity-based sleeper effect. However, the researchers
noted that the expected interaction (between Aad and delay) was limited to
extreme values of Aad- They offer two possible explanations for this: (1)
Although the ads were pre-selected to be affectively extreme, the print ads
nevertheless failed to produce genuinely extreme affective responses; (2) The
experimental manipulation undertaken by the researchers in the form of
"Training" (see previous paragraph) possibly sensitized all the subjects to seek
out brand names in ads, and did not have its intended effect of selectively
maximizing brand familiarity.

However, the relationship between Aad and the two brand familiarity/

knowledge variables did not exhibit the patterns anticipated by Moore and

30
Hutchinson. For subjects highly familiar with the advertised brand names, they

had expected Aad to have no impact on brand familiarity, but for unfamiliar
subjects, brand familiarity was expected to be a U—shaped function of Aad.
Rather than measure prior brand familiarity, Moore and Hutchinson sought to
manipulate it through "Training." (Subjects who memorized the same brand
names for which they later saw ads were expected to be highly familiar with the
advertised brands, while subjects who memorized diflerent brand names were
expected to be unfamiliar.) However, this "Training" resulted sensitized
experimental subjects to watch out for brand names. Thus, the empirical
findings were that there was no main effect of Aad on brand familiarity/
knowledge. Further, the expected interaction between Aad and "Training" was
also not observed (Moore and Hutchinson 1985, p. 77-79). I

The third hypothesis, regarding the moderating impact of brand
familiarity on the Aad —> Ab relationship, could not be tested satisfactorily owing
to the unexpected experimental artifacts resulting from the "Training"
manipulation.

The final section of Moore and Hutchinson‘s paper is concerned with
testing three alternative explanations for their fourth hypothesis (see above).3
c. Chattopadhyay and Nedungadi (1992)

In a recent study, Chattopadhyay and N edungadi tested seven memory-
based hypotheses regarding the effect of different levels of attention and delay on
Aad, and on ad- and brand-directed cognitions. They did not explicitly consider

the impact of attention and delay on the Aad —> Ab relationship.

 

3 The three alternative explanations tested were (1) Mutual Exclusion Hypothesis; (2) Affective
Components Hypothesis; (3) Ad Information Hypothesis. None of the three alternative
explanations was consistent with the data obtained, and the authors suggest that a synthesis
of theories should be sought.

31
The experimental design was a 2 (attention: high versus low) x 2 (delay:

immediate versus 7—day) x 2 (ad type: neutral vs. likable) between-subjects
factorial design. The stimulus materials consisted of two alternative versions
(neutral and likable) of a TV commercial for a real (but unfamiliar) brand of pen,
embedded in a 15-minute television program along with other filler ads. The
target ad was always shown last. Attention was manipulated by giving subjects
different cover stories at the beginning of the experimental session. Dependent
variables (Aad , Ab, ad-directed cognitions and brand-directed cognitions) were
measured either immediately following exposure, or after a seven-day delay.

The researchers' relevant findings were: (1) The more likable ad resulted in
fewer brand cognitions than the neutral ad; (2) The number of brand-directed
cognitions was more resilient to decay over time than the number of ad-directed
cognitions; (3) Ad type had a significant effect on Aad immediately following
exposure, but not after a delay. Specifically, while Aad declined over time for
those exposed to the likable ad, there was an increase in Aad over time for those
exposed to the neutral ad; (4) Ad type had a significant effect on Ab immediately
after ad exposure, regardless of the level of attention paid to the ad; (5) In the
case of the delay group, (a) ad type had no impact on Ab for the high-attention
group; (b) the likable ad resulted in a lower Ab than the neutral ad for the low-
attention group. (The researchers suggest that this is because in the likable-ad,
low—attention condition, neither ad attitude nor brand cognitions are accessible.)

2.4. THE SLEEPER EFFECT

Both Moore and Hutchinson (1985) and ChattOpadhyay and Nedungadi
(1992) found support for the existence of a "familiarity based sleeper effect. "
However, the sleeper effect has had a checkered past, as noted by Alwitt and
Mitchell (1985). The label "sleeper effect" is attributed to Hovland, et al. (1953),

who came across the phenomenon in the course of their persuasion studies

32
during World War 2. They found that a persuasive film, designed to induce

positive regard for their British allies among American soldiers, produced greater
attitude change nine weeks after the message, compared to attitude change
measured one week after the message (Petty and Cacioppo 1981, p.89).

Kelman and Hovland (1953) tried to explain the sleeper effect using
message-learning theory. According to the message-learning theory, message
arguments and message cues are separate. Message cues consist of factors other
than message arguments that cause a person to accept or reject an advocacy.
Message cues include augmenting cues, such as an attractive spokesperson,
visually appealing ad, etc., and discounting cues, such as an untrustworthy
spokesperson. Kelman and Hovland’s (1953) dissociative-cue hypothesis holds that
a sleeper effect occurs because a discounting cue is dissociated from the message
conclusion by the passage of time, while the remaining (more slowly decaying)
association between message arguments and message conclusion produces what
appears to be an "awakening" of attitude change (Petty and Cacioppo 1981, p.
90). The same hypothesis is labeled by Pratkanis and Greenwald (1985) as the
difl’eren tial decay hypo thesis.

Gillig and Greenwald (1974) were among the many investigators who
found that attempts to replicate the sleeper effect did not succeed. Pratkanis and
Greenwald (1985) reviewed the history of the sleeper effect, and reiterated the
findings of Gruder, Cook, Hennigan, Flay, Alessis and Halamaj (1978), who
reasoned that a sleeper effect would occur only under certain restrictive
conditions. According to Gruder, et al. (1978), a sleeper effect would occur when
(a) a persuasive message has a substantial impact on attitudes; (b) this change is
totally inhibited by a discounting cue; (c) the cue and message are dissociated
over time; and (d) the cue and the message are dissociated quickly enough so

that the message by itself still has some impact when dissociation occurs.

33
Pratkanis and Greenwald (1985) point out that such conditions (and therefore a

reliable sleeper effect) are likely to be relatively rare in natural advertising
exposure situations. This study will once again explore if a familiarity-based
sleeper effect can be obtained in a somewhat realistic advertising exposure
environment.
2.5. EFFECTS OF REPETITION

Both Moore and Hutchinson (1983, 1985) and Chattopadhyay and
Nedungadi (1992) have suggested that the impact of repetition should be
considered in memory-based explanations of the impact of Aad- Chattopadhyay
and Nedungadi (1992) posit that repetition of the ad will result in greater
accessibility of ad information, and thus have an impact on the number of ad-
and brand-directed cognitions generated over time, as well as on Aad and Ab.
Moore and Hutchinson (1983) have suggested that the direct influence of Aad on
A}, will increase with repetition. Repetition will also increase the familiarity of
the brand, thus enhancing Ab indirectly as well. Accordingly, this section
reviews a few of the known findings on the effects of repetition on Aad and Ab.

The "Generalization Hypothesis" tested by Moore and Hutchinson (1983)
suggests that the Aad —> Ab relationship works through some kind of classical
conditioning mechanism, at least in the short term. Several other researchers
have also mentioned classical conditioning as an explanation for the effect of Aad
on Ab (e.g., Gardner 1985, Gresham and Shimp 1985, MacKenzie, Lutz and Belch
1986, Mitchell and Olson 1981). If the generalization hypothesis is based upon a
conditioning mechanism, then repeated exposures to an ad should strengthen the
Aad —> A}, relationship.

Cox and Cox (1988) studied the impact of repetition on Aad. They
exposed experimental subjects to a print ad for a fictitious new brand of cola.

The independent variables manipulated were ad complexity (low vs. high

~—

34

complexity), and number of exposures (one vs. two exposures). They found that
repetition had a strong positive effect on subjects’ evaluations of complex ads,
and only a slight (and statistically non-significant) effect on their evaluations of
simple ads. The findings also suggested that brand liking improves with greater
ad exposure. However, the researchers acknowledge that these effects need to be
tested over a wider range of repetitions.

Reviewing earlier literature, Batra and Ray (1986a) predicted that, in
general, intermediate levels of message exposure (two or three) should provide
higher message effects (measured by Ab and Purchase Intention) than either very
low or very high levels. They hypothesized that the effects of advertising
repetition would, however, be moderated by consumers' motivation, ability and
Opportunity to respond. Their study provided limited support for the
moderating effects of these variables.

\/ Machleit and Wilson (1988) tested the effect of repetition on the Aad-Ab
relationship, using three levels of exposure to four separate target commercials
(1, 2 and 3 exposures). They hypothesized that correlations between Aad and Ab
should strengthen with repeated exposure to the commercial. However, contrary
to their expectation, repetition was found to have an impact on affect transfer for
only one of the four TV commercials that were tested. They hypothesize, post
hoc, that direct transfer of affect may occur only for unfamiliar brands which are
also low in involvement.

Other studies have found that increased exposure to ads might actually
produce a lower Aad- Burke and Edell (1986) found that consumers who
reported having seen specific television commercials many times usually had a
negative attitude toward those commercials, but this effect varied substantially
from ad to ad. Calder and Stemthal (1980) found that consumers' liking of the

ads for one product decreased with exposure, but evaluations of ads for another

35
product were unchanged with repetition, and actually increased with exposure

when execution was varied.

Schumann, Petty and Clemons (1990) extended the research of Calder and
Stemthal (1980). They distinguished two alternative types of ad variation
strategy — cosmetic variation and substantive variation. Cosmetic variation was
defined as changes in endorsers, ad layouts, typefaces, and so forth, with no real
changes in message content across different versions of an ad. Substantive
variation was defined as changes in message content, with no changes in
cosmetic features across different ads. Schumann, et al. (1990) found that
cosmetic variation produced more favorable Aad and Ab when the consumers
were exposed to the ads under conditions of low personal relevance (peripheral
processing in ELM terminology), whereas substantive variation produced more
favorable Aad and Ab when the personal relevance of the ads was perceived to be
high (central processing in ELM terminology).

Haugtvedt, et al. (1994) found that subjects exposed to a substantive
variation strategy would have greater ad feature recall, greater brand attribute
recall, and more product related thoughts, compared with subjects exposed to a
cosmetic variation strategy. They also found that attitudes formed as the
consequence of exposure to three repetitions of an ad displayed less decay than
attitudes formed after a single exposure. The substantive and cosmetic variation
strategies resulted in slightly more positive attitudes than the same ad repeated
thrice. It should be noted that both Schumann, et a1. (1990) and Haugtvedt, et al.
(1994) used the same set of print advertisements as stimulus materials.
Therefore, the generalizability of their results to broadcast ads needs to be
investigated.

Based on the above review, one is inclined to agree with Cox and Cox

(1988), who observed that the literature on the effects of repetition on ad liking

36

have produced results that have been mixed and inconclusive. This underlines
the need for ongoing research on this topic. The present study will make a
contribution to the knowledge about the tepic by exploring the impact of
repetition on the Aad —> Ab relationship. The specific research hypotheses

flowing from the literature discussed above are presented in Chapter 3.

37

Table 1.
Personal] Individual Antecedents of

Attitude Toward the Ad. a

ons
ztrons:

tti tow
tow
u tarian

orientations
tow
ent

 

a Adapted from Muehling and McCann (1993)
b Number of studies cited by Muehling and McCann (1993)

38

Table 2.
Ad-Related Factors] Antecedents of Aad a

V

Studies b
5

s
a
arm
easantness

veness

 

a Adapted from Muehling and McCann (1993)
b Number of studies cited by Muehling and McCann (1993)

39

Table 3.
Other Factors Influencing Aad- a

V

Studies b

uct
uct v0 vement

 

a Adapted from Muehling and McCann (1993).
b Number of studies cited by Muehling and McCann (1993).

Table 4.
Consequences] Effects of Aad a

Studies b

VBI‘IESS

 

a Adapted from Muehling and McCann (1993).
b Number of studies cited by Muehling and McCann (1993).

40

Table 5.
Moderators of Aad- a

Studies b

2

tive non-com
tive tive
com

cues
evance
uct
e
v tions
ature concrete

 

a Adapted from Muehling and McCann (1993).
b Number of studies cited by Muehling and McCann (1993).

Chapter 3
RESEARCH HYPOTHESES

This study seeks to examine relationships in the following domains:

1. The immediate and delayed impact of affective responses to the ad (Aad) on
affective responses to the brand (Ab), and ad recall.

2. The impact of ad repetition on immediate and delayed measures of Aad, Ab,
and ad recall.

In order to study these effects, this study involved manipulation of: (1)
delay in measurement level (measures will be taken immediately or seven days
after ad exposure); (2) repetition levels (one, or three exposures); and (3) ad type
(positive, negative, and neutral). Attitude toward the ad (Aad) was measured
rather than manipulated. However, in order to ensure sufficient variance in
affective reactions, stimulus materials (ad types) were pre-selected to include
commercials intended to induce positive, neutral and negative Aad- A more
detailed outline of the methodology is provided in the Chapter 4.

3.1. IMMEDIATE VERSUS DELAYED EFFECTS

The literature review in Chapter 2 leads to the formulation of several
hypotheses which can be empirically tested. The first three hypotheses are based
on the findings of Moore and Hutchinson (1983, 1985) and Thorson and Friestad
(1989). First, the distinctiveness hypothesis suggests that ads which produce a
strong affective response are more likely to be recalled, compared to ads
producing a more neutral affective response . This is similar to a hypothesis

proposed by Page, Thorson and Heide (1990).

H1: Ads which produce a more intensive affective response (regardless of valence) will
produce higher ad recall scores than neutral ads, for both no-delay and seven-day

groups.

41

42
Second, the familiarity-based sleeper effect suggests that, in the short run,

Ab is a linear function of Aad- However, after a delay, the relationship between
Ab and Aad will be U—shaped (see Chapter 2, Figure 5).

H2: For the no-delay group, Ab will be a linear function of Aad: with negatively
evaluated ads producing the lowest Ab, neutral ads producing intermediate levels of
Ab, and positively evaluated ads producing the highest Ab.

H3: For the seven-day delay group, Ab will be a U-shaped function of A ad: with

negatively and positively evaluated ads producing higher levels of Ab compared to
neutral ads.

Chattopadhyay and N edungadi (1992) found that the number of ad-
directed cognitions declined more rapidly than the number of brand-directed
cognitions. A greater number of cognitions is expected to facilitate the process of
retrieval of ad-related information from memory. Accordingly, the following
hypothesis is offered:

H4: Ad recall scores will be lower for the seven-day delay group than for the no-delay
group.

The distinctiveness hypothesis would also appear to suggest that the memory
trace left by affectively extreme ads is stronger than the trace left by neutral ads.
Therefore, the following hypothesis is proposed.

H5: The decline in ad recall between the immediate and seven-day delay groups will be

greater for the neutral ad than the negative or positive ad.

3.2. REPETITION EFFECTS
Cox and Cox (1988) have reported that increased exposure to an ad
improves consumers' evaluation of the ad. Other researchers have also predicted
that moderate levels of repetition will increase consumers' liking for an ad (Batra
and Ray 1988). Apart from increasing consumers' liking for an ad through some

form of the "mere exposure" effect (Zajonc 1968), repetition would also be

43

expected to increase the memory for the ad and the brand in consumers' minds.1
Accordingly, the following hypotheses are offered:

H6: Repetition will result in increased consumer evaluation of the ad (i.e., higher Aad
scores), regardless of ad type.

H7: Repetition will result in increased consumer evaluation of the brand (i.e., higher Ab
scores), regardless of ad type.

H3: Increased exposure to the ad will result in higher ad recall, for both no-delay and
seven-day delay groups, regardless of ad type.

Chattopadhyay and Nedungadi (1992) have suggested that repetition
would help increase the accessibility of ad-related information in consumers'
minds. Based on this, one would expect repetition to attenuate the decay in ad
and brand recall caused by delay, as stated in the following hypothesis.

H9: Increased exposure to the ad will attenuate the decline in ad recall scores between the
no-delay and seven-day delay. In other words, the difi‘erence between no-delay and
seven-day ad recall scores is expected to decrease with the number of exposures to
the ad.

The pattern of interactions predicted by this hypothesis is shown graphically in
Figure 7.

Combining the previous findings on the effect of repetition with the
"Distinctiveness Hypothesis (H 1)," it is hypothesized that repetition will interact
with affective responses to the ad to affect recall, such that:

H10: Increased exposure will produce a greater improvement in ad recall scores for
affectively extreme ads than for neutral ads.

The assumption in this hypothesis is that repetition and affective intensity
combine to enhance consumer memory for the ad and brand. Further, this
hypothesis assumes that there is no "maximum threshold" of awareness beyond
which any increase in either affective intensity or number of exposures will be

counter-productive. The pattern of interaction predicted by this hypothesis is

 

1 Most researchers appear to be of the Opinion that it would take more than three exposures for
"wearout" or irritation to set in. Due to experimental constraints, this phenomenon is beyond
the scope of this study.

44
shown graphically in Figure 8. (Note that in Figure 8, the vertical axis measures

change in recall scores produced by repetition, and not absolute recall scores.)

Figure 7.
Expected interaction between repetition and delay.

High
exposure

Low
exposure

Ad Recall! Brand Recall

 

lrnnlrredate 7418!)! delay

Delay Group

An alternative hypothesis, which makes a totally contradictory prediction,
is also possible. If one were to assume that there is a "maximum threshold" of ad
and brand recall, then it is possible that repetition may not improve recall for ads
evoking strong affective feelings. If such a threshold were to exist, neutral ads
might benefit more from repeated exposures in terms of greater recall than

affectively extreme ads. The methodology used to test these hypotheses is

presented in Chapter 4, and the results are presented in Chapter 5.

Figure 8.
Expected interaction between repetition and ad affect.

Increase In edlbrmd recall
due to ad repetition

 

l I
I

l

| l

Negative Neutral Positive
Ad Affect (Aad)

Chapter 4
METHOD

The hypotheses proposed in Chapter 3 were tested using an experimental

design, with television commercials as stimuli. A 12 cell between-subjects

factorial design was used. The 12 cells were made up of 3 ad types (positive,

negative, neutral), 2 exposure level conditions (one versus three exposures), and

2 measurement delay conditions (immediate measurement vs. one-week delay).

4.1. VARIABLES AND SCALES

The following section describes the independent, dependent and moderating

variables that were used in this study.

4.1.1. Independent Variables

The following independent variables were manipulated:

1.

Ad type. After reviewing several hundred television commercials, three
stimulus ads were selected, such that they would induce positive, neutral
and negative affective responses toward the ad (Aad)- (The procedure for
selection of the stimulus ads is described in greater detail in Section 4.3.)
The objective of varying ad type was to overcome the problems encountered
by Moore and Hutchinson (1985), who felt that their stimulus print ads
possibly did not induce genuinely extreme affective responses. However,
pre-selection is still no guarantee that all subjects will experience the same
valence and intensity of Aad- Therefore, Aad was treated as a measured
variable, not a manipulated one. This is consistent with the
conceptualization of Chattopadhyay and N edungadi (1992), who treated ad
type and Aad as separate constructs.

Repetition. Subjects were exposed to the test commercials either one or three
times. The repetition group was given three rather than two exposures, in
order to enable a greater variance in ad effects due to greater exposure. This
is consistent with Machleit and Wilson's (1988) suggestion that researchers
should include enough exposures to a test ad to investigate repetition
effects. It should be noted that it is not uncommon for viewers of television
programs to be subjected to three (or occasionally even more) repetitions of
the same commercial in a half-hour program, especially while viewing
sponsored programs.

Measurement delay. Measures of the dependent variables were taken either

immediately after exposure, or after a delay of seven days following
exposure. This is consistent with the design used by Chattopadhyay and

45

46

N edungadi (1992). Further, the interaction of repetition and measurement
delay (H10 and H11) can be studied better when the seven-day delay
condition is contrasted with measures taken immediately after exposure,
rather than two days after exposure, as done by Moore and Hutchinson
(1983, 1985).

The following independent variable was measured, and not manipulated:

4. Attitude toward the ad (Aad). Although stimulus ads were pre-selected to
ensure a variance in affective responses, these ads could still have induced
varying intensities (and even valence) of affective response among subjects.
Therefore, Aad were treated as a measured variable. As discussed later, this
measure really constitutes a manipulation check for the effectiveness of the
”ad type” manipulation.

4.1.2. Dependent Variables
The dependent variables which were measured, either immediately after

ad exposure, or after a 7-day delay, are:
1. Attitude toward the brand (Ah).

2. Purchase Intention (PI).

3. Ad recall: Both unaided and aided recall were measured.
4.1.3. Moderating Variables

Many researchers have hypothesized that consumers’ level of involvement
(with the advertised product class) will moderate the impact of Aad on Ab (see
Table 1). The moderating effect of involvement has been explained within the
framework of the Elaboration Likelihood Model (Petty and Cacioppo 1981). Aad
is seen as a peripheral cue, which will play a much greater role in determining
Ab under conditions of low involvement, compared to conditions of high
involvement (Lutz 1985). Accordingly, the product class involvement of
experimental subjects were measured in order to study its moderating effect on
the hypothesized relationships, although no specific hypotheses regarding the
effect of involvement were proposed in Chapter 3. Consumers' product class

involvement was measured using a scale derived from the Personal Involvement

47
Inventory (Zaichkowsky 1985). This sub-scale has been previously used and

validated by Phelps and Thorson (1991). The scales that were used to measure

the constructs described above are summarized in Table 6.

Table 6.
Scales Used to Measure Key Constructs.

 

 

A. Prior Brand Familiarity

1. Please indicate if you are familiar with (Target and decoy brands), by circling
only one of the following responses: 8

a. Have never heard of the brand.

b. Have heard of the brand name, but don't know anything else about it.
c. Have heard of the brand name and know what products it relates to.
d. Know a little bit about the brand and the product class

e. Am extremely knowledgeable about the brand and product class.

B. Attitude Toward the Ad

1. Please indicated your feelings about the ad for (Target Brand), by marking the
most appr0priate spot on each of the following scales: b

Good —: — —: —: —: —: —: Bad
Unpleasant —: — —: —: —: —: —: Pleasant
Favorable —: — —: —: —: —: —: Unfavorable
Enjoyable —: — —: —: —: —: —: Not enjoyable
Disliked it —: — —: —: —: —: —: Liked it
Irritating —: — —: —: —: —: —: Likable
Informative -—: — —: —: —: —: —: Uninformative
C. Attitude Toward the Brand

1. Please indicate your feelings about (Target Brand) by marking the most
appropriate spot on each of the following scales:b

Good — —: —: —: —: —: —: Bad
Unpleasant — —: —: —: —: —: —: Pleasant
Favorable — —: —: —: —: —: —-: Unfavorable
Dislike — —: —: —: —: —: —: Like

Poor quality — —: —: —: —: —: —: High quality

 

(Table continued on next page)

 

 

48

Table 6 (continued).

 

D. Purchase Intention

1. If you were in the market for (Product), how likely is it that you would
consider buying (Target Brand)? Please mark the appropriate spot in each
of the following scales. b

Likely —: —: —: —: —: —: —: Unlikely
Probable —: —: —: —: —: —: —: Improbable
Possible —: —: —: —: —: —: —: Impossible
E. Brand Recall
BreeRecall

1. Please list all the brands for which you remember seeing commercials during
the program you just viewed.

AidedRecall

2. Please indicate if you remember seeing commercials for the following brands
during the program you just viewed.

(Both target and decoy brand names)
Yes No

F. Ad Message Recall

1. Please write down as much as you can remember about the ad for (Target
Brand) that was shown during the program that you just viewed.

 

(Table continued on next page)

 

 

49

Table 6 (continued).

 

G. Product Class Involvement

1. Please indicate your feelings about (target and decoy product categories)
along the following dirnensions:C

Important Unimportant
to me —: —: —: —: —: —: —: to me

Of no concern Of concern

to me —: —: —: —: —: —: —: to me
Irrelevant —: —: —: —: —: —: —: Relevant

Very meaningful Means nothing
to me —: —: —: —: —: —: —: to me

Matters to me —: —: —: —: —: —: —: Doesn't matter
Interesting —: —: —: —: —: —: —: Not interesting
Significant —: —: —: —: —: —: —: Insignificant
Boring —: —: —: —: —: —: —: Exciting

 

a Adapted from Moore and Hutchinson (1985).
b Adapted from MacKenzie and Lutz (1989), Machleit and Wilson (1988).
c Adapted from Zaichkowsky (1985).

4.2 SCALE REUABILITTES
The reliabilities of these scales are reported in Table 7. The value of Cronbach’s
alpha for all the items was above 0.9, well above the acceptable limits for most

published studies in consumer research (Peterson 1994).

Table 7.
Reliability Coefficients for Key Measurement Scales.

Variable of items (Standardized)
0.

on

invo vement

invo vement ice cream
vement
vement instant

 

 

50
4.3. SUBJECTS

The hypotheses were tested in an experimental setting, using student
subjects. The subjects were undergraduate students enrolled in communication
and marketing programs at a small state university on the Gulf Coast. The
experimental design calls for a total of twelve cells. An attempt was made to
obtain 20 subjects per cell, but there was a shortfall in a few cases, due to attrition
between sessions (in the case of the delayed measurement condition), or no-
shows.

Subjects were recruited in classrooms, through a voluntary sign-up
procedure. As an incentive for participating in the study, subjects were told that
their names would be enrolled in a sweepstakes, with a color television set as the
prize. A total of 211 subjects completed both parts of the study. Of the 211
subjects, 87 (41.2 percent) were male, and 124 (58.8 percent) were female. The
age of the participants ranged from a minimum of 17 years to a maximum of 53.
The mean age of the subjects was 24 years, and the median age was 22. They
were randomly assigned to one of twelve experimental conditions. Details of the
experimental procedure are described in Section 4.5.

4.4. STTMULUS MATERIALS
4.4.1. Pretesting
The objective of the pretest was to identify (at least) three commercials that
would induce positive, negative and neutral attitudes, respectively, among the
target audience.
Step 1. After viewing hundreds of (15 and 30 second) TV commercials, twenty
one were shortlisted as candidates for testing. The criteria used for shortlisting
were: (1) the commercials should advertise products, rather than services (a
stipulation borrowed from Machleit and Wilson 1988; the rationale for this

criterion is that it is more valid to expect purchase decisions for products to be

51

made on the basis of advertising, compared to services); (2) the experimental
subjects (i.e., students) should constitute a valid target market for the products
advertised; (3) the brands should be unfamiliar, or only moderately familiar, to
the target audience, to rule out confounds arising from brand familiarity.

Step 2. The twenty one commercials were pretested among a sample of forty nine
undergraduate students. The students were split up into two groups. The first
group (n = 23) was shown ten commercials, and the second group (n = 26) was
shown the remaining eleven commercials. The purpose of "splitting the
commercial pool among two groups was to avoid respondent fatigue.

Before viewing the commercials, respondents answered a question
indicating their familiarity with 26 brands, including the brands featured in the
test commercials. A few well known national brands (such as Folger’s, Wrigley
and Frito Lay) were included in the list to help ”calibrate” their responses. Next,
respondents answered a question reporting their purchase likelihood for 22
product categories, comprising the categories represented in the test
commercials.

After answering the first two questions, respondents were shown the test
commercials one at a time, and asked to immediately record their attitude
toward the ad (Aad), attitude toward the brand (Ab) and purchase intention (PI).
This procedure was repeated until the respondents had been exposed to all the
commercials.

I The results of the pretest are presented in Table 8. Attitude toward the Ad
and Purchase Intention were measured on multi-item 7-point semantic
differential scales, which were then averaged to give the reported score (see
Table 6). Brand familiarity was measured on a 5-point scale. In Table 8, the
brands are arranged in ascending order of the mean Aad score. Based on the

pretest, the following commercials were chosen as stimulus ads for the final

52
experiment: (1) Krunchers Potato Chips (negative Aad), (2) Arnold Bakery Light

Bread (neutral Aad), and (3) Prior Instant Meals (positive Aad)-

Table 8. Pretest Results.

Category Mean Mode Famlllarltyl (Precinct)2

2.1 86
2.286
2.885
2.922
3.267
3.659
3.962

4.422
4.507
4.681
4.681
4.851
4.851
4.851
4.874
4.968
5.01 1
5.198
5.609
5.875

1.129
2.000
1.000

3.571
1 .000
3.286

7.000
4.571
4.571
2.714
4.000
5.571
5.286
6.429
4.429
4.286
5.286
6.143
7.000

1.584
2.288
4.084
1 9
3.703
3.388
4.061

3.146
2.490
3.967
3.233
4.280
1.099
1.159
3.174
1.060
3.352
3.127

4.239

3.662
4.851
5.377

6.796
6.323
6.861

4.904
3.970
2.450
5.491
4.073
5.388
5.784
6.323
5.288
4.383
4.181

6.316

 

Although Claritin produced a more negative Aad than Krunchers, potato

chips are more universally purchased by the target audience than allergy

medication (see PI scores in Table 8). The Gulden’s Mustard commercial also

produced a more negative Aad score than Krunchers, but it was a 15-second spot,

whereas most of the other spots (including Krunchers) were 30 seconds long.

 

1 The brand familiarity and product purchase intention scores have been averaged out across the

two groups (n = 49).

2 The PI scores reported here indicate respondents’ (pre-exposure) likelihood of purchasing the
product category itself (e.g., ”Allergy Medication”) and not the specific brand advertised.

53

The Mentos commercial was ruled out on the grounds that the target audience
was too familiar with the brand. Hence, the Krunchers commercial was selected
for the ”negative Aad” category.

In the ”neutr ” Aad category, the possible candidate commercials were
Sasson (apparel), Payday (candy) and Arnold Bakery Light (bread). Of these,
Payday appeared to have a fairly high brand familiarity among the respondents,
which made it less desirable as a stimulus brand. The Sasson commercial also
seemed to be well qualified, but it had some executional idiosyncrasies that made
it (in the researcher’ 5 opinion) a less suitable candidate commercial than the
Arnold Bakery Light commercial.

Selection of the candidate commercial for the positive Aad category posed
relatively fewer problems. The ad that was evaluated the most positively in the
pretest was a spot for Evian bottled water. However, the brand familiarity scores
for Evian were extremely high. Therefore, the ad for Prior Instant Chicken was
chosen as the stimulus commercial for the positive Aad condition.

The differences in pairwise Aad means were examined using a t-test. The
results showed that there was a statistically significant difference between
pairwise means at an alpha level of 0.05. This confirms that the stimulus ads
were indeed suitable for inducing the desired affective responses among subjects.
The scripts for the three stimulus commercials are attached in Appendix 2.

4.4.2. Stimulus program

The stimulus ads were embedded in a 30-minute television program, in
order to provide for a realistic ad exposure environment. The program was a
documentary titled Southern Voices, Southern Words, which had been produced
for public television by one of the faculty members at the university where the
experimental study was conducted. Although the documentary had been

screened once or twice on regional PBS networks, it was expected that most of

54

the target audience would not have seen the program previously. Machleit and
Wilson (1988) have suggested that the program used for embedding commercials
in a test setting should be relatively neutral in affect, to avoid confounding effects
caused by affective reactions to the program itself. The selected documentary
satisfied this condition by providing a relatively neutral program environment.
The stimulus materials were developed very carefully, so as to realistically
approximate the viewing environment of many network television programs.
This was in response to several calls in the literature urging consumer
researchers to pay attention to ecological validity (D’Souza and Rao 1995,
Pechmann and Stewart 1989). Three commercial breaks were introduced into the
program, with three commercials in each pod. The first break was inserted
approximately ten minutes into the program, the second twenty minutes into the
program, and the final break was inserted just before the final credits. The
sequence of programming and commercial breaks is shown visually in Figure 9.
The target (stimulus) commercial was shown either once (single exposure
condition) or three times (repetition condition). For the single-exposure
condition, the target commercial was placed at the very end of the last
commercial break This placement was consistent with the practice followed by
ChattOpadhyay and Nedungadi (1992). It should be noted here that the
placement of the stimulus commercial in a pod gives rise to questions of
sequencing effects. Placing a commercial at the beginning of a pod could give
rise to primacy effects, and placing it at the end of a pod could give rise to
recency effects. It is beyond the scope of this study to address these issues.
However, the impact of sequencing effects should have been the same across all
three ad types in the single-exposure condition, as the target commercial was

placed at the end of the third pod, regardless of ad type.

55

Figure 9.
Sequence of commercial breaks.

 

I Opening Credits + 10 minutes of program I

l

l Commercial break 1: I

 

3 Filler ads: (Sharp ViewCam, Arrid
Deodorant, Archer Daniels Midland)

l

I 10-minute program segment I

l

Commercial break 2:
3 Filler ads: (Alamo Car Rentals,
Sensodyne Toothpaste, Robitussin Cough

Syrup )

l

l 10-minute program segment

i

Commercial break 3:
2 Filler ads (Hoover Vacuum Cleaners,
Zenith AVI Color TV) + Target Ad
(Krunchers, Arnold Bakery Light, or Prior)

 

 

 

 

 

 

 

 

Closing credits

 

MB: In the 3-exposure (repetition) condition, the Arrid commercial in Break 1, and the
Sensodyne commercial in Break 2 were replaced with the target commercial. (In the
seond commercial break, the target ad ran first, the Alamo ad second and the
Robitussin ad third.)

Along with the target commercial, subjects also saw eight filler
commercials. The filler commercials were chosen according to the following

criteria: (1) they were relatively neutral commercials; (2) there was a mix of

56

commercials for familiar and unfamiliar brands. For the three-exposure
(repetition) condition, the third exposure was placed in the middle of the first
commercial pod, the beginning of the second pod, and at the very end of the
third pod (see Figure 9).

4.5. PROCEDURE

After arriving at the experimental venue, subjects were informed that they.
would be taking part in a research project to learn more about their television
viewing and purchasing habits. They were also told that the study would solicit
their opinion about a program produced by the documentary institute at their
university. Confidentiality of data was guaranteed. However, as an incentive,
all subjects who completed both parts of the study were eligible to have their
names entered in a sweepstakes, with a color television set as the prize.

First, subjects answered a few decoy questions pertaining to their
television viewing habits, and about the subject matter of the documentary.
Second, they answered a battery of questions measuring their familiarity with
various brand names, including the target brands. Third, they answered
questions indicating their purchase intentions — and involvement — with regard
to several product categories, including the categories pertaining to the target
brands. (A copy of the questionnaire is found in Appendix 1. It should be noted
that there were six versions of the questionnaire, with the brand name, category
and instructions modified according to the target ad and delay condition.) Most
subjects took ten to twelve minutes to complete the first part of the study. After
completing these questions, subjects viewed the 30-minute documentary with the
commercials embedded in it. They were exposed to the target commercial either
once or three times.

After viewing the program, subjects were given a distracter task, where

they answered a few questions measuring their responses to (and involvement

57
with) the documentary they had just seen. Subjects in the 7-day delay condition

were then excused, and asked to return one week later. To disguise the purpose
of the study, subjects in the 7-day delay condition were told that they might have
to view another documentary at that time. In reality, when the subjects returned,
they were asked to complete the post-exposure questionnaire. Subjects in the no-
delay condition were given the post-exposure questionnaire immediately after
they had completed the distracter task.

The post-exposure questionnaire consisted of three parts. The first part
measured unaided and aided recall of the target ad. It also measured if the
respondent could successfully retrieve the message in the target ad. The second
part measured the subjects’ attitudes toward the target brand (Ab) and purchase
intentions. After completing the second part, subjects were then shown the
target ad once again, and asked to complete the third part of the post-exposure
questionnaire. The third part of the questionnaire served as a manipulation
check, and measured subjects’ attitude toward the ad (Aad). Subjects were also
asked to respond to 20 Likert items indicating their reactions to the target
commercials. These Likert items were adapted from the Viewer Response Profile
(VRP) scale (Coulson 1989, Schlinger 1979). The objective of these questions was
to get a better understanding of the dimensions of Aad. Questioning during the
debriefing session after the study revealed that there had been no hypothesis-
guessing. In the debriefing session, students were informed about the true
purpose of the study, and given the option to withdraw their questionnaires
from the data used for analysis. None of the subjects Opted to withdraw their
questionnaire. The results of the study with respect to the hypotheses will be

discussed in Chapter 5.

Chapter 5
RESULTS

This chapter reports the results of the study. First, the post-hoc measures
of attitude toward the ad (Aad) across the three ad types are compared to see if
the stimulus ads succeeded in producing the desired affective response. The
data are then examined to see if they support the hypotheses proposed in
Chapter 3.

5.1 MANIPULATION CHECK

As noted in Chapter 4, the three stimulus commercials were selected in
order to induce a positive, neutral or negative Aad among respondents. While
the pretest indicated that the three ads would indeed induce the desired affective
responses, a post-hoc measure of Aad was also taken after the completion of the
main experiment as a manipulation check. Attitude toward the ad was measured
on a seven point scale, with 1 representing a very negative attitude, and 7

representing a very positive attitude. The results are reported in Table 9.

Table 9.
Mean Scores for Attitude toward the Ad,
Attitude toward the Brand and Purchase Intentions Across Ad Types.

 

 

 

 

 

 

 

 

Ad Type Attitude Attitude Purchase ‘
(Intended Affective Response) toward toward Intention
the Ad the Brand (PI)
(Aad) (Ab)
Krunchers (Negative) 335* 3.98 3.45
hArnold Bakery (Neutral) 3.90* 4.31 3.43
Prior (Positive) 5.31* 4.33 3.31

 

 

* Significantly different from the scores for the other two ads at the a = 0.05 level.

Consistent with the results of the pretest, the Krunchers Potato Chips commercial
produced the least positive Aad, with a mean score of 3.35. The ad for Arnold

Bakery Light Bread was evaluated a little more positively, with a mean score of

58

59
3.90, and the ad for Prior Instant Chicken received the highest score of 5.31. A

t-test was used to compare the pairwise means across the three conditions. The
Aad score for each ad type was significantly different from the other two at the 95
percent level of significance. The t-test results confirm that the three stimulus
ads did indeed succeed in inducing the desired affective response among the
respondents. The ad type manipulation may therefore be deemed successful.

In terms of the actual scores, the Krunchers ad (negative ad type) was
evaluated only mildly negatively (its Aad score of 3.35 on a scale of 1 to 7 is just
below the mid-point of the scale). The Arnold Bakery ad (neutral ad type) was
evaluated only slightly more positively than the Krunchers ad. The Prior ad
(positive ad type) was evaluated much more positively than the other two ads.

5.2 GLOBAL CORRELATIONS

The data in Table 9 indicate that the variation in ad type produced the
desired variation in Aad scores. The variation in the Ab scores is in the same
direction as the Aad scores, with the negatively evaluated ad producing the least
positive Ab, and the positively evaluated ad producing the most positive Ab.
However, a t-test showed that the difference in Ab scores across the three ad
types was not statistically significant at the 95 percent level of significance.

The direction of the PI scores is not consistent with the direction of the Aad
and Ab scores. The most favorably evaluated ad (Prior) produced the lowest PI
score (3.31), while the least favorably evaluated ad (Krunchers) produced the
highest PI score (3.45) (see Table 9). The mean scores for PI across the three ad
types are clustered together very tightly, with a difference of only 0.14 (on a
7-point scale) between the highest and lowest mean scores, suggesting that the
differences in mean PI scores across the three ads can be attributed to random

differences. This is confirmed by the results of the t-test, indicating that the

60

purchase intention (PI) scores across the three ad types were not significantly
different from each other at the 0.05 alpha level.

Although the hypotheses in Chapter 3 do not specifically include PI as a
criterion variable, the relationship between PI and the other variables was
examined to see if the data were consistent with previous research on the subject.
The data reported in Table 9 seem to indicate an inconsistency in the direction of
Aad and PI. On the other hand, most of the previous literature on the topic has
suggested a positive, statistically significant correlation between Aad, Ab and PI.
In an attempt to investigate this issue further, the correlations between the three
constructs were examined, pooling the data from subjects across all the
experimental conditions. This data indicates that there is indeed a positive and

statistically significant correlation between Aad, Ab and PI (see Table 10).

Table 10.
Correlations between Attitude toward the Ad,
Attitude toward the Brand and Purchase Intention
(Data pooled across all experimental conditions, n = 205).

 

| Variable I Aad [ Ah I

Ab | 0.5726**| — J
PI | 0.3633**| 0.6206*]

 

** Significant at the a = 0.01 level.

5.2.1. Tercile Split

Taken together, the data from Tables 9 and 10 indicate that, while there is
an overall positive correlation between Aad, Ab and PL the differences in Ab and
PI obtained across the three ad types are not large enough to be statistically
significant. An explanation that can help resolve these seemingly contradictory
results is that, while the ad type manipulation achieved its goal of producing
differential levels of Aad, there is a wide variance in the respondents’ Aad scores

within each ad type. To verify this explanation, the data was split into three

61

terciles, according to respondents’ post-hoc Aad scores. This resulted in three
respondent groups: one with low (or ”negative”) Aad (mean Aads 3.50), one
with medium Aad (mean Aad between 3.51 and 4.82), and one with positive Aad
(mean Aad24.83). When the mean scores for A}, and P1 are compared using
these post-hoc ”ad groups,” rather than the predetermined ad types, the results
are consistent with the positive correlations found in Table 10. The mean scores
for Ab and PI for the three ad groups are reported in Table 11. In contrast with
the data in Table 9, it can be seen that the direction of the PI scores is consistent
that of the Ab and PI scores. However, the difference in the PI scores between
Ad Groups 2 and 3 is not statistically significant. In the analysis that follows, the
data will be analyzed according to both ad type and ad group, to examine if the

results are consistent across both types of analysis.

Table 11.
Mean Scores for Attitude toward the Ad,
Attitude toward the Brand and Purchase Intentions across ”Ad Groups.”

 

 

 

 

 

 

 

 

Ad Group Attitude Attitude Purchase
(T ercile Split along Aad scores) toward toward Intention
the Ad the Brand (PI)
(Aad) (Ab)
Ad Group 1 (Aads 3.50) 2.03* 333* 267*
Ad Group 2 (Aad >351 and <4.82) 4.15* 4.28* 3.59
Ad Group 3 (Aad 2 4.83) 5.99* 4.92* 3.97

 

 

* Significantly different from the scores for the other two ad groups at the or = 0.01 level.

5.3 HYPOTHESIS TESTING
5.3.1. Unaided recall across ad types
The study measured both unaided and aided recall of the target ad. To
measure unaided recall, subjects were given a free recall task, which asked them
to list all the brands for which they remembered seeing commercials. After this,

they were given the aided recall task, which asked them to indicate if they had

62

seen a commercial for the brand(s) listed in the questionnaire. To avoid biasing
the respondents, the list included brands from the filler commercials, as well as
various ”decoy” brands, in addition to the target brand (see Question 30 in the
Questionnaire included in Appendix 1).

The open-ended responses to the recall question fell into four categories:
(I) recall of both the brand and product category, (2) recall of the product
category but not the brand, (3) recall of the product category and the wrong
brand, and (4) recall of neither the product category nor the brand. In the
analysis below, recall is treated as dichotomous, i.e. recall was either successful
or unsuccessful. Thus responses from subjects in Categories 2 and 3 are not
reported below.

Based on the distinctiveness hypothesis (Moore and Hutchinson 1983,
1985), H1 predicted that ads inducing a more intensive affective response
(regardless of valence) would produce higher ad recall scores than neutral ads,
for both no-delay and 7-day groups. The data for ad recall for the three ad types
are presented in Table 12. It can be seen from the table that the three different ad
types had substantially different rates of recall, an observation confirmed by a
chi-square test, which was significant at an alpha level of 0.02.

 

 

 

 

 

Table 12.
Unaided Recall of Different Ad Types.
Ad Type Recalled Did Not Total
Ad Recall Ad
Negative (Krunchers) 56 (60.9%) 36 (39.1%) 92
Neutral (Arnold Bakery) 25 (51.0%) 24 (49.0%) 49
Positive (Prior) 18 (35.3%) 33 (64.7%) 51
Trotal 99 (51.6%) 93 (48.4%) 192

 

 

 

 

 

 

Chi-square (d.f. 2) = 8.60 (P <0.02)
N.B.: Does not include data for respondents who recalled just the brand name or
product category.

63

According to H1, both the negative (Krunchers) ad and the positive (Prior)
ad should have had a higher percentage of recall than the neutral (Arnold
Bakery) ad. The data in Table 12 shows that the negative ad did indeed have a
higher percentage of recall (60.9%) than the neutral ad (51.0%), but the positive
ad had an even lower percentage of recall (35.3%) than the neutral ad. Thus, H1
is not supported.

The data was re-analyzed using the post hoc ”ad groups” (tercile split of
post hoc Aad scores) instead of the predetermined ad types (see Table 13). The
pattern of results was very similar to that found in Table 12. However, a
chi-square test showed that the differences across cells were not statistically
significant at the 0.05 alpha level, suggesting that the recall percentages for the

neutral and positive ads are very similar.

 

 

 

 

 

Table 13.
Unaided Recall of Different Ads Classified by Ad Group.
Ad Group Recalled Did Not Total
(Aad terciles) Ad Recall Ad

Negative (Tercile 1) 43 (60.6%) 28 (30.1%) 71‘
Neutral (Tercile 2) 28 (49.1% 29 (50.1%) 57
Positive (Tercile 3) 28 (43.8%) 36 (56.2%) 64
Total 99 (51.6%) 93 (48.4%) 192

 

 

 

 

 

 

Chi-square (d.f. 2) = 4.00 (P <0.15)
N.B.: Does not include data for respondents who recalled just the brand name or
product category.

The pattern of recall scores across the three ad types and ad groups is depicted
visually in Figure 10. The figure indicates a negative linear relationship between
Aad and recall, whereas H1 suggested a U-shaped relationship. Thus H1 is not

supported by the recall scores re-analyzed according to ad group.

64

Figure 10.
Comparison of Unaided Recall by Ad Type and Ad Group.

70

Unaided Recall %
OI
O

 

Recall (Ad Type)
Recall (Ad Group)

 

F3—

 

Positive

Neutral
Ad Typel Group

Negative

5.3.2. Aided recall across ad types

The data for aided recall across ad types and ad groups follows the same

pattern as the data for unaided recall. As might be expected, the percentage of

aided recall is greater than unaided recall across all three ad types and ad

groups. The data for aided recall is presented in Tables 14 and 15. Once again,
the negatively evaluated ad had the highest aided recall (92.6% for Krunchers,
91.4% for Tercile 1), followed by the neutral ad (80.0% for Arnold Bakery, 79.4%
for Tercile 2), followed by the positive ad (56.4% for Prior, 67.2% for Tercile 3).

 

 

 

 

 

Table 14.
Aided Recall of Different Ad Types.
Ad’rype RLecalled Did Not Total

Ad Recall Ad
Negative (Krunchers) 88 (92.6%) 7 (7.4%) 95
Neutral (Arnold Bakery) 44 (80.0%) 11 (20.0%) 55
Positive (Prior) 31 (56.4%) 24 (57.1%) 55
”Total 163 (79.5%) 42 (20.5%) 205

 

 

 

 

 

 

Chi-square (d.f. 2) = 28.14 (P <0.01)

65

 

 

 

 

 

Table 15.
Aided Recall of Different Ads Classified by Ad Group.
Ad Group Recalled Did Not Total
(Aad terciles) Ad Recall Ad
Negative (Tercile 1) 64 (91.4%) 6 (8.6%) 70
Neutral (Tercile 2) 54 (79.4%) 14 (20.6%) 68
Positive (Tercile 3) 45 (67.2%) 22 (32.8%) 67
'TBtal 163 (79.5%) 42 (20.5%) 205'

 

 

 

 

 

 

Chi-square (d.f. 2) = 12.37 (P <0.01)

5.3.3. Overall effect of Aad on A],

It has already been seen that there is a strong (and statistically significant)
positive correlation between Aad and Ab for all subjects taken together (see Table
10). To further examine the overall effects of Aad on Ab, a simple linear
regression analysis was conducted, with Ab as the dependent variable and Aad as
the independent variable. The results are reported in Table 16. The regression
analysis confirms that Aad has a positive and statistically significant impact on
A5. The value of the regression coefficient (R2) is 0.33, indicating that Aad

accounts for 33 percent of the variance in Ab.

Table 16.
Regression Analysis Results: Standardized Regression Coefficients
(Data pooled across all conditions).

 

 

 

 

 

 

 

 

 

 

 

Dependent Aad A}, F R2 n
Variable
Ab 0.57 _ 97.08* 0.33 200
P1 — 0.62 123.32* 0.38 198
** P<0.01

A multiple regression analysis was also conducted with P1 as the criterion
variable, and Aad and Ab as the predictor variables (see Table 16). Using a
stepwise procedure, Aad was eliminated from the model, leaving Ab as the only
statistically significant predictor of PI for the global model. The value of R2 is

0.38, indicating that Ab accounts for 38 percent of the variance in P1.

66

5.3.4 Immediate and Delayed Effects of Aad on A],

Based upon the previous literature on the effects of Aad on A1,, H2
predicted that, for the no-delay group, Ab would be a linear function of Aad, with
negatively evaluated ads producing the least positive Ab, and positively
evaluated ads producing the most positive Ab. The Ab and PI mean scores for

the three ad types split by delay condition are reported in Table 17.

 

 

 

 

 

 

 

 

Table 17.
Mean A}, and PI Scores by Ad Type and Delay Condition.
A], PI
Ad Type No Delay No Delay

delay delay

Krunchers (Negative) 3.85 4.08 3.33 3.55

Arnold Bakery (Neutral) 4.30 4.31 3.15 3.70

Prior (Positive) 4.39 4.27 3.21 3.40

 

 

 

 

 

t-values for pairwise differences between ad types are not significant at a = 0.05

For the no-delay condition, the data in Table 17 appear to indicate that Ab
is a linear function of ad type, while PI is a U-shaped function of ad type.
However, a t-test shows that the differences in Ab and PI scores (between pairs of
ad types) are not statistically significant. Thus, H2 is only partially supported by
the data.

Based upon the distinctiveness hypothesis, H3 predicted that, for the delay
condition, Ab would be a U-shaped function of ad type. After examining the
delay condition data in Table 17, no clear pattern emerges in terms of the
functional relationship between ad type and Ab. The mean Ab scores are almost
equal across all the three ad types. A t-test shows that the differences in Ab and
PI scores (between pairs of ad types) are not statistically significant. Thus, H3 is
not supported.

When the data are re-analyzed along the tercile split ad group, instead of

the predetermined ”ad type,” a slightly clearer pattern emerges. The mean

67

scores for Ab and PI grouped according to ad group and delay condition are
reported in Table 18. As can be seen from this table, Ab appears to be a linear
function of Aad (ad group) for the non-delay condition. Moreover, the

differences in Ab scores across the three Aad terciles are statistically significant.

 

 

 

 

 

Table 18.
Mean A}, and PI Scores by Ad Group and Delay Condition.
A], PI
Ad Group No Delay No Delay
(Aad terciles) delay delay
Negative (Tercile 1) 3.369C 3.3099 2.61“! 2.749.C
Neutral (Tercile 2) 4.22“ 4.34“l 3.34a 3.85a
Positive (Tercile 3) 5.1243 4.79a 4.06a 3.92a

 

 

 

 

 

 

 

a Significantly different from Ad Group 1 at a = 0.01
b Significantly different from Ad Group 2 at or = 0.01
C Significantly different from Ad Group 3 at or = 0.01

Figure 11.
Observed relationship between Aad (Ad Group) and A1,,
for no-delay and delay conditions.

 

 

 

 

 

 

 

 

 

Attitude toward the Brand (Ab)

 

 

 

 

 

 

 

l
Negative Neutral Positive
Attitude toward the Ad

It may be noted that purchase intention also appears to be a linear
function of Aad, for both the non-delay and delay conditions. However, a t-test

indicates that the difference between the PI scores for the neutral and positive

68

conditions is not statistically significant (see Table 18). The data from Table 18
are presented visually in Figure 11, to facilitate a visual comparison of the results
with the trends predicted by H2 and H3 (shown in Figure 5). It is clear that Ab
appears to be a linear function of Aad for both the no-delay and delay conditions.
Thus, H2 is supported, but Hg is not supported by the data.

To further explore if delay moderated the Aad —> Ab relationship
significantly, an analysis of variance (ANOVA) was conducted, with Ab as the
criterion variable, and Ad Type and Delay as predictor variables. The main effect
for ad type (F=1.612; P>0.20) and delay (F=0.166; P>0.68), and the two-way
interaction effects between ad type and delay (F=0.296; P>0.74) were all not
statistically significant. However, an AN OVA using Ad Group (Aad terciles)
instead of ad type, yielded slightly different results. The results of the ANOVA
are reported in Table 19. The data in the table indicate that Aad had a statistically
significant effect on Ab, but that the main effect of delay, and of the interaction of
delay and Aad were not statistically significant. Therefore, the data in Table 19
further confirm the trend seen visually in Figure 11, namely, that delay seems to

have had no moderating effect on the linear relationship between Aad and Ab.

Table 19.
Effects of Aad and Delay on A1,.

F Degrees
Freedom

 

69

The effects of Aad and delay on Ab were also examined using a multiple
regression analysis. In the regression model, delay was represented by a dummy
variable, which was assigned a value of 1 for the 7-day delay condition, and 0 for
the no—delay condition. The regression confirmed the results of the AN OVA,
namely, that only Aad had a statistically significant impact on A5. Neither delay,
nor the interaction term had a statistically significant beta values. As noted
earlier in Table 16, Aad alone accounts for 33 percent of the variation in Ab.

5.3.5. Effect of delay on ad recall

Based upon the findings of Chattopadhyay and N edungadi (1992), H4
predicted that ad recall would be lower for the 7-day delay condition, compared
to the no-delay condition. The data for unaided ad recall analyzed by no-delay

and 7-day delay conditions is presented in Table 20.

Table 20.
Unaided Recall of Ads by Delay Condition.
(Data pooled across all ad types)

 

 

 

 

Delay Condition Recalled Did Not Total
Ad Recall Ad
No-delay 57 (64.0% 32 (36.0% 89
Seven-day delay 40 (41.2%) 57 (58.8%) 97
TSEal 97 (52.2%) 89 (47.8%) 186

 

 

 

 

 

 

Chi-square (d.f. l) = 9.68 (P < 0.01)
A majority (64 percent) of the subjects in the no-delay condition were able to
recall the target ads, whereas only a minority (41.2 percent) of the subjects in the
seven-day delay condition recalled the ads. A chi-square test shows that these
differences are statistically significant at an alpha level of 0.01. Thus, H4 is
supported by the data. As might be expected, recall of all three types of target ad
was lower after a seven-day delay than immediately after exposure. This leads to
the question of whether there is a difference in recall levels across the three

different ad types, as predicted by H5.

70

The data on immediate and delayed recall levels of the three different ad
types is presented in Table 21. In the no-delay condition, all the three types of
target ad had high levels of recall, with the neutral ad (Arnold Bakery Light
bread) having substantially higher recall (72%) than either the negative ad (61%)
or the positive ad (61.4% ). A chi-square test for the no-delay condition was not
significant at an alpha level of 0.05, indicating that the observed differences in

recall across the ad types was probably due to random variances.

 

 

 

 

 

 

Table 21.
Unaided Recall of Ad Types by Delay Condition.
No-delay Seven-day delay
Ad Type Recalled Did Not Total Recalled Did Not Total
Ad Recall Ad Ad Recall Ad

Krunchers 25 (61.0%) 16 (39.0%) 41 31 (60.8%) 20 (39.2%) 51
(Negative) ‘ ‘
Arnold
Bakery 18 (72.0%) 7 (28.0%) 25 7 (29.2%) 17 (70.8%) 24
(Neutral)
Prior 16 (61.5%) 10 (38.5%) 26 2 (8.0%) 23 (92.0%) 25
(Positive)
Total 59 (64.1%) 33 (35.9%) 92 40 (40.0%) 60 (60.0%) 100

(100%) (100%)

 

 

 

 

 

 

 

 

 

For the no-delay group, Chi-square (d.f. 2) = 0.93 (n.s.)
For the delay group, Chi-square (d.f. 2) = 21. 02 (P < 0.001)

In the seven-day delay condition, there was a substantial difference in the
recall rates across the three ad types. The negative ad had the highest recall
(60.8% ), the neutral ad had a substantially lower recall (29.2% ), and the positive
ad had extremely low recall (8.0%) after a week’s delay. A chi-square test for the
seven-day delay condition shows that these differences were statistically
significant at an alpha level of 0.01. The data from Table 21 are presented
visually in Figure 13. The data appear to suggest that there is slower forgetting

of the negative ad, compared to the neutral and positive ads.

71

Figure 12.
Unaided recall of different ad types by
no-delay and seven-day delay groups.

(D
O

O)
O

 

—O— Negative Ad
—'D— Neutral Ad
—fi- Positive Ad

N
C

UnaidedAdRecaIl%
A
o

 

 

O

No delay 7-day delay
Delay

When the data in Table 21 are re—analyzed using ad group (Aad terciles),
rather than ad types, the trend is similar in direction, but less dramatic. Recall
data for the no-delay and seven-day delay group are presented in Table 22. Here
again, the recall levels for all three ad groups are very close together, with the
neutral ad being recalled by slightly more subjects (66.7%) compared with the
positive ad (65.4%) and the negative ad (61.1%). A chi-square test shows that any
observed differences are not significant at the 0.05 alpha level, indicating that any
observed variations in recall level in the no-delay group could be due to random
variations in the data. In the seven-day delay group, however, the negative ad
has substantially higher recall (60.0% ), compared with the neutral ad (29.6%) and
the positive ad (28.9% ). Thus, the analysis by ad group confirms the pattern
found in the analysis by ad type, namely, forgetting of the negative ad is slower
than for the neutral and positive ads. However, the decay in recall seems to be
approximately equal for the neutral and positive ad groups. This is illustrated

visually in Figure13.

72

 

 

 

 

 

 

Table 2.2.
Unaided Recall of Ad Groups by Delay Condition.
No-delay group Seven-day delamoup
Ad Group Recalled Did Not Total Recalled Did Not Total
(ABS terciles) Ad Recall Ad Ad Recall Ad
Negative 2 (61.1%) 14 (38.9%) 36 21 (60.0%) 14 (40.0%) 35
(Tercile 1)
Neutral 20 (66.7%) 10 (33.3%) 30 8 (29.6%) 19 (70.4%) 27
(Tercile 2)
Positive 17 (65.4%) 9 (34.6%) 33 11 (28.9%) 27 (71.1%) 38
(Tercile 3)
Total 59 (64.1%) 33 (35.9%) 92 40 (40.0%) 60 (60.0%) 100
(100%) (100%)

 

 

 

 

 

 

 

 

 

For the no-delay group, Chi-square (d.f. 2) = 0.24 (n.s.)
For the delay group, Chi-square (d.f. 2) = 8.98 (P < 0.02)

Figure 13.
Unaided recall of different ad groups
by no-delay and seven-day delay groups.

\I
O

O)
C

(II
C

 

—O- Negative Ad Group I
—0- Neutral Ad Group
—a— Positive Ad Group

Unaided Recall %
A
o

00
o

 

 

N
o

No delay 7-day delay
Delay

It had been hypothesized (in H5) that the recall of affectively extreme ads
would decay slower than the recall of neutral ads. This hypothesis is supported
only partially by the data in Tables 20 and 21, and Figures 12 and 13. The decay
in recall of the negative ad is indeed far less than the decay in recall of the neutral
ad. However, unaided recall of the positive ad has decayed at approximately the
same rate as recall of the neutral ad. This result could be the effect of the
executional characteristics of the stimulus ads. The Krunchers potato chips ad,

which was intended to induce the negative Aad condition, had a high noise level

73

and constant repetition of the brand name. While the neutral ad (Arnold Bakery
Light Bread) also involved frequent mention of the brand name, it was not done
in such a highly intrusive fashion. In the positive ad (Prior Instant Chicken), the
noise level was relatively low, and the brand name was mentioned only visually.
The constant and noisy repetition of the brand name in the negative (Krunchers)
ad may have contributed to a stronger memory trace for this ad, compared with
the other two types of ad.

The data for the effect of delay on aided recall mirror the trends observed
with the data on unaided recall. The data on immediate and delayed aided recall
levels of the three different ad types is presented in Table 23. In the no-delay
group, aided recall is highest for the neutral ad (96.4% ), followed by slightly
lower recall for the negative ad (90.5% ), and considerably lower recall for the
positive ad (73.1%). In the seven-day delay group, however, recall is highest for
the negative ad (94.3% ), followed by lower recall for the neutral ad (63.0% ), and

even lower recall for the positive ad (41.4% ). These trends are presented visually

 

 

 

 

 

 

 

 

 

 

 

 

 

in Figure 14.
Table 23.
Aided Recall of Ad Types by Delay Condition.
No-delay Seven-day delay
Ad Type Recalled Did Not Total Recalled Did Not Total
Ad Recall Ad Ad Recall Ad
Krunchers 38 (90.5%) 4 (9.5%) 42 50 (94.3%) 3 (5.7%) 53
(Negative)
Arnold 27 (96.4%) 1 (3.6%) 28 17 (63.0%) 10 (37.0%) 27
Bakery
(Neutral)
Prior 19 (73.1%) 7 (26.9%) 26 12 (41.4%) 17 (58.6%) 29
(Positive)
Total 84 (87.5%) 12 (12.5%) 96 79 (72.5%) 30 (27.5%) 109
(100%) (100%)

 

For the no-delay group, Chi-square (d.f. 2) = 7.32 (P<0.03)
For the delay group, Chi-square (d.f. 2) = 27.98 (P < 0.001)

 

74

Figure 14.
Aided recall of different ad types
by no-delay and seven-day delay groups.

 

_0— Negative Ad
""‘U"" Neutral Ad
—A— Positive Ad

Aided Recall%

 

 

 

 

No delay 7-day delay
Delay

The data for aided recall analyzed by ad group (Aad terciles) follow a
trend similar to the data analyzed by ad type (see Table 24). In the no-delay
group, the negative ad group had the highest recall (94.6% ), followed by the
neutral ad group (88.6%), followed by the positive ad group (75.0% ). In the
seven-day delay group, the three ad groups retain the same rank ordering in
terms of aided recall. However, it is evident that aided recall for the negative ad
group has decayed much less due to delay, than aided recall for the neutral or

positive ads (see Table 24 and Figure 15).

Table 24.
Aided Recall of Ad Groups by Delay Condition.

Ad Group

 

For the no-delay group, Chi-square (d.f. 2) = 5.168 (P>0.07)
For the delay group, Chi-square (d.f. 2) = 6.07 (P < 0.05)

75

Figure 15.
Aided recall of different ad groups
by no-delay and seven-day delay groups.

100

CO
C

 

—D— Mgative Ad
—'0— Neutral Ad
-'¢— Positive Ad

AidedRecall%
8

\I
O

 

 

8

No delay 7-day delay
Delay

5.3.6. Effects of repetition on Aad and Ab

Based on the mere exposure effect, and the findings of previous
researchers such as Cox and Cox (1988), H6 predicted that repetition would
result in an increase in Aad scores, regardless of ad type. Similarly, H7 predicted
that repetition would result in an increase in Ab scores, regardless of ad type. To
test these hypotheses, t-tests were conducted to examine the differences in the
mean Aad and Ab scores, for the subjects that were exposed to the target ad once,
compared with those who were exposed to the ad three times. The results are
reported in Table 25. First, the data for all three ads were pooled together and
the effects of repetition were examined. The mean Aad score for the single-
exposure group was 3.60 (on a 7-point scale), whereas for the three—exposure
group, it was 4.40. The mean Ab score for the single-exposure group was 3.89,
whereas for the three-exposure group, it was 4.34. The t-test results indicate that
these differences are statistically significant at an alpha level of 0.05. Thus, H6

and H7 are supported when the data are pooled across all ad types.

76

Table 25.
Mean Aad and A1, Scores by Repetition Condition.

Ad Type Single Single

3.60 4.402' 3.89
egative 3.83a 4.36a
en 3 8
five 4.99 5.65 4.22 4.28
a Significantly different between repetition groups at or = 0.01
b Significantly different between repetition groups at a = 0.05

 

When the data for the effect of repetition are analyzed using the mean Aad
and Ab scores for the individual ad types, a slightly different picture emerges
(see Table 25, and Figures 16 and 17). Repetition seems to have resulted in
increased Aad and Ab scores for all three ad types. However, the magnitude of
increase in Aad and Ab is the greatest for the negative ad, and the least for the
positive ad. The results of the t-test show that, while the increase in Aad and Ab
scores is statistically significant for the negative ad, only the difference in Aad
scores is statistically significant for the neutral ad at the alpha level of 0.05. The
increase in the scores for the positive ad is not statistically significant at the alpha

level of 0.05.

Figure 16.
Effect of repetition on Aad for each ad type.

 

—0— Negative Ad
—0— Neutral Ad
—a— Positive Ad

MeanAadScores

 

 

1 Exposure 3 Exposures
No. of Exposures

77

Figure 17.
Effect of repetition on A}, scores for each ad type.

 

 

 

 

 

 

 

 

 

 

5
.31
34
§ —0- NegativeAd
5 —O- NeutralAd
—O— PositiveAd
3

 

1 Exposure 3 Exposures
No. of Exposures

In sum, the data from Table 25, and Figures 16 and 17, appear to suggest
that ads that are seen as annoying during the first exposure, become more
tolerable after three exposures. However, repeated exposure does not seem to
improve people’s liking for ads that are neutrally or positively evaluated. In
evaluating these results, one must bear in mind the limitation voiced in Chapter 3
(Section 3.2, footnote 1) that the phenomenon of wearout of bad (and even good)
advertising cannot be ruled out on the basis of a repetition condition involving
just three exposures. It probably needs more than three exposures for wearout to
set in. Nevertheless, the data provide limited support for H6 and H7, even
though the increment in Aad and Ab is not substantial (or statistically significant )
for the neutral and positive ads.

To further explore if repetition moderated the Aad —> Ab relationship in a
statistically significant manner, an analysis of variance (ANOVA) was conducted,
with Ab as the criterion variable, and Ad Type and Repetition as predictor
variables. The main effect for ad type was not statistically significant (F=1.63;
P>0.19). The main effect for repetition was statistically significant (F=4.87;
P<0.03). The interaction effects between ad type and repetition did not have a

78

statistically significant effect on Ab (F=2.13; P>0.12). An ANOVA using Ad
Group (Aad terciles) instead of ad type, yielded slightly different results. The
results of the ANOVA are reported in Table 24. The data in the table indicate
that Aad had a statistically significant effect on Ab, but that the main effect of
repetition, and of the interaction of repetition and Aad were not statistically
significant. Therefore, the data in Table 26 further confirm that repetition seems

to have had no moderating effect on the linear relationship between Aad and Ab.

 

 

 

 

 

 

Table 26.
Effects of Aad and Repetition on A1,.
Source F Statistic Degrees of Significance of
Freedom F

Main Effects
Ad Group (Aad terciles) , 20.49 3 <0.01
Repetition 0.807 2 >036
2-way interactions
Ad Group x Repetition 0.151 2 >085

 

 

 

 

 

 

As in the case of delay, the effects of Aad and repetition on Ab were also
examined using a multiple regression analysis. In the regression model,
repetition was represented by a dummy variable, which was assigned a value of
1 for the three-exposure (repetition) condition, and 0 for the single-exposure
condition. The regression confirmed the results of the ANOVA reported in Table
26, namely, that only Aad had a statistically significant impact on Ab. Neither
repetition, nor the interaction term had a statistically significant beta values. It
has already been seen in Table 16 that Aad alone accounts for 33 percent of the
variation in Ab.

5.3.7. Effect of repetition on overall ad recall

Based on the expectation that increased exposure to the same ad would

create a stronger trace for the ad in the viewer’ s memory, H3 predicted that

greater exposure would result in greater recall for the ad. The data for unaided

79

recall of the target ad (pooled across all ad types) by repetition condition is
presented in Table 27. The data shows that the proportion of successful unaided
recall of the target ad was 69.0% among subjects who saw the ad three times,
compared with 38.2% among those who saw the ad once. A chi-square test
showed that these differences are statistically significant at an alpha level of 0.01.
Thus, H3 is supported by the data, if unaided recall is used as the criterion.

Table 27.
Unaided Recall of Target Ad by Repetition.

 

 

 

 

 

Recalled Did Not Total
Repetition Condition Ad Recall Ad
We Exposure 39 (38.2%) 63 (61.8%) 102
Three Exposures 58 (69.0%) 26 (31.0%) 84
Total 97 (52.2%) 89 (47.8%) 186

 

 

 

 

 

Chi-square (d.f. 1)= 17.53 (P <0.01)
N.B.: Does not include data for respondents who recalled just the brand name or
product category.

When the data are analyzed for aided recall, the improvement in the recall
levels is in the direction predicted by H3. However, the value of chi-square is not
significant at an alpha level of 0.05, indicating that the observed differences in
aided recall across exposure conditions could be attributed to random variations
(see Table 28).

The data on the effect of repetition on ad recall was also analyzed to see if
there were any observable differences across the different ad types. The results
of this analysis are presented in Table 29. The data once again provide support
for H3. Repetition has resulted in improved unaided recall for all three ad types,
although in the case of the positive ad, the difference in recall scores is
statistically significant only at an alpha level of 0.08.

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 28.
Aided Recall of Target Ad by Repetition.
Recalled Did Not Total
‘Repetition Condition Ad Recall Ad
One Exposure 41 (83.7%) 8 (66.7%) 49
Three Eflaosures 43 (91.5%) 4 (8.5%) 47
Total 84 (87.5%) 12 (12.5%) 96
Chi-square (d.f. 1) = 1.34 (P > 0.24)
N.B.: Does not include data from subjects who responded ”not sure.”
Table 29.
Unaided Recall of Ad Types by Repetition Condition.
Sin e Exposure Three exposures
Ad Type Recalled Did Not Total Recalled Did N of Total
Ad Recall Ad Ad Recall Ad
Krunchers 24 (51.1%) 23 (48.9%) 47 32 (72.7%) 12 (27.3%) 44
(Negative)
Arnold 9 (31.0%) 20 (69.0%) 29 16 (84.2%) 3 (15.8%) 19
Bakery
(Neutral)
Prior 6 (23.1%) 20 (76.9%) 26 10 (47.6%) 11 (52.4%) 21
(Positive)
Total 39 (38.2%) 63 (61.8%) 102 58 (69.0%) 26 (31.0%) 84

 

 

 

 

 

 

 

 

 

For the negative ad, Chi-square (d.f. 1) = 451 (P < 0.05)

For the neutral ad, Chi-square (d.f. 1) = 13.01 (P < 0.001 )

For the positive ad, Chi-square (d.f. 1) = 3.11 (P < 0.08)
5.3.8. Effect of repetition on immediate versus delayed recall

It was hypothesized that repetition would help increase the accessibility of

ad-related information in consumers' minds. Based on this, H9 predicted that the
decline in recall between the immediate and 7-day delay groups would be
attenuated by repetition. The data to test this hypothesis are presented in Table
30. For respondents who were exposed to the target ad(s) once, recall levels have
declined from 60.0% for the no-delay condition, to 21.1% for the seven-day delay

condition. For respondents who were exposed to the ad three times, however,

the recall level for the 7-day delay group is 70.0%, which is marginally greater

81

than the recall level for the no-delay group, which is 68.2%. A one-way
chi-square analysis shows that this difference in recall levels is not statistically
significant. Therefore, one may infer that there is no substantial difference
between the levels of immediate and delayed recall for subjects in the three-

exposure condition. Therefore, the data in Table 30 does provide at least partial

 

 

 

 

 

support for H9.
Table 30.
Unaided Recall of Ads by Delay and Repetition.
N o-delay Seven-day delay

Repetition Recalled Did N of Total Recalled Did N of Total

Condition Ad Recall Ad Ad Recall Ad
One exposure 27 (60.0%) 18 (40.0%) 45 12 (21.1%) 45 (78.9%) 57
3 exposures 30 (68.2%) 14 (31.8%) 44 28 (70.0%) 12 (30.0%) 40
Total 57 (64.0%) 32 (36.0%) 89 40 (41.2%) 57 (58.8%) 97

 

 

 

 

 

 

 

 

 

For the no-delay group, Chi-square (d.f. 1) = 0.65 (P > 0.42)
For the delay group, Chi-square (d.f. 1) = 23.24 (P < 0.001)

Figure 18.
Unaided Recall of Ads by Delay and Repetition.

 

Unaided ad recall %

—P- One—exposure group
—O— Three-exposure group

 

 

No delay 7-day delay
Delay

The data from Table 30 are presented visually in Figure 18. This figure
can be compared with Figure 7, which provided a visual presentation of the
interaction pattern predicted by H9. The pattern of interaction is in the expected
direction, although in Figure 7, both the lines had a downward slope.

82

5.3.9. Effect of repetition on recall across ad types

Using the distinctiveness hypothesis combined with H3 and H9, H10 had
predicted that the improvement in recall scores produced by repetition would be
greater for affectively extreme (positive and negative) ads, compared to neutral
ads. The assumption in this hypothesis was that repetition and affective
intensity would combine to enhance consumer memory for the ad and brand.
The expected pattern of interactions was presented visually in Figure 8. The data
pertaining to this hypothesis are presented in Table 29. The observed pattern of
interactions is presented visually in Figure 19. It is obvious that the data do not
support H10. The observed interactions are precisely the inverse of what was
predicted in Figure 8: the interaction curve was expected to be U-shaped (or

J-shaped); instead, the data follows an inverted-U shape.

Figure 19.
Observed improvement in recall across ad types.

:2: /e\
/ \
.J . ix rmm

Negative Neutral Positive
Ad type

The pattern in Figure 19 does support the alternative hypothesis speculated upon

 

 

 

 

 

 

 

Inoreaseinrecall%duebrepeti1ion

 

 

in Chapter 3 (Section 3.2). That hypothesis was based on the assumption that
there could be a ”maximum threshold" of ad recall. Recall of the affectively
extreme ads is already high due to their distinctiveness, and close to the

maximum threshold. Therefore, repetition cannot increase the recall level very

83

greatly. On the other hand, recall of the neutral ad is relatively low after one
exposure, therefore repetition might result in a substantially increased recall.
However, the notion of a ”maximum threshold” level of recall is not supported
by the absolute levels of recall for the positive ad (23.1% for single exposure;
47.6% for three exposures), which is the lowest among the three ad types (see
Table 29).

5.4 EFFECT OF INVOLVEMENT

As noted in Chapter 2, several researchers have hypothesized that product
category involvement would moderate the Aad —> Ab relationship (Muehling
and McCann 1993). Typically, Aad is expected to have a stronger influence on Ab
for low involvement products, and a relatively weaker influence for high
involvement products. Although no hypotheses were proposed in Chapter 3
regarding the effects of involvement on the relationship between Aad and A5,
product category involvement was measured during the study, to see if
variations in involvement could explain any of the observed results. The scales
for measuring involvement have already been described in Chapter 4 (see Table
6).

The effect of involvement was examined using a multiple regression
model. The criterion variable was Ab, and the predictor variables were Aad;
product category involvement, and a multiplicative term representing interaction
between Aad and involvement. The results of the regression are presented in
Table 31. It is clear from the table that the inclusion of involvement and the
interaction term did not improve the predictive power of the original model
using just Aad- Therefore, one may conclude that product category involvement

did not have a moderating effect on the relationship between Aad and Ab.

84

 

 

 

 

 

 

 

 

 

 

Table 31.
Regression results for the effect of involvement and Aad on A1,.
_Variables in Equation R2 fiF Significance ‘
of F

Aad 0.329 97.22 <0.01
Aad, Involvement 0.334 49.44 <0.01
Aad: Involvement,

Aadxlnvolvement 0.335 32.94 <0.01

5.5 SUMMARY

The results of the study lend further support to the contention that Aad
has a direct, linear influence on Ab. Of the hypotheses proposed in Chapter 3,
only H4 and H3 were fully supported by the data. Partial support was found for
H2, H5, H6 and H7. The remaining hypotheses, H1, H3, H9 and H10, were not

supported by the data. The implications of these results are discussed in
Chapter 6.

Chapter 6
DISCUSSION AND CONCLUSIONS

This chapter includes a discussion of the findings reported in the previous
chapter. The managerial implications of some of the findings are discussed in
each section. The chapter concludes with a discussion of the limitations of the
study and directions for future research.

6.1 EFFECT OF Aad ON Ab

The findings of this study lend support to the findings in the literature
that Aad has a direct effect on A1,. The correlation coefficients reported in Table
10 show that there is a strong positive correlation between Aad and Ab scores.
Further, Ab is seen to have a strong positive correlation with purchase intention
(PI). The step-wise regression results reported in Table 16 show that Aad does
not have a direct effect on P1; instead this effect is mediated by A1,.

The relationship between Aad and Ab appears to be linear, regardless of
whether Ab is measured immediately after exposure, or after a 7-day delay (see
Figure 10). Moore and Hutchinson (1985) reported that Ab and ”change in
purchase likelihood” were both linear functions of Aad in the short run, but were
U-shaped functions of Aad when measured after a 7-day delay. However, they
also reported that the expected linear-curvilinear interaction was not statistically
significant in their study (Moore and Hutchinson 1985, p. 76). They
hypothesized that the interaction they obtained was not as ”robust” as expected
because the print ads that they used as stimuli did not produce genuinely
extreme affective reactions. One of the objectives of the present study was to
extend the research in this domain by using broadcast ads instead of print ads, to
see if a stronger contrast emerged between the immediate and delayed effect of
Aad on Ab and PI. Judging from the results in Table 18 and Figure 11, the

television commercials used as stimuli in this study have not produced the

85

86
predicted interaction between Aad and delay. This leads us to the question of

whether the stimulus commercials were indeed suitable for producing the
desired effects.

As discussed in Chapter 4, the stimulus commercials were pretested to
determine their suitability. As discussed in Chapter 5, a manipulation check was
also employed in order to determine if the three commercials did indeed produce
differing levels of Aad (see Table 9). The mean scores for each ad were compared
with the tercile ranges for all Aad scores taken together. If all Aad scores are
arranged from the least to the greatest, the first tercile consists of Aad scores less
than 3.50. The mean Aad score for the Krunchers (negative) ad was 3.35. The
second tercile consists of ads with an Aad score between 3.51 and 4.82. The mean
score for the Arnold Bakery Light (neutral) ad was 3.90. The third tercile consists
of Aad scores above 4.83. The Prior (positive) commercial had a mean Aad score
of 5.31. Thus, the stimulus ads were well within their expected tercile ranges.
However, as noted in Chapter 5 (see Section 5.1), the Krunchers ad, which was
intended to produce a negative affective response, was rated only mildly
negatively, with a mean score of 3.35 on a scale ranging from 1 to 7. Moore and
Hutchinson (1985) have noted that the use of real ads in experimental situations
presents a problem, to the extent that ads generating a significant negative affect
seldom get beyond the copy testing phase of production. Thus, there appears to
be a trade-off between ecological validity and the strength of experimental
manipulation.

Another limitation of using real ads is that executional equivalence could
not be ensured across all stimulus commercials. Other researchers have
managed to overcome this limitation through the use of specially prepared
stimuli, where all executional elements save one or two remained unaltered (see,

for example, Chattopadhyay and N edungadi 1992, Gorn 1982, Haugtvedt, et al.

87

1994). In the present study, real — rather than manipulated — commercials were
used in order to ensure greater ecological validity. The three commercials used
in this study varied in terms of the use of music, the number of times the brand
name was displayed and repeated, and the ”intrusiveness” of the message. As
noted in Chapter 5, one of factors that made the Krunchers potato chips (negative
ad type) annoying was its repetitive and loud chanting of the brand name. In
contrast, the Prior instant chicken (positive) commercial had a very restrained
execution, and the brand name was not mentioned on the audio track: it was
superimposed visually in the last frame. These executional differences could
explain some of the differences found in the results when the data was analyzed
using ad type, instead of using post hoc Aad scores. In order to explore this issue
further, the stimulus commercials were analyzed according to the different
dimensions of Aad, as discussed in the next section.
6.2 DIMENSIONS OF Aad

The question of whether attitude toward the ad is a unidirnensional or
multi-dimensional construct has been a subject of some debate in the literature.
Lutz (1985) has defined Aad as "a predisposition to respond in a favorable or
unfavorable manner to a particular advertising stimulus during a particular
exposure occasion.” Burton and Lichtenstein (1988) stated that Aad should be
treated as a multi-dimensional construct. Their factor analysis indicated that a
two-factor (cognitive and affective) model was a better representation of Aad
than a unidirnensional model. On the other hand, Machleit and Wilson (1988)
have reported that they did not find any evidence to support discriminant
validity between the affective and cognitive dimensions of Aad- Miniard, Bhatla
and Rose (1990) have suggested that Aad should be decomposed into two

components: attitude toward the actual (message) claims in the ad, and attitude

88

toward the non-claim elements (i.e., execution), and the two should be measured
separately. This contention was supported by Yoon (1991).

The factor structure of Aad was also studied by Olney, Batra and Holbrook
(1990). They suggested that Aad consists of three dimensions: a ”hedonic” aspect
(e.g., pleasant-unpleasant), a ”utilitarian” aspect (e.g., useful-useless), and an
"interestingness" (sic) aspect. Olney, et al. (1990) used a 16-item semantic
differential scale to measure these three dimensions. A global measure of Aad
was also made. The three dimensions of Aad were found to account for 90
percent of the variance in Aad, leading the researchers to conclude that there was
strong support for a three-component model of Aad- The 16-item scale semantic
differential scale used by Olney, et al. (1990) was very similar to the Likert items
used in the viewer response profile (VRP) scales tested earlier by Schlinger (1979)
and Coulson (1989). Although the present study did not offer any hypotheses
regarding the dimensions of Aad, an attempt was made to further explore the
structure of Aad, using a variation of the VRP scales used previously in the
literature.

6.2.1. Factor analysis

In the present study, a global measure of Aad was taken at the conclusion
of the main experiment. Respondents were shown the target commercial once
again, and asked to respond to a battery of 6 semantic differential items about
their global Aad (see Table 6). The primary purpose of this measure was to
provide a manipulation check for the ad type variation (see Chapter 5, Section
5.1). However, in order to enable further analysis of the dimensionality issue,
subjects also responded to a battery of 20 VRP items adapted from Coulson
(1989) (see Appendix 1, Q. 37 through 56). A principal components factor
analysis was performed using the varimax procedure on SPSS-X. (An

exploratory factor analysis procedure was used, as opposed to a confirmatory

89

analysis, as this was an attempt to investigate the structure of consumers’
responses to ads). The scale items and their factor loadings are shown in Table

32.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 32.
Loadings on varimax rotated principal components.
Factort Tum Tutors Factor—4“
(Entertaln- (Negative (Cognitlve (Distinctive-
mont value! Evaluations of value! ness)
SCGIC mm Hedonism) ad) Blitz?
Utilltarlanlsm)

The commercial was lots of fun to
watch and listen to. 0.79 0.25 0.006 0.13
The commercial was entertaining. 0.77 0.35 0.03 0.16
I was involved in the commercial. 0.70 0.19 0.08 0.14
I like the mood of the commercial. 0.67 0.40 0.095 0.22
I felt the commercial was acting
out what I feel like at times. 0.67 -0.18 0.07 0.05
I would be interested in more
information about the brand. 0.63 0.15 0.23 -0.07
This was a puslg commercial. 0.11 0.75 -0.16 -0.04
The commercial insults my
intelligence. 0.28 0.69 0.13 0.06
The commercial described
characteristics undesirable to me. 0.11 0.65 0.199 0.003
The commercial was annoying. 0.55 0.62 0.004 0.17
As I watched, I thought, ”Who
cares?” 0.48 0.62 0.21 0.095
The commercial made exaggerated
claims. -0.08 0.51 0.24 -0.27
The commercial showed me the

roduct had certain advantages. 0.06 0.20 0.79 -0.02
I learned something from the
commercial that I did not know —0.02 0.05 0.72 0.11
before.
I think the advertised brand is a
good brand 0.34 0.30 0.54 0.005
The ad showed me a real
difference between the brand and 0.41 -0.26 0.53 -0.08
competition.
This commercial is different from
the commercials of its competitors. 0.07 -0.01 0.009 0.87
This commercial stands out from
other commercials. 0.22 0.00 0.07 0.85
Eigen Value 5.89 1.98 1.69 1.26

I Percent of variance accounted for I 32.8% I 11.0% I 9.4% r 7.0% I

 

 

90

The analysis yielded a five-factor model. As only one item loaded on the
fifth factor (which contributed 5.1 percent of the total variance), this factor was
eliminated, yielding a four-factor model of Aad- Taken together, the four factors
account for 60.2% of the total variance.

The first factor consists of six items, and accounts for 32.8% of total
variance. Upon examination of Table 32, it is evident that all the items loading
on Factor 1 pertain to the entertainment value or enjoyment of the commercial.
This factor can be labeled ”entertainment value.” Olney, Batra and Holbrook
(1990) labeled this factor ”hedonism.” In their study, the semantic differential
items used to measure this dimension were pleasant/ unpleasant, enjoyable/ not
enjoyable, fun to watch / not fun to watch, and entertaining/ not entertaining.
The first two scale items (”The commercial was lots of fun...,” and ”The
commercial was entertaining...”) were labeled by Coulson (1989) as measuring
”stimulation.” He labeled the next three items as measuring ”empathy/
identification” with the ad. However, in the present study, these three items
loaded very strongly on the same factor as the items measuring stimulation/
hedonism. This suggests that consumers are likely to be more involved with
commercials that they find entertaining. This is consistent with the
conceptualization of Ducoffe (1989) that the ”consumer utility” of an
advertisement is a (summative) function of its entertainment value and its
information content. Finally, the item ”I would be interested in more
information...” also loaded quite unambiguously on the first factor (factor
loading = 0.63), although Coulson (1989) and Schlinger (1979) formulated the
item as being indicative of the ”relevance of news (i.e., information contained in

the ad).”

91

The second factor consists of six items, and accounts for 11.0 percent of the
total variance. All six items in Factor 2 reflect negative evaluations of the ad (the
items were reverse coded for consistency of direction with the others). Coulson
(1989) labeled the items ”commercial was annoying,” and ”insults my
intelligence” as being measures of what he labeled ”negative commercial
evaluation.” The other four items loading on this factor are labeled differently by
him. Olney, Batra and Holbrook (1990) did not have any corresponding factor.
At first glance, negative evaluations of the ad would appear to be a dimension of
Factor 1, that is, the flip-side of entertainment value, reflecting the not-so-
enjoyable aspects of the ad. Indeed, two items with relatively weak loadings on
this factor (”The commercial was annoying,” and ”Who cares?”) also cross-load
somewhat strongly on Factor 1, indicating that the two constructs may be
correlated. (Interestingly, the ”annoying” item pertains to negative
entertainment value, while ”who cares” pertains to lack of involvement, which is
consistent with the observation in the previous paragraph that entertainment
value and ad involvement seem to be closely inter-related.) However, the items
that load more strongly on this factor seem to reflect a slightly different
dimension of negative evaluation, and relate more directly to consumers’
resistance to be persuaded by a pushy message that insults their intelligence or
describes characteristics that are undesirable to them.

The third factor consists of four items, and accounts for 9.4% of the total
variance. It is notable that all the items that load on this factor are cognitive.
Hence this factor may be labeled ”cognitive value.” Two of the items loading on
this factor (”product had certain advantages,” and ”learned something”) fall
under Coulson’s labels of ”relevant news” and ”relevance of news.” The
remaining two items fall under his labels of brand acceptance (”the advertised

brand is good”) and brand differentiation (”ad showed me a real difference

92

between brand and competition”) respectively. As noted above, the common
characteristic of all the items loading on Factor 3 is that they all pertain to
cognitive, rather than affective, responses to the ad. This ”cognitive value” factor
corresponds quite closely to the factor labeled ”utilitarianism” by Olney, Batra
and Holbrook (1990). The semantic differential items used by them to measure
utilitarianism were helpful/ not helpful, useful / not useful, informative / not
informative, and important/ not important.

The fourth factor, consists of two items and accounts for 7 percent of the
total variance. Both the items loading on this factor pertain to the dimension that
Coulson (1989) refers to as ”distinctiveness.” The Likert-scale items pertain to
the distinctiveness of the advertisement itself, and not to the novelty or news
value of the brand.

6.2.2. Regression analysis

Following the procedure used by Olney, Batra and Holbrook (1990), a
regression analysis was conducted in order to test the contribution of the
different VRP dimensions to the global measure of Aad (see Table 6 for the scale
used to measure global Aad)- The results of the regression are reported in Table
33. The adjusted value of R2 was 0.68, suggesting that the four factors together
account for 68 percent of the total variance in global Aad. While high, this figure
is substantially lower than the extremely high R2 of 0.90 reported by Olney, Batra
and Holbrook (1990, p. 277).

It is evident from Table 33 that only Factor 1 (entertainment value/
hedonism) and Factor 2 (negative evaluations) contribute in a statistically
significant manner toward predicting Aad- A stepwise regression resulted in the
elimination of Factor 3 (cognitive value / utilitarianism) and Factor 4 (distinctive-
ness) from the equation. The final model, incorporating only entertainment

value/ hedonism, and negative evaluation, had an R2 of 0.69, indicating that

93

entertainment value/ hedonism, coupled with negative evaluations, account for
69 percent of the variation in global Aad- It is notable that the cognitive value of

the ad seems to make the least contribution toward predicting overall Aad-

 

 

 

 

Table 33.
Standardized Regression Coefficients: Effect of Factors on Global Aad-
. Beta Significance
Independent Vanables of T
Entertainment value/
hedonism 0.47 <0.01
Negative evaluation 0.45 <0.01
Cognitive value/ news 0.002 >0.94
Distinctiveness 0.05 >024

 

 

 

 

 

 

Adjusted R2 = 0.68
F (4, 202) = 112.70 (P < 0.01)

While Olney, Batra and Holbrook (1990) found that all three of their
components (hedonism, utilitarianism and interestingness) had statistically
significant regression coefficients, the utilitarianism factor (cognitive value) had a
lower standardized regression coefficient (0.28), compared to hedonism (0.74)
and interestingness (0.53).

In sum, the data provide some support for a four-factor model of Aad:
consisting of entertainment value / hedonism, negative evaluations, cognitive
value/ utilitarianism, and distinctiveness. Although two of these factors parallel
the three-factor structure suggested by Olney, Batra and Holbrook (1990), the
data does not appear to offer support for ”interestingness” of the ad as a factor
distinct from entertainment value/ hedonism. Olney and his colleagues state
that the use of ”interesting” as a scale dimension suggests the existence of a third
(”collative”) aspect of Aad; which is conceptually different from simple
evaluative or affective responses. It must be borne in mind that the scales used in

this study are not identical to those used by Olney, et al.. There is need for

94

further exploration of the issue using two complete sets of alternative scales.
Likewise, the distinct dimension of ”negative evaluations,” which was reported
by Coulson (1989), but not by Olney, et al. (1990), requires further exploration. It
is surprising that ”irritation,” which is not a distinct conceptual construct, should
emerge as a factor in itself, with some correlation to the ”entertainment value /
hedonism” factor. This still raises some questions about the face validity of the
two dimensions, and needs to be explored in future research.

Recent researchers (Muehling and McCann 1993, Percy and Rossiter 1992)
have stressed the need to obtain a better understanding of the dimensions of Aad
in order to make the whole construct more meaningful for research on
advertising response. While it was not directly related to the hypotheses in this
dissertation, the exploration of the dimensions of consumer responses to
advertising serves two purposes. It provides partial support for some of the
findings in the literature, to the extent that it closely parallels the findings on
Olney, et al. (1990). It also provides new directions for further exploration, which
should help resolve some of the doubts that people have about the utility of this
stream of academic research.

6.3 RECALL OF DIFFERENT AD TYPES

Based on the distinctiveness hypothesis (Moore and Hutchinson 1983,
1985; Page, Thorson and Heide 1990), H1 had predicted that the affectively
extreme (negative and positive) ads would have a higher recall than the neutral
ad. This hypothesis was not supported by the data. The negative ad had the
highest overall recall, followed by the neutral ad, followed by the positive ad.
This could be due to the relatively greater intrusiveness of the commercial used
to induce the negative Aad response. The data leads to the rather disturbing
conclusion that disliked ads are more likely to be remembered than positive or

neutral ones. This conclusion is supported by the data on delayed recall. After a

95
seven-day delay, recall of the neutral and positive ad had declined substantially,

whereas recall of the negative ad declined at a much slower rate (see Figures 12,
13 and 14). The practical application of this finding could result in a further
proliferation of loud, intrusive ads. However, such a strategy would be
beneficial to advertisers only if there was evidence for a familiarity-based sleeper
effect, as was hypothesized in H3. This hypothesis was based on the critical
assumption articulated by Moore and Hutchinson (1983) that ”ad affect is
forgotten more readily than brand familiarity. In short, immediately following
exposure to an ad, brand attitudes may be directly linked to ad affect but, after
some delay brand attitudes may be more a function of brand familiarity than the
initial affective reaction to the ad (p. 530).” It has already been noted that the
data in the present study do not support H3. An examination of Figure 11
indicates that even after a seven-day delay, A5 is still a linear function of Aad-
Therefore, better recall of an ad does not necessarily translate to improved liking
for the advertised brand. This means that the interests of advertisers would not
be well served by creating intrusive ads merely in order to induce recall.
6.4 FAMILIARITY BASED SLEEPER EFFECT

The data in this study did not offer support for the pattern of interaction
between Aad and delay in influencing Ab, that was predicted by H3. This
hypothesis was based on what Moore and Hutchinson (1983) referred to as a
familiarity-based sleeper effect. It has been noted in Chapter 2 that the sleeper
effect has had a checkered past. Citing the findings of Gruder and his colleagues
(1978), Pratkanis and Greenwald (1985) stated that a sleeper effect would occur
only under certain restrictive conditions that are unlikely to be representative of
actual advertising viewing conditions (see Chapter 2, Section 2.4). Although
Chattopadhyay and Nedungadi (1992) and Moore and Hutchinson (1983, 1985)

found some support for a familiarity based sleeper effect, the data in the present

96
study failed to support the existence of such an effect under ad exposure

conditions that had a fairly high degree of ecological validity. In managerial
terms, this finding reinforces the need for advertisers to create ads that are at
least somewhat likable. They cannot hope to get away with creating irritating or
intrusive ads, in the hope that consumers will remember just the brand, while
forgetting their negative affect toward the ad that made them aware of the brand.
If the ad evoked a negative emotional response, some amount of residual
negative affect toward the brand is likely to remain even over time.

6.5 REPETIT ION: WEARIN AND WEAROUT

It had been predicted by H6 and H7 that repeated exposures to an ad
would result in increased liking of the ad and brand, regardless of whether the
ad was initially evaluated to be negative, neutral or positive. As seen in Figures
16 and 17, the data provide limited support for these hypotheses. Repetition
resulted in increased liking for both the ad and the brand, for all three ad types.
However, the increment in liking was not statistically significant for the neutral
and posifive ad types. This appears to indicate that consumers become more
tolerant of a negative ad due to moderate levels of repetition. Such a result is
consistent with the findings of Batra and Ray (1988), and Cox and Cox (1988). It
is also consistent with the mere exposure effect (Zajonc 1968).

It was also found that repetition has the effect of attenuating the decline in
recall caused by delay (see Figure 18). This finding indicates that repetition can
make the impact of an ad last longer, possibly by reinforcing the trace of an ad in
memory. Thus, advertisers may receive some benefit by sponsoring programs,
as sponsorship allows them to repeat commercials several times during the
program, at a price that is relatively economical compared with the up front cost

of equivalent spots.

97
The findings regarding the beneficial effects of repetition should not be

treated by advertisers as carte blanche for bombarding viewers with massive
levels of advertising. A repetition level consisting of three exposures to an ad in
a single half-hour program may seem rather high at first sight. However, this
was the first time the subjects had seen the target ads. In a widely cited study of
effective frequency, Naples (1979) concluded that the optimal level of advertising
repetition is three exposures within a purchase cycle. Even after three exposures,
advertising becomes more effective as frequency is increased, but at a decreasing
rate. Naples also stated that wearout of an advertising campaign is not caused
by too much frequency per se; it is caused by copy and content problems. On the
other hand, Blair (1987) conducted a longitudinal study and reported that highly
persuasive ads declined in their effectiveness as an exponential function of GRP
levels (low-persuasion ads did not get better or worse over time, or with
repetition). In the current study, a three-exposure level of repetition is probably
too low to permit any conclusions about the onset (or lack thereof) of wearout.

6.6 LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH
6.6.1. Stimulus materials

This study has several limitations. First, in spite of the care taken in
choosing the stimulus materials, there were noticeable executional differences in
the three ads. This limitation can be overcome in a purely experimental setting
by creating different versions of a custom-made commercial with variations in
specific, carefully chosen elements. Such a procedure has been followed by
several consumer researchers (e. g., Chattopadhyay and N edungadi 1992, Gorn
1982). The use of such a procedure might compromise ecological validity
somewhat, but it would result in greater experimental rigor.

Apart from ensuring executional equivalence, control of specific ad

elements would also permit the use of more extreme negative and positive ad

98
affect. As noted earlier in this chapter, the use of actual ads in this study places

some limits on the extremity of negative affect that can be induced.
6.6.2. Subjects

Second, despite the attempts to achieve a high degree of ecological
validity through the embedding of commercials within a half-hour program, any
advertising response experiment under forced-exposure conditions suffers from
somewhat limited external validity. The use of student subjects further limits the
external validity of the results, although this limitation is mitigated somewhat by
two considerations: (1) the advertised products were chosen such that they were
indeed likely to be purchased by the student audience (see Table 8); (2) the
students who participated in the study were enrolled in a regional university
with a fairly large commuter population, therefore, they represented a more
diverse mix of ages than might be found in a traditional university campus. (The
ages of students ranged from 17 to 53; the mean age was 24, and the median age
was 22.) Future research in this area should try to use non-student subjects — a
task rendered somewhat difficult by the fact that subjects in the delay condition
would be required to participate in two half-hour long sessions exactly one week
apart.
6.6.3. Brand familiarity effects

Third, this study did not take into account the effect of prior brand
familiarity in the Aad —> Ab relationship. Like many other studies in the field,
the present study made use of commercials for unfamiliar brands. MacKenzie
and Lutz (1989) note that this parallels c0pytesting situations, where consumers
are given little or no information about a product other than the ad, and subjects
process ads more attentively than they normally would. There are valid
experimental reasons for using fictitious or unfamiliar brands. The use of such

brands ensures tight control over the attributes that subjects can consider as

99
decision inputs. As noted by Alba, Hutchinson and Lynch (1991), "if real and

familiar brands were used, subjects might use idiosyncratic inputs that create
error variance in decision outcomes. . . However, these controls also exact a cost
(p.2)."

One of the costs that such an experimental control can exact is in terms of
overlooking the role that prior brand attitude plays in moderating the Aad —> Ab
relationship. Several researchers have attempted to overcome this shortcoming
by using familiar as well as unfamiliar brands (Edell and Burke 1986, Kent and
Allen 1994, Machleit and Wilson 1988, Phelps and Thorson 1991).

Machleit and Wilson (1988) hypothesized that Aad would not have a
significant effect on Ab when the effect of prior brand attitude was controlled.
Their model was found to fit the data. However, Edell and Burke (1986) found
that Aad had a significant effect on Ab even for familiar brands, although the
effect was greater for unfamiliar brands. Phelps and Thorson (1991) also
reported that Aad had a statistically significant effect on Ab, even for familiar
brands, although the effect of Aad is indeed attenuated by prior brand attitude.
Thus, the results obtained by Phelps and Thorson (1991) are more in line with
those of Edell and Burke (1986), and do not validate the findings of Machleit and
Wilson (1988).

In a recent study, Machleit and Sahni (1992) argued that measurement
context has an impact on the Aad —> Ab relationship for familiar brands. They
hypothesized that for familiar brands, Aad will have a significant impact on A},
only when Aad and Ab are measured contiguously. For unfamiliar brands, they
theorized that Aad would have an effect on Ab regardless of measurement
context. Their hypotheses were supported only in part by experimental data.

The relatively limited number of studies on the moderating impact of

brand familiarity, as well as the mixed findings reported in the literature, suggest

100

that there is need for more investigation in this area. Even when previous
studies on the impact of delay on the Aad —> Ab relationship have used a mix of
ads for familiar and unfamiliar brands (Moore and Hutchinson 1985), the
interaction of prior brand familiarity has not been explicitly reported. Moore and
Hutchinson (1985) exposed experimental subjects to projected print ads for 20
real brands. However, in reporting their results, they note that ”the critical
hypothesized effects were more evident when initial familiarity with the brand
was low. .. Therefore in all of the analyses that follow, data were included for a
given respondent only for [those] brands [for which the respondent was initially
not aware of its product category] (Moore and Hutchinson 1985, p.75).”
6.6.4. Variation in delay

In the present study, measures were taken from subjects in the delay
condition seven days after their exposure to the ad. This was done in order to
ensure comparability of the results with earlier studies that have used a 7-day
delay period (Chattopadhyay and Nedungadi 1992, Moore and Hutchinson
1985). It would be informative in future studies to see the longitudinal impact of
varying the delay between exposure and measurement.
6.6.5. Ad and Brand Cognitions

Several models of the Aad-Ab relationship have posited an important role
for ad cognitions and brand cognitions in the determination of ad and brand
affect (e.g., Lutz 1985, Miniard, Bhatla and Rose 1990, Yoon 1991). Ad and brand
cognitions were not measured during this study. Nevertheless, their role is
germane to the subject of the sleeper effect. Chattopadhyay and Nedungadi
(1992) found that ad-related thoughts decayed at a faster rate than brand-related
thoughts, thus providing partial support for a familiarity-based sleeper effect.
The differing rates of decay of ad cognitions and brand cognitions should be

investigated in future research.

101
6.7 CONCLUSION
This study makes several contributions to our knowledge about the
impact of consumers’ attitude toward the ad on their brand attitudes and
purchase intentions. First, it reaffirms the important role that Aad plays in
determining Ab and PI. Second, it explores the impact of delay and repetition on
ad recall, and on the relationship between Aad and Ab. This interaction is a
relatively less-studied area in the academic literature. The results appear to
reaffirm that a familiarity-based sleeper effect probably does not occur under
realistic advertising exposure conditions. Third, the study confirms that
increased recall of an ad does not necessarily result in increased liking for the ad
or brand. This result casts further doubts on the efficacy of recall based copy-
testing measures, and suggests that alternative c0pytesting measures, based on
affective responses, might be more appropriate. Fourth, the study expands the
literature on wearin and wearout of advertising. The results offer some support
for the contention that repetition enhances liking for an ad. Finally, the study
enhances our understanding of the different dimensions of Aad, an area that
recent researchers have considered worthy of further attention. While no single
study can resolve the issue of the dimensionality of Aad, the finding of some
commonalty with that of previous researchers (Olney, Batra and Holbrook 1990)
is encouraging. It indicates that we are in a better position to define the
components in an ad that will result in favorable responses from consumers. It is

this last aspect that is probably of greatest managerial interest.

APPENDICES

APPENDIX 1

102
APPENDIX 1

TELEVISION PROGRAM STUDY
QUESTIONNAIRE —v. 1.N.S

 

_To allow us to match the two parts of your questionnaire, please enter the last
four digits of your social security number below:

 

 

 

 

The purpose of this study is to learn more about your television viewing habits,
and to get your reactions to some of the television programming being produced
at the University of West Florida. You will also be asked a few questions that
pertain to your usage of products that are commonly advertised on television.

Your participation in this study is purely voluntary. If for any reason, you are unwilling
or unable to complete the questionnaire, there will be no penalty. The information in
this questionnaire will be kept completely confidential.

Please answer all questions carefully and COMPLETELY, and do not leave any
line blank. Please answer honestly. There are no ”right” or ”wrong” answers.

For example, if there is a question containing three items, then please mark each

of the three items (as shown in the sample below), even if several items appear to
be repetitive.

Ex. 1. Please indicate below your feelings about the TV program, Sesame Street:

1 2 3 4 5 6 7
Good _: X _: _: _: _: _: Bad
Unpleasant _. _ _. _. _. _ X Pleasant
Poor quality _: _: _: _: X _: _: High quality

103

In an average week, how many hours do you spend watching television?
hours

Please indicate below the names of the three channels (or networks) that
you view most often:

a. View most often
b. View second most often
c. View third most often

Do you ever watch the following channels/ networks (check all that apply)?

PBS

Arts & Entertainment (A&E)
The Discovery Channel

The Learning Channel

UWF Closed Circuit Channel

Have you ever heard of author Terry Kay?

Yes No

Have you read any of the works of author Terry Kay?

Yes No

Do you believe that Public Broadcasting System (PBS) stations should start
accepting advertising to supplement their income?

Yes No _ No Opinion _

What do you think of television advertising in general? Please respond
using the following scales.

1 2 3 4 5 6 7
Good —: —: —: —: —: —: —: Bad
Pleasant —: —: —: —: —: —: —: Unpleasant
Truthful —: —: —: —: —: —: —: Deceptive

Informative —: —: —: —: —: —: —: Not informative

104

8. Given below is a list of some brands that are advertised on television.
(Some of them may not be advertised in your area.) Please indicate if you
are familiar with the following brands by checking only ONE of the boxes
for EACH brand. '

 

Brand

1. l have never
heard of the
brand.

2. l have heard
the name but
don’t know
anything about
the predict

category.

3. l have heard of
the brand name
and know what
prodrcts it relates
to.

4. I know a Ira;
bit about the
brand and the
predict category.

 

{I am extremely
familiar with the
brand and its
predict category.

 

figer’s

 

Krunchers

 

Listerine

 

Prior

 

Claritin

 

Wrigley’s

 

Gulden's

 

Arnold
Bakery

 

Tone

 

r
Sasson

 

T’ayday

 

LMentos

 

 

Darigold

 

 

 

 

 

 

 

105

9. How likely is it that you will purchase (for yourself, or someone you know)
the following types of products within the next six months?

Product 1: Extremely 7: Extrennly
likely unlikely
7

_a
N
(a)

- ‘5
0|
0

Beer (domestic)
Allergy medicine
Jeans

Potato chips
Ice cream
Instant coffee

Breakfast cereal
Mustard (packaged)
Dishwashing liquid

Bread
Candy bars
Heat-and-eat meals

Bath soap
AppareV clothing

10. Please indicate your feelings about potato chips along the following

11.

dimensions:

Important

tome

Ofnoconcem
tome

Irrelevant

Very meaningful

to me
Matters to me
lnterefiing
Significant
Boring

1

2

3

4

5

6

7

Unimportant

tome

Of concern
to me

Relevant

Means nothing
to me

Doesn't matter
Not intereaing
Insignificant
Exciting

Please indicate your feelings about ice cream along the following

dimensions:

Interesting

Ofnoconcem
tome

Irrelevant
Matters to me
Significant
Boring
Important

IOITIO

Very meaningful

tome

Not interesting

Ofconcem
to me

Relevant
Doesn't matter
Insignificant
Exciting

Unimportant
to me

Means nothing
to me

12. Please indicate your feelings about bread along the following dimensions:

13.

lrnportant

IONIC

Ofnoconcem
to me

Irrelevant

Very meaningful

to me
Matters to me
Interesting
Significant
Boring

1

2

3

4

5

6

7

Unimportant
to me

Of concern
to me

Relevant

Means nothing
to me

Doesn't matter
Not interesting
Insignificant

Exciting

Please indicate your feelings about heat-and—eat meals along the following

dimensions:

Interesting

Ofnoconcem
tome

Irrelevant
Matters to me
Significant
Boring

Important
to me

Very meaningful

IONIC

1

2

3

4

5

6

7

Not interesting

Of concern
to me

Relevant
Doesn't matter
Insignificant
Exciting
Unimportant

to me

Means nothing

IO me

108

Please indicate how strongly you agree or disagree with the following
statements, by circling the number that best represents your feelings.

14. There is no substantial difference between different brands of potato chips.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

15. I pay little or no attention to the brand of potato chips that I buy.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

16. Some brands of potato chips are definitely superior to other brands.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

17. There is no substantial difference between different brands of bread.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

18. I pay little or no attention to the brand of bread that I buy.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

19. Some brands of bread are definitely superior to other brands.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

20. There is no substantial difference between different brands of ready-to-eat/

microwave meals.

Strongly SII'OI'Ieg
agree 1 2 3 4 5 6 7 disagree

21.

22.

109

I pay little or no attention to the brand of ready-to-eat/ microwave meals
that I buy.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

Some brands of ready-to—eat/ microwave meals are definitely superior to

other brands.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

This concludes the first part of this study.

You will now view a 30-minute television program, titled Southern Voices,
Southern Words.

This documentary was produced by faculty and students at The University of
West Florida for airing on PBS Stations. It is now being adapted for airing on
commercial cable stations, therefore the version you will see may have some
commercial breaks.

PLEASE DO NOT LEAVE THE ROOM AFTER THE PROGRAM.
YOUR FEEDBACK ABOUT THE PROGRAM IS IMPORTANT.

 

Please do not turn the page until the program has ended.

110
We would like to ask you about your opinions regarding the program you just

saw. Please answer honestly. There are no ”right” or “wrong" answers.
Please answer carefully and COMPLETELY; do not leave any line blank.

23. The program, Southern Voices, Southern Words, is:

1 2 3 4 5 6 7

Interesting —: —: —: —: —: —: —: Boring
Informative —: —: —: —: —: —: —: Uninformative
\Neflnwdb —a ——: —< -—< ——: —a -—< Poomynwde
Unappealing —: —: —: —: —: —: —: Appealing
Fascinating —: —: —: —: —: —: —: Mundane
Likable —: —: —: —: —: —: —: Not likable
Emotional —: —: —: —: —: —: —: Neutral
Involving —: —: —: —: —: —: —: Not involving
Held my Did not hold

attention —: —: —: —: —: —: —' my attention

24. After viewing the program, are you more motivated than before, to read the
works of the featured author, Terry Kay?

Yes _ No _
25. Did the interruptions in the program due to the commercial breaks bother
you?
1 2 3 4 5 6 7
Bothered me —: —: —: —: —: — — Did not bother me

26. Before viewing this program, were you aware that the University of West
Florida makes programs (other than Nautilus News ) for airing on public
television stations?

Yes No

27. Do you think that the University of West Florida should make more such
documentaries?

Yes No

28. Would you be interested in viewing some of the other documentaries
produced at UWF?

Yes No

 

Now, please open the envelope that you have been given, and
answer the questionnaire in the envelope. Thank you.

 

111

TELEVISION PROGRAM STUDY PART 2
QUESTIONNAIRE — v. 1.N.S

 

For identification purposes only, please enter the last four digits of your social
security number below:

 

 

We would now like to ask you some questions about the commercials that you
saw just now along with the program, Southern Voices, Southern Words. Please
answer honestly. There are no ”right” or ”wrong" answers. Please answer
carefully and COMPLETELY, and do not leave any line blank.

29. Please list all the brands and products for which you remember seeing
commercials along with the program, Southern Voices, Southern Words.
Please list only the brands and products; do not describe the commercials.

Do not turn to the next question until you have completed your response
to this question.

 

112

30. Did you see commercials for the following brands along with the program
Southern Voices, Southern Words?
(Please respond without referring to the previous page!)

Brand Yes No
Panasonic Camcorders
Arrid Deodorant

Archer Daniels Midland (ADM)

Avis Car Rentals
Robitussin Cold Medicine

Alamo Car Rentals

Sensodyne Toothpaste _
Hoover Vacuum Cleaners

Zenith Color TV

Sharp Camcorders
Krunchers Potato Chips
Arnold Bakery Light Bread

Prior Chicken American

113

31. Can you remember seeing a commercial for Krunchers potato chips?

Yes No _ Not sure _

(If you answered NO, please go to Question 33.)

32. Please describe, in as much detail as you can, the commercial that you have
seen for Krunchers potato chips:

114

33. Please indicate your impressions about Krunchers brand potato chips by
marking the most appropriate spot on each of the following scales:

1 2 3 4 5 6 7
Good —: —: —: —: —: —: —: Bad
Unpleasant —: —: —: —: —: —: —: Pleasant
Favorable —: —: —: —: —: —: —: Unfavorable
Dislike —: -—: —: —: —: —: —: Like
Poor quality —: —: —: —: —: —: —: High quality
Well known —: —: —: —: —: —: —: Unknown

34. If you were in the market for potato chips, how likely is it that you would
consider buying Krunchers? Please mark the appropriate spot in each of
the following scales.

1 2 3 4 5 6 7
Likely —: —: —: —: —: —: —: Unlikely
Probable —: —: —: —: —: —: —: Improbable
Possible —: —: —: —: —: —: —: Impossible

This concludes the second part of this study.
PLEASE REMAIN SEATED, AND DO NOT LEAVE THE ROOM.

In a few minutes, we will once again show you one of the commercials that
you saw with the program, and request your reactions. Please wait for the
commercial to be screened again before turning the page.

The study is almost over. Thank you for your patience!

We will now show you once again one of the commercials that you saw with

115

the program. Please give us your responses below with respect to the
commercial that you will be shown.

35.

36.

What is the brand that is being advertised?

 

Please let us know your feelings about the commercial that you just saw,

along the following scales:
1 2 3 4 5 6 7

Good —: —: —: —: —: —: —: Bad
Unpleasant —: —: —: —: —: —: —: Pleasant
Favorable —: —: —: —: —: —: —: Unfavorable
Enjoyable —: —: —: —: —: —: —: Not enjoyable
Disliked it —: —: —: —: —: --: —: Liked it
Irritating —: —: —: —: —: —: —: Likable
Informative —: —: —: —: —: —: —: Uninformative

Please respond to Questions 37-56, using the following scale.

For each statement, please circle the number that most closely represents your

I strongly disagree with the statement

I disagree with the statement

I somewhat disagree with the statement

I neither agree nor disagree with the statement
I somewhat agree with the statement

I agree with the statement

I strongly agree with the statement

NGUbQN-‘r

feelings.
37. I learned something from the commercial for Krunchers that I did not know
before.
Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

38.

The commercial showed me the product had certain advantages.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

39.

41.

43.

45.

46.

47.

48.

116
As I watched, I thought, ”Who cares?”

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

I would be interested in more information about the advertised brand.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

I think the advertised brand is a good brand.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial described characteristics undesirable to me.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The ad showed me a real difference between the brand and competition.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial made exaggerated claims.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial insults my intelligence.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial was annoying.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

It was difficult to understand the commercial.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial was lots of fun to watch and listen to.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

49.

50.

51.

52.

53.

55.

56.

117

The commercial was entertaining.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

This was a pushy commercial.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

This commercial is different from the commercials of its competitors.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

This commercial stands out from other commercials.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

I felt the commercial was acting out what I feel like at times.

Strongly Strongly
agree . 1 2 3 4 5 6 7 disagree

I was involved in the commercial.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

I like the mood of the commercial.

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

The commercial made me ”feel” rather than ”think.”

Strongly Strongly
agree 1 2 3 4 5 6 7 disagree

 

YOU’RE ALMOST DONE! We just need a little more information about

yourself, so please turn the page and finish the questionnaire.

118

57. Your gender: Male _ Female _
58. What was your age last birthday? years
59. On average, how many times a month do you shop for groceries?

times.

 

 

THANK YOU VERY MUCH FOR YOUR PARTICIPATION!

As a reward for your efforts, you have been entered in a sweepstakes, with a
chance of winning a free color TV.
Results will be declared in six to eight weeks.

 

 

APPENDIX 2

119

APPENDIX 2
SCRIPTS OF THE STIMULUS COMMERCIALS

KRUNCHERS POTATO CHIPS
30-second TV Spot

VIDEO

AUDIO

1. Open on close up of man beating a drum that MAN (chants): No more wimpy chips! No more

has the printed slogan, “No more wimpy
chips.” Camera zooms out slowly.

2. A few people gather around drum-beating
man.

3. Cut to long shot of the mall around the
Washington memorial, crowded with
thousands of demonstrators.

4. Cut to CU of demonstrators wearing T-shirts
and holding placards with the slogan, “No
more wimpy chips.”

5. Cut to CU of a child on demonstrator’s
shoulder looking up at the sky and pointing.

6. Cut to long shot of airplane dropping packets
from the sky. As one packet rus es toward
camera, it is shown to be a bag of Krunchers.

7. Cut to CU of young woman biting into a
potato chip

8. Dissolve to CU of Kruncher’s Potato Chips
bag.

9. Cut to long shot of large crowd of
demonstrators

10. Cut to MS of man biting into a potato chip.

11. Quick cut to medium shot of demonstrators
exultantly enjoying chips.

12. Dissolve to pack shot of three bags of
Krunchers.

wimpy chipsl...

ANNOUNCER (V 0): It started as a one-man
crusade...

and became a national obsession.

CROWD (chants): No more wimpy chips! No
more wimpy chipsI...

CHILD: Look!

ANNOUNCER (V0): And the nation’s cries
were answered...

SF X: Crunch (crunching of chip)

ANNOUNCER (V 0): with a resounding
crrrunch!

ANNOUNCER (V 0): Krunchers Potato Chips.
Cooked in peanut oil for a taste that’s
outrageously bold...

SFX: Crunch.
CROWD (exultantly): Yayyy!

ANNOUNCER (V0): A chip that’s incredibly
crunchy.

SFX: Crunch.
CROWD (exultantly): Yayyy!

ANNOUNCER (V0): Krunchers mesquite
barbecue, alfredo and jalapeno flavors.

120

13. Dissolve to CU of several demonstrators
joyously beating drum with the printed
slogan, “N o more wimpy chips.”

14. Super on video of demonstrators: ‘fjoin the
Krunchers Crusade.”

ANNOUNCER (V0): So say good bye to
wimpy chips.

DEMONSTRATORS (chanting): Kmnchers,
Krunchers, Krunchers...

AN NOUNCER (V0):]oin the Krunchers
crusade.

121

APPENDIX 2 (contd.)

SCRIPTS OF THE STIMULUS COMMERCIALS

ARNOLD BAKERY LIGHT BREAD
30-second TV Spot

VIDEO

1. Open with close-up of Woman—l with a
sandwich in her hand. One mouthful has
been bitten off.

2. Text (in strong, heavy, reverse typeface):
You’re wrong

. Text (in light, reverse typeface): We’re light
. Cut to close-up of loaf of bread.

. Cut to medium-shot of Man-l with sandwich
in hand. One mouthful has been bitten off.

. Cut to medium shot of brick oven in
background, with silhouette of loaves of
bread in foreground.

aim-BOD

7. Dissolve to product shots of bread, with and
without wrapper.

8. Cut to close-up of Woman-l

9. Cut to CU of slices of bread.

10. Cut to CU of Woman-2 shaking her head
and smiling as she chews on bread.

11. Cut to CU of Woman-3 shaking her head
and smiling.

12. Cut to CU of Man-2 with bread in hand.

13. Text (in strong, heavy, reverse typeface):
You’re wrong

14. Text (in light, reverse typeface): We’re light
15. Pack shot of five loaves of bread
16. Super: Arnold Bakery Light

AUDIO

WOMAN -l (high-pitched, arrogant voice): This
is light bread?! No way! Unh, unh!

JINGLE: You’re wrong!

JINGLE: We’re light!

JINGLE: Arnold Bakery Light!

MAN-l: I don’t believe ya. This is... that’s
ridiculous!

V0: At Arnold’s Bakery, we don’t bake light
bread like everybody else.

Our five bakery flavors are so delicious, you
won’t behave they’re light.

WOMAN -l: I know it doesn’t taste like light
bread, so forget it!

V0: Packed with fresh bakery taste, each full
slice is just forty calories light.

WOMAN -2 (with mouth full): No!

WOMAN -3: You’re kidding me! Get outta here!
MAN -2: It’s not light bread.
JINGLE: You’re wrong!

.IINGLE: We’re light!
jINGLE: Arnold Bakery Light!
V0: From Amold’s Bakery.

122
APPENDIX 2 (contd.)

SCRIPTS OF THE STIMULUS COMMERCIALS

PRIOR INSTANT CHICKEN DINNERS
30-second TV Spot

VIDEO AUDIO

1. Open with medium shot of balding, middle- SFX: Striking of clock.
aged man sitting at kitchen table with a bored
expression.

2. Man removes glasses as his middle-aged, SFX: Ticking of clock.
grey-haired wife walks in with two plates
loaded with food.
3. Wife lays plates on table. SF X: Ticking of clock.
4. Cut to CU of plate with chicken dinner on it. SFX: Ticking of clock.

5. Cut back to CU of man. Man looks sideways SFX: Ticking of clock.
at wife, with bored expression.

6. Cut to CU of wife. She looks anxiously at SFX: Ticking of clock.
man, as if awaiting approval.

7. CU of man as he cuts into chicken, takes a SF X: Knife clattering against plate.
mouthful and chews.

8. Cut to CU of wife looking anxiously at man.
9. Cut back to MS of man eating. As he chews MUSIC (comes up slowly): “Night Fever” disco

food, a smile slowly spreads across his face. track from the movie Saturday Night Farm.
10. Man rises from chair, rolling arms in MUSIC: “Night Fever” track.
rhythm with the music.

11. Man starts dancing. Woman also rises from Music continues.
chair, casts her apron aside and joins in
dance.

12. Couple dance together (comical but cute). Music continues.

13. Cut to ack shot of instant dinner with Music continues.
super: “ aturday Night Fever in less than an
hour.”

The descriptive name, “American Weekend
Chicken” is visible on the pack, but the place
for the brand name is blank.

14. The brand name, “Prior” dissolves onto the Music continues.
pack shot.

LIST OF REFERENCES

UST OF REFERENCES

Aaker, David A., and John G. Myers (1987). Advertising Management, Third
Edition. Englewood Cliffs, NJ: Prentice-Hall.

Aaker, David A. and Alexander L. Biel (eds.) (1993). Brand Equity and Advertising.
Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

Alba, Joseph W., J. Wesley Hutchinson, and John G. Lynch (1991). "Memory and
Decision Making," in Handbook of Consumer Behavior, Thomas S. Robertson
and Harold H. Kassarjian, eds. Englewood Cliffs, NJ: Prentice-Hall Inc., 1-
49.

Alwitt, Linda F ., and Andrew A. Mitchell (1985a). "Introduction," in Psychological
Processes and Advertising Efl'ects, Linda Alwitt and Andrew A. Mitchell,
editors. Hillsdale, NJ: Lawrence Erlbaum Associates Inc., 1-10.

Alwitt, Linda R, and Andrew A. Mitchell (1985b). "Concluding Remarks," in
Psychological Processes and Advertising Efi‘ects, Linda Alwitt and Andrew A.
Mitchell, editors. Hillsdale, NJ: Lawrence Erlbaum Associates Inc., 272-293.

Batra, Rajeev and Michael L. Ray (1986a). "Situational Effects of Advertising
Repetition: The Moderating Influence of Motivation, Ability and
Opportunity to Respond," Journal of Consumer Research, 12 (March), 432-445.

Batra, Rajeev and Michael L. Ray (1986b). ”Affective Responses Mediating
Consumer Acceptance of Advertising," Journal of Consumer Research, 13
(September), 234-249.

Blair, Margaret Henderson (1987). ”An Empirical Investigation of Wearin and
Wearout," Journal of Advertising Research, (December 1987/ January 1988),
45-50.

Brown, Steven P. and Douglas M. Stayman (1992). "Antecedents and
Consequences of Attitude toward the Ad: A Meta-Analysis," Journal of
Consumer Research, 19 (June), 34-51.

Burke, Marian C. and Julie A. Edell (1986). "Ad Reactions Over Time: Capturing
Changes in the Real World," Journal of Consumer Research, 13 (June), 114-118.

Burton, Scott and Donald R. Lichtenstein (1988). ”The Effect of Ad Claims and
Ad Context on Attitude toward the Advertisement," Journal of Advertising,
17, 1, 3-11.

124

Cacioppo, John T. and Richard E. Petty (1985). "Central and Peripheral Routes to
Persuasion: The Role of Repetition," in Psychological Processes and Advertising
Efiects, Linda Alwitt and Andrew A. Mitchell, editors. Hillsdale, NJ:
Lawrence Erlbaum Associates Inc., 91-112.

Calder, Bobby J. and Brian Stemthal (1980). "Television Commercial Wearout: An
Information Processing View," Journal of Marketing Research, 17 (May), 173-
186.

Chattopadhyay, Amitava and Prakash N edungadi (1992). "Does Attitude
Toward the Ad Endure? The Moderating Effects of Attention and Delay,"
Journal of Consumer Research, 19, 1 (June), 26-33.

Colley, Russell H. (1961). Defining Advertising Goals for Measured Advertising
Results. New York, NY: Association of National Advertisers.

Coulson, John S. (1989). "An Investigation of Mood Commercials,” in Cognitive
and Aflective Responses to Advertising, editors, Patricia Cafferata and Alice M.
Tybout. Lexington MA: D.C. Heath and Company, 21-30.

Cox, Dena Saliagas and Anthony D. Cox (1988). "What Does Familiarity Breed?
Complexity as a Moderator of Repetition Effects in Advertisement
Evaluation," Journal of Consumer Research, 15 (June), 111-116.

Cox, Dena Saliagas and William B. Locander (1987). ”Product Novelty: Does it
Moderate Relationship between Ad Attitudes and Brand Attitudes?”
Journal of Advertising, 16 (3), 39-44.

Droge, Cornelia and Rene Y. Darmon (1987). ”Associative Positioning Strategies
Through Comparative Advertising: Attribute versus Overall Similarity
Approaches," Journal of Marketing Research, 24 (November), 377-388.

D’Souza, Giles and Ram C. Rao (1995). ”Can Repeating an Advertisement More
Frequently Than the Competition Affect Brand Preference in a Mature
Market?” Journal of Marketing, 59, 2 (April), 32-42.

Ducoffe, Robert Hal (1989). ”Evaluating the Utility of Advertising: Theory and
Measurement," Unpublished doctoral dissertation. East Lansing, MI:
Michigan State University.

Edell, Julie A. and Marian C. Burke (1986). "The Relative Impact of Prior Brand
Attitude and Attitude Toward the Ad on Brand Attitude After Ad
Exposure," in Advertising and Consumer Psychology, 3, Jerry Olson and Keith
Sentis, editors. New York, NY: Praeger Publishers, 93-107.

125

Edell, Julie A. and Marian C. Burke (1987). ”The Power of Feelings in
Understanding Advertising Effects," Journal of Consumer Research, 14
(December), 421-33.

Edell, Julie A. and Marian C. Moore (1993). "The Impact and Memorability of Ad-
Induced Feelings: Implications for Brand Equity," in Brand Equity and
Advertising, David A. Aaker and Alexander L. Biel, editors. Hillsdale, NJ:
Lawrence Erlbaum Associates, Publishers.

Fishbein, Martin and Icek Ajzen (1974). "Attitudes toward objects as predictors of
single and multiple behavioral criteria," Psychological Review, 81, 59-74.

Freiden, Jon B. (1984). ”Advertising Spokesperson Effects: An Examination of
Endorser Type and Gender on Two Audiences," Journal of Advertising
Research, 24 (5), 33-41.

Gardner, Meryl Paula (1985). "Does Attitude toward the Ad Affect Brand
Attitude Under a Brand Evaluation Set?" Journal of Marketing Research, 22
(May), 192-198.

Gillig, P. M., and Anthony G. Greenwald (1974). "IS it time to lay the sleeper
effect to rest?" Journal of Personality and Social Psychology, 29, 132-139.

Corn, Gerald J. (1982). "The Effects of Music in Advertising on Choice Behavior:
A Classical Conditioning Approach," Journal of Marketing, 46 (Winter 1982),
94-101.

Greenwald, Anthony G. (1968). "Cognitive learning, cognitive response in
persuasion, and attitude change," in Psychological Foundations of Attitudes,
A.G. Greenwald, T.C. Brock and TM. Ostrom, editors. New York, NY:

Academic Press.

Gresham, Larry C. and Terence A. Shimp (1985). "Attitude toward the
Advertisement and Brand Attitudes: A Classical Conditioning Perspective,"
Journal of Advertising, 14, 1, 10-17, 49.

Gruder, C.L., T.D. Cook, K.M Hennigan, B.R. Flay, C. Alessis and J. Halamaj
(1978). "Empirical tests of the absolute Sleeper effect predicted from the
discounting cue hypothesis," Journal of Personality and Social Psychology, 36,
1061-1074.

Hastak, Manoj and Jong-Won Park (1990). ”Mediators of Message Sidedness
Effects on Cognitive Structure for Involved and Uninvolved Audiences,"
Advances in Consumer Research, 17, Marvin E. Goldberg, Gerald Corn and
Richard W. Pollay, eds. Provo, UT: Association of Consumer Research, 329-
336.

126

Haugtvedt, Curtis P., Clark Leavitt and Wendy L. Scheiner (1993). "Cognitive
Strength and Established Brands: Memory, Attitudinal and Structural
Approaches," in Brand Equity and Advertising, David A. Aaker and
Alexander L. Biel, editors. Hillsdale, NJ: Lawrence Erlbaum Associates,
Publishers.

Haugtvedt, Curtis P., David W. Schumann, Wendy L. Scheiner and Wendy L.
Warren (1994). "Advertising Repetition and Variation Strategies:
Implications for Understanding Attitude Strength," Journal of Consumer
Research, 21, 1, 176-189.

Haugtvedt, Curtis P. and Duane T. Wegener (1994). "Message Order Effects in
Persuasion: An Attitude Strength Perspective," Journal of Consumer Research,
21, 1, 205-218.

Homer, Pamela M. (1990). "The Mediating Role of Attitude toward the Ad: Some
Additional Evidence," Journal of Marketing Research, 27 (February), 78-86.

Hovland, C.I., Janis, LL. and Kelley, J.J. (1953). Communications and Persuasion.
New Haven, CT: Yale University Press.

Kamins, Michael A. (1989). ”Celebrity and Non-celebrity Advertising in a Two-
Sided Context," Journal of Advertising Research, 15 (3), 34-42.

Kamins, Michael A., Lawrence J. Marks and Deborah Skinner (1991). ”Television
Commercial Evaluation in the Context of Program Induced Mood:
Congruency versus Consistency Effects," Journal of Advertising, 20 (2), 1-14.

Keller, Kevin Lane (1991). ”Cue Compatibility and Framing in Advertising,"
Journal of Marketing Research, 28 (February), 42-57.

Kent, Robert J. and Chris T. Allen (1994). "Competitive Interference Effects in
Consumer Memory for Advertising: The Role of Brand Familiarity," Journal
of Marketing, 58 (July 1994), 97-105.

Krugman, Herbert E. (1965). "The Impact of Television Advertising: Learning
Without Involvement," Public Opinion Quarterly, 29, 349-356.

Lavidge, Robert J. and Gary A. Steiner (1961). "A Model for Predictive
Measurements of Advertising Effectiveness," Journal of Marketing, 25
(October 1961), 59-62.

Leigh, Thomas W., Arno J. Rethans and Tamatha R. Whitney (1987). ”Role
Portrayals of Women in Advertising: Cognitive Responses and Advertising
Effectiveness," Journal of Advertising Research, 27(5), 54-63.

127

Lutz, Richard J. (1985). "Affective and Cognitive Antecedents of Attitude Toward
the Ad: A Conceptual Framework," in Psychological Processes and Advertising
Eflects, Linda Alwitt and Andrew A. Mitchell, editors. Hillsdale, NJ:
Lawrence Erlbaum Associates Inc., 45-63.

Machleit, Karen A. and R. Dale Wilson (1988). "Emotional Feelings and Attitude
Toward the Advertisement: The Roles of Brand Familiarity and Repetition,"
Journal of Advertising, 17, 3, 27—35.

Machleit, Karen A. and Arti Sahni (1992). "The Impact of Measurement Context
on the Relationship Between Attitude Toward the Ad and Brand Attitude
for Familiar Brands," Advances in Consumer Research, 19, John F. Sherry and
Brian Stemthal, eds. Provo, UT: Association for Consumer Research, 279-
283.

MacKenzie, Scott B. and Richard J. Lutz (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.

MacKenzie, Scott B., Richard J. Lutz and George E. Belch (1986). "The Role of
Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A Test
of Competing Explanations," Journal of Marketing Research, 23 (May), 130-
143.

Macklin, M. Carole, Norman T. Bruvold and Carol Lynn Shea (1985). ”IS it
always as simple as , ’Keep It Simple!’?” Journal of Advertising, 14 (4), 28-35.

Madden, Thomas J., Kathleen Debevec and Jacquelyn L. Twible (1988).
”Assessing the Effects of Attitude toward the Ad on Brand Attitudes: A
Multitrait-Multimethod Design,” in Marketing Communications: Theory and
Research, Michael J. Houston and Richard J. Lutz, editors. Chicago: AMA,
109-13.

McGuire, William J. (1976). ”Some Internal Psychological Factors Influencing
Consumer Choice," Journal of Consumer Research, 2 (March), 302-319.

Miniard, Paul W., Sunil Bhatla and Randall L. Rose (1990). ”On the formation
and relationship of ad and brand attitudes: an experimental and causal
analysis," Journal of Marketing Research, 27, 290-303.

Mitchell, Andrew A. and Jerry C. Olson (1981). ”Are Product Attribute Beliefs the
Only Mediators of Advertising Effects on Brand Attitude?” Journal of
Marketing Research, 18 (August), 318-32.

128

Mittal, Banwari (1990). ”The Relative Roles of Brand Beliefs and Attitude
Toward th eAd as a Mediator of Brand Attitudes," Journal of Marketing
Research, 27 (May), 209-219.

Moore, Danny L. and J. Wesley Hutchinson (1983). "The Effects of Ad Affect on
Advertising Effectiveness," Advances in Consumer Research, 10, RP. Bagozzi
and Alice Tybout, edS.. Ann Arbor, MI: Association for Consumer Research.

Moore, Danny L. and J. Wesley Hutchinson (1985). "The Influence of Affective
Reactions to Advertising: Direct and Indirect Mechanisms of Attitude
Change," in Psychological Processes and Advertising Effects, Linda Alwitt and
Andrew A. Mitchell, editors. Hillsdale, NJ: Lawrence Erlbaum Associates
Inc., 65-87.

Muehling, Darrel D. (1987). ”Comparative Advertising: The Influence of
Attitude toward the Ad on Brand Evaluation," Journal of Advertising, 16 (4),
43-49.

Muehling, Darrel D. (1987) and Russell N. Laczniak (1992). ”An Examination of
Factors Mediating and Moderating Advertising’ S Effect on Brand Attitude
Formation," Journal of Current Issues and Research in Advertising, 14 (1), 23-
34.

Muehling, Darrel D., Russell N. Laczniak and Jeffrey J. Stoltrnan (1991). ”The
Moderating Effects of Ad Message Involvement: A Reassessment," Journal
of Advertising, 20 (2), 29-38.

Muehling, Darrel D. and Michelle McCann (1993). "Attitude toward the Ad: A
Review," Journal of Current Issues and Research in Advertising, 15, 2 (Fall 1993),
25-58.

Naples, Michael (1979). Effective Frequency: The Relationship between Frequency and
Advertising Effectiveness. New York, NY: Association of National
Advertisers, 63-82.

Olney, T. J., Rajeev Batra and Morris B. Holbrook (1990). ”A Three-Component
Model of Attitude Toward the Ad: Effects of the Zipping and Zapping of
Television Commercials,” in Emotion in Advertising: Theoretical and Practical
Explorations, Stuart J. Agres, Julie A. Edell and Tony M. Dubitsky, editors.
New York NY: Quorum Books, 269-281.

Page, Thomas J., Esther Thorson and Maria Papas Heide (1990). ”The Memory
Impact of Commercials Varying in Emotional Appeal and Product
Involvement,” in Emotion in Advertising: Theoretical and Practical
Explorations, Stuart J. Agres, Julie A. Edell and Tony M. Dubitsky, editors.
New York NY: Quorum Books, 255-268.

129

Palda, Kristian S. (1966). "The hypothesis of a hierarchy of effects: a partial
evaluation," Journal of Marketing Research, 3 (February), 13-24.

Pechmann, Cornelia and David W. Stewart (1989). ”Advertising Repetition: A
Critical Review of Wearin and Wearout,” Current Issues and Research in
Advertising 1988, James H. Leigh and Claude R. Martin, Jr., editors. Ann
Arbor, MI: University of Michigan, 285-329.

Percy, Larry and John R. Rossiter (1992). "Advertising Stimulus Effects: A
Review," Journal of Current Issues and Research in Advertising, 14, 1, 75-92.

Peter, J. Paul and Jerry Olson (1990). Consumer Behavior and Marketing Strategy,
Second Edition. Homewood, IL: Richard D. Irwin.

Peterson, Robert A. (1994). ”A Meta-analysis of Cronbach’s Coefficient Alpha,”
Journal of Consumer Research, 21, 2, 381-391.

Petty, Richard E. and John T. Cacioppo (1981). Attitudes and Persuasion: Classic and
Con temporary Approaches. Dubuque IA: William C. Brown.

Petty, Richard E., John T. Cacioppo and David Schumann (1983). ”Central and
Peripheral Routes to Advertising Effectiveness: The Moderating Role of
Involvement,” Journal of Consumer Research, 10 (September), 135-146.

Phelps, Joseph and Esther Thorson (1991). ”Brand Familiarity and Product
Involvement Effects on the Attitude Toward the Ad-Brand Attitude
Relationship,” Advances in Consumer Research, 18, Rebecca H. Holman and
Michael R. Solomon, edS.. Provo UT: Association for Consumer Research,
202—209.

Pratkanis, Anthony R. and Anthony G. Greenwald (1985). ”A Reliable Sleeper
Effect in Persuasion: Implications for Opinion Change Theory and
Research,” in Psychological Processes and Advertising Efi‘ects, Linda Alwitt and
Andrew A. Mitchell, editors. Hillsdale, NJ: Lawrence Erlbaum Associates
Inc., 157-173.

Schlinger, Mary Jane (1979). ”A Profile of Responses to Commercials,” Journal of
Advertising Research, 19, 2, 37-46.

Schumann, David W., Richard E. Petty and D. Scott Clemons (1990). "Predicting
the Effectiveness of Different Strategies of Advertising Variation: A Test of
the Repetition-Variation Hypotheses," Journal of Consumer Research, 17, 2
(September), 192-202.

130

Shimp, Terence A. (1981). "Attitude Toward the Ad as a Mediator of Consumer
Brand Choice," Journal of Advertising, 10, 2, 9-15.

Soldow, Gary F. and Victor Principe (1981 ). ”Response to Commercials as a
Function of Program Context," Journal of Advertising Research, 21 (2), 59-65.

Stewart, David W., Connie Pechmann, Srinivasan Ratneshwar, Jon Stroud and
Beverly Bryant (1985). "Methodological and Theoretical Foundations of
Advertising Copytesting: A Review," Current Issues and Research in
Advertising.

Stewart, David W. (1986). "The moderating role of recall, comprehension and

brand differentiation on the persuasiveness of television advertising,"
Journal of Advertising Research, (April/ May), 43-47.

Stout, Patricia and John D. Leckenby (1986). ”Measuring Emotional Response to
Advertising," Journal of Advertising, 15 (4), 35-42.

Thorson, Esther and Marian Friestad (1989). "The Effects of Emotion on Episodic
Memory for Television Commercials,” in Cognitive and Aflective Responses to
Advertising, editors, Patricia Cafferata and Alice M. Tybout. Lexington MA:
D.C. Heath and Company, 305-326.

Thorson, Esther, Annie Chi and Clark Leavitt (1992). "Attention, Memory,
Attitude and Conation: A Test of the Advertising Hierarchy," Advances in
Consumer Research, 19, John F. Sherry and Brian Stemthal, eds. Provo, UT:
Association for Consumer Research, 366-379.

Thorson, Esther, and Thomas J. Page, Jr. (1988). ”Effects of Product Involvement
and Emotional Commercials on Consumers’ Recall and Attitudes,” in
Nonverbal Communication in Advertising, Sidney Hecker and David W.
Stewart, eds., Lexington, MA: Lexington Books, 111-126.

Villareal, Angelina (1985). ”Message Context and Content as Meidators of
Advertising Evaluation: The Influence of Feeling States," AMA Educators’
Conference Proceedings, Michael J. Houston and Richard J. Lutz, eds., 167-
171.

Yoon, Kak (1991). ”A Pretesting Model of Advertising Effectiveness: The
Mediating Effects of Involvement, Picture and Message," Unpublished
doctoral dissertation. East Lansing MI: Michigan State University.

Zaichkowsky, Judith Lynne (1985). "Measuring the Involvement Construct,"
Journal of Consumer Research, 3 (December), 341-352.

131

Zajonc, Robert B. (1968). "Attitudinal Effects of Mere Exposure," Journal of
Personality and Social Psychology, 9 (2, Part 2), 1-28.

Zinkhan, George M., William B. Locander and James H. Leigh (1986).
”Dimensional Relationships of Aided Recall and Recognition," Journal of
Advertising, 15(1), 38-46.

Zinkhan, George M., and Claude R. Martin, Jr. (1983). ”Message Characteristics
and Audience Characteristics: Predictors of Advertising Response,"
Advances in Consumer Research, 10, Richard P. Bagozzi and Alice M. Tybout,
eds. Ann Arbor, MI: Association for Consumer Research, 27—31.

TE UNIV. LIBRRR

nICHrcaN sra
llllllllllllllHUI|
31293

ll lullIllllll“

015796562