University of Southern California Dissertations and Theses

How sequels seduce: consumers' affective expectations for entertainment experiences

(USC Thesis Other)

How sequels seduce: consumers' affective expectations for entertainment experiences

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content HOW SEQUELS SEDUCE: CONSUMERS’ AFFECTIVE EXPECTATIONS FOR
ENTERTAINMENT EXPERIENCES
by
Justin Anderson
____________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)
August 2007
Copyright 2007 Justin Anderson
ii
DEDICATION
With thanks to my former students: My enjoyment teaching you motivated me to
finish this dissertation and pursue a career in academia.
And to my future students: May I always be worthy of you.
iii
ACKNOWLEDGEMENTS
Many thanks are due to my dissertation committee chair, Deborah J. MacInnis, and
committee members, C. Whan Park, Valerie S. Folkes, Rand Wilcox, and Peter
Vorderer, for their guidance and contributions to this dissertation.
iv
TABLE OF CONTENTS
Dedication ii
Acknowledgements iii
List of Tables v
List of Figures vi
Abstract vii
Introduction
The Entertainment Industry
Purpose and Overview
1
2
6
Chapter 1: Affective Expectations
Theoretical Development
Methodology: Study 1
Discussion
12
12
16
22
Chapter 2: Brand Extension Forecasts and Confidence
Theoretical Development
Methodology: Study 2
Discussion
27
27
35
47
Chapter 3: Expectations of Defensive Pessimists
Theoretical Development
Methodology: Study 3
Discussion
51
51
57
68
Discussion
Contributions
Limitations
Future Directions
73
73
76
77
References 85
Appendices
Appendix A: Study 1 Manipulations
Appendix B: Study 1 Measures
Appendix C: Study 2 Manipulations
Appendix D: Study 2 Measures
Appendix E: Study 3 Manipulations
Appendix F: Study 3 Measures
92
92
93
94
96
97
98
v
LIST OF TABLES
Table 1: Study 1 Confirmatory Factor Analysis Model Comparison 19
Table 2: Study 1 Confirmatory Factor Analysis Factor Loadings 20
Table 3: Study 1 Manipulation Check Results 21
Table 4: Study 1 Hypothesis Test Results 22
Table 5: Study 2 Confirmatory Factor Analysis Model Comparison 39
Table 6: Study 2 Confirmatory Factor Analysis Factor Loadings 40
Table 7: Study 2 Manipulation Check Results 42
Table 8: Study 2 Hypothesis Test Results 44
Table 9: Study 2 Hypothesis H3 Mediation Test Results 46
Table 10: Study 3 Confirmatory Factor Analysis Model Comparison 60
Table 11: Study 3 Confirmatory Factor Analysis Factor Loadings 61
Table 12: Study 3 Hypothesis H4 Test Results 64
Table 13: Study 3 Hypothesis H5 Test Results 65
Table 14: Study 3 Hypothesis H6 Test Results 67
Table 15: Study 3 Hypothesis H7 Test Results 68
vi
LIST OF FIGURES
Figure 1: Study 1 Conceptual Model 14
Figure 2: Study 2 Conceptual Model 33
Figure 3: Study 3 H5 and H6 Conceptual Model 54
Figure 4: Study 3 H7 Conceptual Model 56
Figure 5: Study 3 H8 Conceptual Model 56
vii
ABSTRACT
Entertainment goods have two important characteristics: 1) they primarily provide
affective benefits and 2) their quality cannot be determined prior to consumption.
Therefore, when evaluating future entertainment experiences, consumers form
expectations about the emotional gratification that they will receive, which depends
upon the entertainment brand being evaluated and the consumer’s characteristics,
among other factors. This dissertation explores the process by which consumers
form affective expectations and examines the impacts of brand strategy and
consumer characteristics upon this process. The first study explores the process by
which consumers form affective expectations, finding that consumers form affective
expectations by weighting their affective forecast by their confidence in that forecast.
The second study examines the influence of brand extensions on this process, finding
that the similarity of a brand extension, such as a movie sequel or new book by a
serial author, to its parent brand facilitates the transfer of associations from the
parent brand to the brand extension, and that these associations form the basis for the
affective forecast for the extension, which, in turn, impact affective expectations for
the extension. Findings also suggest that similarity increases consumers’ confidence
in their affective forecasts for brand extensions, which impacts affective expectations
for the extensions. The third study addresses a consumer characteristic, finding that
consumers who are defensively pessimistic generate more negative thoughts about
brand extensions than new brands, resulting in lower affective forecasts and
expectations. Furthermore, this study finds that defensive pessimism strengthens the
role of similarity on the transfer of associations from parent brands to brand
viii
extensions, and also strengthens the role of forecast confidence in the formation of
consumers’ affective expectations. These findings contribute to affective forecasting
literature by addressing the impact of forecast confidence, and by proposing a model
of how consumers form affective expectations. Findings also contribute to this
literature by addressing the impact of defensive pessimism on affective expectations.
Finally, findings contribute to the literature on brand extensions by demonstrating
the role that brand strategy plays in the affective expectations process, and how this
role is influenced by defensive pessimism.
1
INTRODUCTION
The following is a description for a fictional movie called Confederate Gold:
“After decoding several clues hidden in artifacts on display in the Smithsonian
museum, an archaeologist and his colleagues embark on an adventure to find a
treasure that was hidden long ago by the leaders of the Confederacy during the
American Civil War. However, another team of treasure hunters is also searching
for the same treasure. After his friends are kidnapped by the bad guys, the
archaeologist must choose between salvaging the lost treasure and saving his friends’
lives.” Based on this description, what emotional experience do you forecast that the
movie would provide you? How confident are you of that forecast? What are your
expectations about how Confederate Gold would make you feel? Now, imagine that
the movie is actually a sequel to National Treasure, in which an archaeologist
decodes clues hidden in the Declaration of Independence and other historic artifacts
to find a treasure chest rumored to have been hidden by the leaders of the American
Revolution. How do your associations for National Treasure impact your forecast of
how Confederate Gold would make you feel, and your confidence in that forecast?
How would your forecast change if you very much enjoyed National Treasure, and
your positive memory for that movie would be harmed if Confederate Gold failed to
satisfy you? This dissertation begins to answer these questions. It contributes to
marketing literature by exploring the process by which consumers form affective
expectations, and examining the impact of brand extensions and consumer
characteristics on this affective expectations process. It also identifies managerial
implications for the marketing of entertainment experiences.
2
The Entertainment Industry
The entertainment industry is an important part of the U.S. economy. Census
data indicates that the motion picture and book publishing industries earned $62.9
billion and $27.9 billion, respectively, in 2002 (Census Bureau 2002). Broadcast and
cable television accounted for $58.4 billion, combined. Another $141.9 billion went
to the arts, entertainment, and recreations industries, which include performing arts,
spectator sports, museums, gambling, and amusement parks. These figures total
$291.1 billion, representing a sizable part of the estimated U.S. GDP of $12.4 trillion
(CIA 2005). Given the importance of the entertainment industry to our economy, it
is imperative to understand factors that might impact consumers’ choices in an
entertainment context, such as their expectations for entertainment experiences, and
how marketing principles impact these expectations in this important industry.
Academic research on entertainment consumption is mainly conducted in two
disciplines: communications and marketing. Entertainment-focused communications
scholars have explored motivations for consuming entertainment (e.g., Bosshart and
Macconi 1998; Vorderer 2001), the processes by which entertainment provides
pleasure (e.g., Green et al 2004, Tannenbaum 1980, Zillman 1994), and the effects of
entertainment consumption on individuals and society (e.g., Hansen and Hansen
2000, Mundorf and Laird 2002, Zillmann 1988). In examining these issues,
communications scholars have developed a body of knowledge about why people
consume entertainment, in general. Notably, these studies do not address questions
of interest to marketers, such as factors that impact consumers’ choice of
entertainment experiences (e.g., movie, book, video game, etc.). Thus, this literature
3
is informative for the study of entertainment consumption behavior, but it leaves the
study of consumption-related questions to marketing scholars.
Marketing scholars have taken up this task, to a small degree. Several studies
have developed econometric models of motion picture box office revenues (e.g.,
Eliashberg and Shugan 1997, Neelamegham and Jain 1999, Sawhney and Eliashberg
1996). Another study explores the competitive issues involved in motion picture
release timing (Krider and Weinberg 1998). Although these studies address issues
relevant to managers in the motion picture industry, they focus on movies, as
opposed to other entertainment contexts. As a result, neither their empirical
measures (e.g., actors’ “star power,” MPAA rating, etc.) nor their findings generalize
beyond the motion picture category. These studies also focus on an aggregate level,
with movies as the unit of analysis. The psychological forces that drive consumers’
choices are not examined.
One exception to the above is the research conducted by Holbrook, who
explored such entertainment-related issues as emotions (Havlena and Holbrook
1986), experiential consumption (Holbrook and Hirschman 1982), and the play
processes involved with video game consumption (Holbrook et al 1984). However,
rather than exploring consumers’ evaluation and choice processes, those studies
resemble communications literature in their focus on how consumers experience
entertainment. Other studies explore popular-versus-expert judgments of motion
pictures (Holbrook 1999), consumers’ use of leisure time (Holbrook and Lehmann
1981), and the differences between hedonic and utilitarian consumption (Hirschman
and Holbrook 1982). Although these papers address entertainment issues, the
4
analyses are category-specific, limiting the generalizability of the findings to
entertainment consumption behavior, in general.
In summary, the existing literature on entertainment consumption is either (1)
too broadly theoretical about how consumers experience entertainment, without
providing specific marketing implications, or (2) too narrowly managerial, lacking
generalizable findings that would apply in multiple entertainment categories. The
lack of a theoretical understanding of the factors that influence consumers’
entertainment choices presents a daunting challenge to marketing scholars who
would pursue research into consumer behavior in the entertainment industry.
The characteristics of entertainment experiences, in relation to product and
service categories, further illustrate both the importance and the challenge of
researching consumer behavior in the entertainment industry. Unlike many products
and services that are chosen primarily for their functional or symbolic benefits,
entertainment experiences are often chosen primarily for the emotional gratification
they are expected to provide (Tannenbaum 1980, Zillman 1988). While some
products and services may provide affective benefits in addition to their utilitarian
value (e.g., soup may evoke happiness due to nostalgia), and other goods may
provide hedonic sensory pleasure (e.g., ice cream, a massage, etc.), entertainment
experiences primarily provide affective benefits. Entertainment experiences are also
experience goods, meaning that consumers cannot determine their quality prior to
consumption (Nelson 1970). As a result, consumers’ choices for entertainment
experiences are based on expectations of their affective benefits, rather than on
attributes of the entertainment experience obtained through information search.
5
Therefore, to understand consumers’ entertainment choices, it is necessary to
examine how consumers form expectations of the affective gratification that they
expect an entertainment experience will provide. While affective expectations have
started to penetrate the marketing literature, much remains to be learned about
consumer and marketing relevant factors that impact these expectations.
Another important characteristic of entertainment experiences is that they
tend to have short introductory periods, during which consumers form relatively
stable mental representations of the experience. These brief periods allow marketers
limited opportunity to change an unsuccessful positioning strategy. In other product
and service categories, a firm might introduce a brand with one intended brand
image, or value proposition. If consumers fail to respond to that brand image, the
firm can reposition the brand until it hits on a value proposition that generates
consumer interest. However, entertainment brands cannot be easily repositioned
once they have been produced, distributed, and promoted. It is therefore crucial that
marketers know how a chosen brand strategy will impact consumers’ affective
expectations of the entertainment experience. Entertainment marketers need to
understand the way consumers think about entertainment brands and brand
extensions, so that they may develop successful marketing programs that effectively
communicate desirable affective benefits. This dissertation addresses one such brand
strategy – the use of brand extensions, often called sequels in an entertainment
context.
6
Purpose and Overview
The purpose of this dissertation is to explore how consumers form affective
expectations of entertainment experiences, as well as to examine how brand and
consumer characteristics impact these expectations.
Chapter 1 explores how consumers evaluate potential entertainment
experiences by examining the process by which consumers form affective
expectations. Affect is a term that encompasses both emotion and mood (Bagozzi et
al 1999). Whereas affective forecasts are consumers’ beliefs about their future
emotional states (Richard et al 1996), affective expectations are defined as
probabilistic evaluations of future affective states (Mellers et al 1997). Most studies
in the realm of affective forecasting have examined the outcomes of affective
forecasts on choice behavior (e.g., Lopes 1984, 1987, 1990) and post-consumption
behavior (e.g., MacInnis et al 2005, Patrick et al 2007). A few papers have
examined the moderating effects of specific emotions (e.g., regret) on the
relationship between affective forecasting and choice (e.g., Simonson 1992,
Zeelenberg et al 1996). However, little attention has been devoted to the process by
which consumers form affective expectations. Drawing on decision affect theory
(Mellers et al 1997), this study proposes that consumers form affective expectations
by weighting their affective forecasts by their confidence in those forecasts, or the
perceived probability that the forecasted outcome will occur.
Study 1 used a 2 (positive versus neutral affective forecast) x 2 (high versus
low forecast confidence) experimental design to test the hypothesis that affective
expectations are determined by consumers’ affective forecasts and the confidence
7
with which these forecasts are held. Analysis of this empirical data supports the
hypothesis. This study contributes to the literature on affective expectations by
proposing a conceptual model of how consumers form affective expectations. It also
contributes to entertainment research in marketing by proposing an affective
expectations process in an entertainment context. This study also provides a
foundation for the analyses in the subsequent studies of this dissertation, which
explore brand and consumer effects on affective expectations.
Chapter 2 examines how brand characteristics impact affective expectations.
Specifically, it suggests that the perceived similarity between the brand extension
and the parent brand impacts the two antecedents to affective expectations –
affective forecasts and forecast confidence. Studies consistently find that the
similarity between the parent brand and brand extension impacts consumers’ brand
extension evaluations (Aaker and Keller 1990, Boush et al 1987). This study extends
these findings by focusing on affective expectations (vs. evaluations) and by
suggesting that similarity influences affective expectations by impacting affective
forecasts. The transfer of associations from the parent brand to the brand extension
also provides consumers with more information on which to evaluate the brand
extension. Since belief strength increases with more information (Berger and
Mitchell 1989) and prior brand experience (Fazio and Zanna 1978, Wright and
Lynch 1995), this study also proposes that greater similarity between a parent brand
and brand extension impacts affective expectations by increasing consumers’
affective forecast confidence.
8
Study 2 used a 2 (positive versus neutral parent brand affective experience) x
2 (high versus low brand extension similarity) experimental design to test the
hypotheses that parent-extension similarity moderates the impact of parent brand
affective experiences on affective forecasts for a brand extension, and directly
impacts consumers’ confidence in their affective forecasts for a brand extension.
Analyses of this empirical data support the hypotheses. Affective forecasts for brand
extensions are positively related to the recalled affective experiences of the parent
brand, but this effect only holds for extensions that are perceived to be similar
(versus dissimilar) to the parent brand. Parent brand experiences also influence
affective expectations for brand extensions, as mediated by affective forecasts. The
study also supports the proposition that increased perceived similarity between the
parent brand and brand extension increases the consumer’s confidence in their
affective forecast. This study contributes to the affective forecasting and
expectations literatures by suggesting that judgments of brand extensions are based
on consumers’ affective experiences with the parent brand. It contributes to brand
extension literature by suggesting that (1) affective experiences of the parent brand
impact brand extension judgments, and (2) similarity impacts not only brand
extension evaluations, but also affective forecasts, forecast confidence, and affective
expectations.
Chapter 3 adopts a consumer (versus managerial) focus to investigate how a
consumer characteristic – defensive pessimism - impacts affective expectations.
Defensive pessimism is a mindset by which consumers manage satisfaction by
setting artificially low expectations about a future outcome (e.g., believing they will
9
fail an exam) and taking defensive action (e.g., study harder) to avoid a negative
outcome (Sanna 1996). In the context of entertainment consumption, the negative
outcome might be dissatisfaction with the affective experience. However,
consumers cannot take action to change the quality of the entertainment experience.
Instead, this study proposes that defensively pessimistic consumers will manage their
satisfaction with an entertainment experience by reducing their affective expectations
for the entertainment experience. These lowered affective expectations maximize
the likelihood of a positive gap between experienced and expected affect, which may
in turn improve their satisfaction with the entertainment experience. Brand
extensions have implications not only for consumers’ experience of the extension,
but also for their brand image of the parent brand (Boush et al 1987). Therefore, this
study proposes that defensive pessimism will have stronger effects on consumers’
negative thoughts and affective forecasts for brand extensions than for new brands,
because brand extensions have greater potential for creating consumer
dissatisfaction. Furthermore, because defensive pessimists have been shown to be
more involved in their evaluations than other consumers (Norem 2001, Sanna 1996),
the study proposes that defensive pessimists place greater emphasis on their
similarity judgments when creating affective forecasts, and on their forecast
confidence when forming affective expectations.
Study 3 used a 2 (brand extension versus new brand) x 2 (high versus low
defensive pessimism) quasi-experimental design to test the hypotheses that defensive
pessimism and brand strategy impact the number of negative thoughts generated,
affective forecasts, and expectations, and that defensive pessimism impacts the
10
moderating effects found in Studies 1 and 2. Analyses of the empirical data support
the hypotheses. Consumers who are defensively pessimistic about a brand extension
generate more negative thoughts about the extension, resulting in lower affective
forecasts for the extension, compared to defensive pessimists evaluating a new brand
and compared to non-defensive pessimists. Furthermore, defensive pessimists rely
more on perceived similarity between the parent brand and brand extension when
creating affective forecasts and on forecast confidence when forming affective
expectations. This study contributes to affective forecasting and expectations
literatures by investigating the influence of brand extensions and defensive
pessimism on affective forecasts and expectations. It also demonstrates the
implications of defensive pessimism for brand extension similarity and affective
forecast confidence. Results suggest that defensive pessimism may be an important
phenomenon that deserves much greater study.
These three studies propose, test, and support a conceptual model of how
consumers form affective expectations for entertainment experiences, and how brand
extension similarity and defensive pessimism influence those expectations. These
findings contribute to extant research on affective forecasting, affective expectations,
and brand extensions. The findings also offer implications for marketers of
entertainment experiences. Because forecast confidence has a significant impact on
affective expectations, marketers should try to make consumers more confident
about their affective forecasts, provided that these forecasts are positive. One way to
do so is to make brand extensions seem more similar to their parent brands, although
this strategy also increases the threat of brand dilution due to dissatisfaction (Boush
11
et al 1987), which should also be taken into account. Another way, even outside of a
brand extension context, is to make reference to consumers’ past experiences, so
they may more confidently create their affective expectations, even for a new brand.
Because defensive pessimism negatively impacts affective forecasts, marketers
should be aware of this phenomenon, and measure consumers’ defensive pessimism.
This would help marketers to understand how defensive pessimism may impact
consumers’ affective forecasts and expectations, and their entertainment purchase
intentions. (However, it is not suggested that marketers try to make consumers more
or less defensively pessimistic – only that they should be aware of consumers’
defensive pessimism and its potential impacts on their affective forecasts and
expectations). It is hoped that both researchers and practitioners will benefit from
this dissertation.
12
CHAPTER 1: AFFECTIVE EXPECTATIONS
As described in the Introduction, the entertainment industry is a large and
growing part of the U.S. economy. Recognizing the importance of this industry,
several marketing studies have developed econometric models to predict box office
revenue for motion pictures (e.g., Eliashberg and Shugan 1997, Neelamegham and
Jain 1999, Sawhney and Eliashberg 1996). Although these studies have revealed
some interesting findings, they offer little psychological rationale to explain why
consumers choose to see some movies and not others. Because entertainment
experiences are consumed primarily for their affective benefits (Tannenbaum1980,
Zillmann 1988), and because consumers cannot determine their quality prior to
consumption (Nelson 1970), this study looks for answers to the question of why
consumers choose to consume certain entertainment experiences over others by
focusing on the realm of affective expectations.
Theoretical Development
Affective expectations are defined as probabilistic evaluations of future
affective states (Mellers et al 1997). Affect is an abstract term that includes mood
and emotion (Bagozzi et al 1999). Mood is an enduring, low-intensity, general
feeling of positivity, neutrality, or negativity, such as feeling “good” or “unpleasant.”
In contrast, emotion is a short-lived, high-intensity, specific response to a stimulus,
such as feeling “happy,” “proud,” or “guilty” about something. This dissertation
proposes that a major antecedent to an affective expectation is an affective forecast,
defined as a consumer’s prediction of a future affective state (Richard et al 1996),
which includes valence (i.e., positive, neutral, or negative), intensity, and duration
13
(Gilbert et al 1998). The theoretical arguments developed in this dissertation are
intended to apply to affective forecast valence, intensity, and duration, but the
empirical studies will not address duration.
Affective forecasts and affective expectations are similar, yet distinct,
concepts. Whereas an affective forecast represents a consumer’s belief about a
potential affective outcome, an affective expectation incorporates the perceived
probability with which that forecasted outcome will occur. Thus, affective
expectations should be influenced by affective forecasts that have been adjusted by
their probability of occurrence. This probability, or likelihood of outcome
occurrence, can be thought of as a consumer’s confidence in their affective forecast
(Bell 1982).
While limited research has examined affective expectations, there is a small
but growing body of literature on affective forecasting. Some of this research
examines the affect associated with a product or service offering, and how the affect
forecasted to occur with product choice impacts actual choices (Lopes 1984, 1987,
1990). Other affective forecasting research has explored the role of specific
emotions (e.g., regret) as a moderating influence on the relationship between
affective forecast and choice (Simonson 1992, Zeelenberg et al 1996). Still other
articles have explored the importance of consumers’ confidence in their forecasts
(Bell 1982, Mellers et al 1997). Even the post-consumption influences of affective
forecasts on consumer satisfaction have received study (Patrick et al 2007). Finally,
consumers’ inaccuracy at affective forecasting has been explored. Findings suggest
that consumers are notoriously poor at forecasting their affective valence (Gilbert et
14
al 2004, Woodzicka and LaFrance 2001), intensity (Buehler and McFarland 2001,
Mitchell et al 1996), and duration (Gilbert et al 1998, Wilson et al 2000). Despite
these examinations of the outcomes of affective forecasts and misforecasts, little is
known about the branding and consumer variables that might impact affective
forecasts, forecast confidence, and affective expectations.
Decision affect theory provides some insight into the relationship between
affective expectations and affective forecasts. The theory proposes that affective
expectations are probabilistic assessments of the likelihoods of possible affective
responses (Mellers et al 1997). Although not explicitly stated in the theory, this idea
suggests that affective expectations are distinct from affective forecasts, such that
affective expectations are confidence-weighted averages of affective forecasts. This
theory has received empirical support in the context of affective expectations
associated with monetary gains and losses, but the theory should also apply to
strictly affective outcomes without any ties to utilitarian outcomes.
Building upon decision affect theory, Study 1 suggests that consumers form
an affective expectation (i.e., probabilistic evaluation) as an interaction between their
affective forecast (i.e., predicted future outcome) and their confidence in that
forecast (i.e., the perceived probability that the outcome will occur).
FIGURE 1: STUDY 1 CONCEPTUAL MODEL
Affective
Expectations
Affective
Forecast
Forecast
Confidence
H1 +
15
The proposed effect illustrated by Figure 1 is specified by the first
hypothesis:
H1: Consumers’ affective expectations are positively related to their affective
forecast, and the strength of this relationship is positively related to the
consumers’ forecast confidence.
It is important to distinguish this model of affective expectations from two
similar concepts, expected value and multi-attribute attitude models. In the former,
consumers calculate expected value as the multiplicative combination of a possible
outcome and the probability associated with that outcome, summed over all possible
outcomes (Cobb and Douglas 1928). However, the expected value model assumes
that the possible outcomes are known, and have a known probability distribution. In
contrast, the affective forecast includes many unknown factors, including the specific
emotions that will be experienced, the intensity of those emotions, and their duration.
Furthermore, the probability distributions of those outcomes are not known. Thus,
the expected utility model is a computation of “known unknowns,” or unknown
outcomes with known likelihoods, while the affective expectations model is an
“unknown unknown,” or unknown outcomes with unknown likelihoods.
The affective expectations model also differs from multi-attribute attitude
models. In these models, attitude toward an object is computed as the multiplicative
combination of a belief about an attribute of that object and the evaluation of that
attribute, summed over all of the object’s attributes (Fishbein and Ajzen 1972,
Wilkie and Pessemier 1973). The difference between this concept and that of
affective expectations is the difference between evaluation and confidence, where
16
evaluation is the consumers’ judgment of whether an attribute is desirable or
undesirable, and confidence represents the consumers’ judgment of the likelihood of
the forecasted outcome. The multi-attribute attitude model weights beliefs by their
evaluation. Although affective forecasts are beliefs about the consumer’s future
affective state, the affective expectations model weights these beliefs by their
likelihood. Thus, the three concepts of expected values, multi-attribute attitudes, and
affective expectations are conceptually distinct, despite their similarities.
Methodology: Study 1
The hypothesis proposed above suggests a model of affective expectations
whereby affective forecasts and forecast confidence interact to impact consumers’
affective expectations. This model was tested using a 2x2 between-subjects
experimental design. Four conditions were created to represent two levels of
affective forecast (positive versus neutral) and two levels of forecast confidence
(high versus low).
Data were collected from 127 subjects sampled from a population of
undergraduate students enrolled in an introductory marketing course at a large,
private university of the Western U.S. Subjects were randomly assigned to one of
the four conditions. The manipulations for each condition are shown in Appendix 1.
All subjects were asked to imagine a scenario in which they were planning to
see a comedy movie. Subjects in the positive (neutral) affective forecast conditions
were informed that the movie stars an actor whom they think is extremely funny
(somewhat funny, but not extremely funny), and that the movie should be able to
make the subject feel extremely happy (somewhat happy, but not extremely happy).
17
Subjects in the high (low) forecast confidence conditions were told that the actor’s
movies always (never) make the subject feel the same way, and that critics agree
(disagree) about how similar this movie is to the actor’s previous movies, making the
subject very confident (not at all confident) about how funny the movie will be.
Subjects were then asked to respond to fifteen closed-ended items (see
Appendix 2). These items measured (a) affective expectations for the movie, (b)
affective forecasts for the movie, (c) confidence in these affective forecasts, (d)
overall attitude toward the movie, (e) the risk probability, or likelihood of a negative
outcome occurring, and (f) the risk consequence, or negative impact of a negative
outcome. The attitude variable was included to demonstrate its distinction from
affective expectations and forecasts, and the risk variables were included to
demonstrate their distinction from forecast confidence.
Affective Expectations was measured with three 7-point items ranging from
strongly disagree to strongly agree in response to the following statements: (i) I
expect that watching this movie will make me very happy, (ii) Watching this movie
probably won’t put me in a positive mood, and (iii) There is a very good chance that
watching this movie will make me laugh.
Affective Forecast was measured with three 7-point items ranging from (i)
neutral to very happy, (ii) neutral to very good, and (iii) neutral to very positive in
response to the statement, “I forecast that after watching this movie, I would feel….”
Forecast Confidence was measured with three 7-point items ranging from
strongly disagree to strongly agree in response to the following statements: (i) I am
very confident about how this movie will make me feel, (ii) I am very uncertain
18
about how this movie will affect my mood, and (iii) I can accurately predict how this
movie will affect my emotions.
Attitude was measured with the following two items: (i) My overall opinion
of this movie is …, rated on a 7-point range from very bad to very good and (ii) In
general, I would (like/dislike) this movie, rated on a 7-point range from dislike to
like.
Risk probability was measured with two 7-point items ranging from strongly
disagree to strongly agree in response to the following statements: (i) There is a very
good chance that this movie will not satisfy me and (ii) I am likely to be
disappointed by this movie.
Risk consequence was measured with two 7-point items ranging from
strongly disagree to strongly agree in response to the following statements: (i) If I
see this movie and don’t like it, I will feel bad and (ii) If I see this movie and don’t
like it, I will be very unhappy.
The first three constructs - affective expectations, affective forecast, and
forecast confidence - are the key variables that will be used to test Hypothesis 1. The
other three constructs - attitude, risk probability, and risk consequence - are
potentially similar to the relevant constructs, and are measured here to determine the
discriminant validity of affective expectations, affective forecast, and forecast
confidence from attitude, risk probability, and risk consequence.
Confirmatory Factor Analyses: The first step of this analysis was to test the
hypothesized factor model against a model in which all items loaded onto a single
construct. Table 1 displays the model fit statistics for a single-factor solution and for
19
the hypothesized factor solution. The hypothesized model has a better model fit,
according to the AIC index and other parameters. Therefore, it model can be
accepted over the single-factor model. This indicates that there is discriminant
validity between the three constructs of interest. It also indicates that the constructs
of affective expectations, affective forecast, and forecast confidence are distinct from
attitude, risk probability, and risk consequence.
TABLE 1: STUDY 1 CONFIRMATORY FACTOR ANALYSIS MODEL
COMPARISON
Model Fit Parameter Single-factor Model Hypothesized Model
Chi-square 5.764, df = 90, p<0.000 1.576, df = 75, p<0.001
RMSEA 0.194 0.068
AIC (saturated) 578.801 (240.000) 208.169 (240.000)
Note: Single-factor model: all items loading onto one latent factor
Hypothesized model: items loading onto their hypothesized latent construct
The second step of the CFA is to examine the factor loadings of each item
onto its hypothesized construct. Table 2 displays the factor loadings of each item
onto its latent constructs. (Note: confirmatory factor analysis only generates factor
loadings for an item’s hypothesized construct; off-factor loadings are not available.)
As indicated, each item loads onto its hypothesized construct with a significance of
p<0.000. Furthermore, all of the positively-worded items have strong loadings,
while the two negatively-worded items (Q2, Q8) have slightly weaker – but still
significant – loadings. This analysis suggests that there is convergent validity
between the items that measure each construct. Given the discriminant validity
between constructs and the convergent validity within each construct, the CFA
indicates that the measured items are valid measures for this analysis.
20
TABLE 2: STUDY 1 CONFIRMATORY FACTOR ANALYSIS FACTOR
LOADINGS
Item Factor 1 Factor 2 Factor 3
1. I expect that watching this movie will make
me very happy.
0.797
2. Watching this movie probably won’t put me
in a positive mood. (RS)
0.706
3. There is a very good chance that watching
this movie will make me laugh a lot.
0.819
4. I forecast that after watching this movie, I
would feel: Very Sad – Very Happy
0.919
5. I forecast that after watching this movie, I
would feel: Very Negative – Very Positive
0.829
6. I forecast that after watching this movie, I
would feel: Very Bad – Very Good
0.872
7. I am very confident about how this movie will
make me feel.
0.930
8. I am very uncertain about how this movie will
affect my mood. (RS)
0.714
9. I can accurately predict how this movie will
affect my emotions.
0.822
Note: Factor 1: Affective Expectations
Factor 2: Affective Forecast
Factor 3: Forecast Confidence
Factor loadings indicate standardized regression weights
All loadings are significant at p<0.000
RS indicates reverse-scored items
Manipulation Checks: Table 3 displays the manipulation checks for Study 1.
Results indicate a successful affective forecast manipulation. Affective forecast was
significantly greater in the positive affective forecast condition (M = 5.60) than the
neutral affective forecast condition (M = 4.77; F = 36.879, p<0.000). There were no
significant effects of the forecast confidence manipulation on affective forecast (F =
3.547, p<0.062) or their interaction (F = 0.354, p<0.553). Results also indicate a
successful forecast confidence manipulation. Forecast confidence was significantly
greater in the high forecast confidence condition (M = 5.20) than the low forecast
21
confidence condition (M = 4.26; F = 14.750, p<0.000). There were no significant
effects of the affective forecast manipulation (F = 1.552, p<0.215) or their interaction
(F = 0.008, p<0.927).
TABLE 3: STUDY 1 MANIPULATION CHECK RESULTS
Neutral Affective
Forecast
Positive Affective
Forecast
Measure
Low
Forecast
Confidence
High
Forecast
Confidence
Low
Forecast
Confidence
High
Forecast
Confidence
Affective Forecast 4.68
a
(0.64)
4.85
a
(0.75)
5.43
b
(0.69)
5.77
b
(0.97)
Forecast Confidence 4.10
a
(1.46)
5.06
b
(1.27)
4.43
a
(1.38)
5.34
b
(1.37)
Note: Results are shown as mean (standard deviation)
Means in a given row with the same superscript are not significantly different
at p<0.05.
Hypothesis Testing: Hypothesis 1 proposes that consumers’ affective
expectations are influenced by the interaction of their affective forecast and their
confidence in that forecast. Table 4 displays the hypothesis test results. There is a
significant positive effect of the affective forecast on affective expectations, such
that affective expectations are greater when the affective forecast is positive than
when it is neutral (F = 68.484, p<0.000). Additionally, there is a marginally
significant interaction effect, such that the effect of the affective forecast on affective
expectations is stronger when forecast confidence is high than when it is low (F=
3.607, p<0.060). Although this interaction effect is only marginally significant when
conducting an ANOVA analysis using discrete independent variables, a regression
analysis using continuous measures of affective forecast and forecast confidence
yields a significant effect of both affective forecast ( = 0.467, p<0.000) and its
22
interaction with forecast confidence ( = 0.211, p<0.017). Therefore, results support
Hypothesis 1, suggesting that affective forecasts positively influence affective
expectations, and that forecast confidence has a positive impact on this relationship.
TABLE 4: STUDY 1 HYPOTHESIS TEST RESULTS
Neutral Affective
Forecast
Positive Affective
Forecast
Measure
Low
Forecast
Confidence
High
Forecast
Confidence
Low
Forecast
Confidence
High
Forecast
Confidence
Affective Expectations 4.45
a
(1.02)
4.51
a
(1.09)
5.55
b
(0.71)
6.10
c
(0.79)
Note: Results are shown as mean (standard deviation)
Means in a given row with the same superscript are not significantly different
at p<0.05.
Discussion
The results of Study 1 support the theoretical model proposed in Figure 1.
Consumers’ affective expectations are positively impacted by their affective
forecasts, and this relationship is strengthened by forecast confidence. This finding
provides a specific contribution to understanding how consumers form expectations
about entertainment experiences, such as movies. Consumers forecast the emotional
experience that the movie will provide, determine how confident they are in that
forecast, and combine the two to generate an expectation for the affective
experience.
This process should easily generalize to other entertainment categories, such
as literature, video games, and music. However, at a theoretical level, this affective
expectations model can be generalized to any category in which there is a significant
affective component to the value proposition. For instance, if a consumer desires to
23
eat a bowl of soup not only for its nutritional value and hunger-relief, but also to
obtain feelings of warmth and nostalgia, then these affective benefits have an impact
on the consumers’ expectations for the consumption experience, and the affective
expectations model developed in Study 1 should apply. Therefore, while this model
has been developed and tested in an entertainment context, it is proposed to apply to
a much broader array of consumption categories.
One limitation to this study is that the manipulations were artificial. Subjects
were told to imagine themselves in a situation, to imagine a forecast of how a movie
would make them feel, and that they were very confident (or not) about their
forecasts. However, subjects were not presented with the description of an actual
movie to evaluate for themselves. In a realistic context, consumers would probably
demonstrate much more individual variation in their affective forecasts and forecast
confidence than is observed in this experimental study. Testing this model with a
survey design and having consumers rate their affective expectations for a real movie
would provide a more realistic test of this model. However, reducing this individual
variance allows for a better test of the hypothesized model, and allows the effects of
affective forecast and forecast confidence to be examined without interference from
other, unexplained factors.
A second potential limitation to this study is that it was only tested in the
context of movies. This context was purposely used because most of the subjects in
the sample were ages 18-22, which is part of the heaviest group of movie consumers
in the U.S. (MPAA 2004). Because these subjects can be expected to be very
knowledgeable about their consumption behavior in this entertainment category,
24
motion pictures seems to be an appropriate context in which to conduct this study. It
is further noted that this contextual limitation is prevalent in other academic
marketing studies on entertainment. However, whereas those studies have tended to
use very specific measures related to movies (e.g., MPAA rating, star actor power,
etc.), this study uses measures that can be generalized to any other entertainment
category (e.g., literature, video games, music, etc.), and even non-entertainment
categories (e.g., soup). Testing this model in other entertainment and non-
entertainment categories will be left for future exploration.
These findings are intended to advance the literature on affective
expectations by proposing and testing a model of how consumers form affective
expectations. Although several other studies have examined the effects of affective
expectations on choice decisions (e.g., Lopes 1984, 1987, 1990) and post-
consumption satisfaction (Patrick et al 2004), very little work has explored the
process by which consumers create these affective expectations. This study provides
such a model.
Using this model as a framework, other affective forecasting effects may be
examined in greater detail. For example, one finding has been that anticipated regret
can alter the effect of affective forecasts on consumers’ choices (Simonson 1992,
Zeelenberg et al 1996). It is possible that the model developed in this study can help
to explain that effect. Specifically, it may be the case that when consumers desire to
avoid feeling regret about their consumption choices, they may place greater weight
on their forecast confidence in creating their affective expectations. That is, a
25
consumer may form more positive affective expectations for an experience that they
are more confident will not produce regret than one about which they are less certain.
These findings are also intended to advance the literature on entertainment
consumption by suggesting a framework for studying how consumers develop
affective expectations for entertainment experiences. As stated previously,
entertainment experiences provide affective benefits (Tannenbaum 1980, Zillman
1988) and are experience goods that cannot be evaluated prior to consumption
(Nelson 1970). As a result, prior to purchase, consumers cannot fully evaluate the
affective experience that an entertainment experience will provide. The model
developed here suggests how consumers develop those affective expectations, which
can help explain their entertainment consumption behavior. While many previous
marketing studies of entertainment issues have focused on econometric analyses of
motion picture box office revenue, they have been lacking in providing a theoretical
foundation for their findings. This study proposes to advance this literature by
suggesting that affective expectations can be used as a consumer psychology variable
that may help to explain past empirical findings, as well as to stimulate interesting
research questions for the future study of entertainment-related consumer behavior in
marketing.
One area in which affective expectations can be used to explore other
marketing phenomena is in the realm of brand extensions, which are commonly
referred to as sequels in the entertainment category. Brand strategy is a managerial
decision with significant implications for consumer response. The next chapter
examines the impact that brand extensions have on affective expectations, affective
26
forecasts, and forecast confidence. Understanding these impacts may have important
implications for marketers’ brand strategy decisions.
27
CHAPTER 2: BRAND EXTENSION FORECASTS AND CONFIDENCE
Study 1 finds that affective forecasts and forecast confidence impact
consumers’ affective expectations. An interesting follow-up to this study would
therefore examine managerially-relevant factors that influence those two inputs. One
such factor is the use of a brand extension strategy. As discussed in the Introduction,
one interesting characteristic of the entertainment industry is that marketers often get
only one chance to position their offering and generate a consumer purchase
response. Therefore, entertainment marketers face an important need to brand their
offerings properly right from the start. One way that they attempt this is to introduce
brand extensions, often called sequels in the entertainment industry. Movie sequels
have been found to generate significant consumer interest, tending to earn greater
box office revenue than original films, after controlling for promotional expenditures
and distribution intensity (Anderson 2003). Evidence suggests a similar
phenomenon for books by serial authors, as well (New York Times 2006). This
study intends to address this important phenomenon, by proposing that brand
extensions can strongly impact affective expectations through their antecedents,
affective forecasts and forecast confidence.
Theoretical Development
The study of brand extensions has been extensive. Some studies explore the
motivations for firms to introduce brand extensions instead of new brands (Sullivan
1992, Wernerfelt 1988). Others explore consumers’ attitudes toward brand
extensions as a function of parent brand characteristics (Aaker and Keller 1990,
Dacin and Smith 1994, Gurhan-Canli 2003, Park et al 1991). Still others explore the
28
feedback effects of brand extensions on parent brand evaluations (Gurhan-Canli and
Maheswaran 1998, John et al 1998, Loken and John 1993, Milberg et al 1997). The
following section reviews the literature on brand extensions, and discusses its
applicability to affective forecasts and forecast confidence.
Considerable research explores how consumers evaluate brand extensions.
One factor that impacts consumers’ brand extension evaluations is their evaluation of
the parent brand (Aaker and Keller 1990, Bhat and Reddy 2001, Bottomley and
Doyle 1996, Bottomley and Holden 2001, Broniarczyk and Alba 1994, DelVecchio
2000, Keller and Aaker 1992). Categorization theory suggests that consumers
categorize a brand extension as belonging to the parent brand family. Since
consumers have no experience with the brand extension, they draw on their
evaluations of the parent brand to form evaluations of the brand extension.
Consumers transfer their affect for the parent brand to the brand extension through a
process known as affect transfer. As a result, there is a positive relationship between
parent brand evaluations and evaluations of brand extensions. That is, the more
favorable the evaluations of the parent brand, the more favorable the evaluations of
the brand extension.
An important factor moderating the relationship between parent brand
attitude and brand extension evaluation is the degree of similarity between the two.
In general, the more closely-related or similar the parent brand and brand extension
are perceived to be, the more easily parent brand associations are transferred to the
brand extension (Aaker and Keller 1990, Bottomley and Doyle 1996, Bottomley and
Holden 2001, Boush et al 1987, Boush and Loken 1991, Czellar 2003). This finding
29
is derived from categorization theory (Boush et al 1987, Lakoff 1987), which
suggests that the more similar or related a brand extension is to the parent brand, the
more closely the two will be categorized in the consumer’s mind, and the more
readily associations will be transferred between them. Thus, the relationship
between parent brand evaluations and brand extensions will be stronger when there
is greater similarity between the products than when similarity is lower.
Similarity has been further deconstructed into various dimensions. One
dimension is the fit between the physical product features of the parent brand and
those of the brand extension (Aaker and Keller 1990, Boush et al 1987, Boush and
Loken 1991, Park et al 1991). Another dimension is the consistency between the
brand concepts associated with the parent brand and the brand extension (Park et al
1991). A final type of similarity is goal congruency (Martin and Stewart 2001),
which refers to extensions that satisfy common goals as the parent brand. Any of
these similarity dimensions may enhance the relationship between parent brand
evaluations and evaluations of the brand extension. Ultimately, similarity is
perceived and evaluated by each individual consumer, leaving the consumer free to
determine similarity in whatever way they deem best.
As this brief literature review demonstrates, research examining brand
extensions is considerable. However, the extent to which these studies of brand
extensions are generalizable to entertainment brand extensions is limited by the fact
that they primarily explore consumers’ attitude toward an external object (i.e., the
brand extension or the parent brand), as opposed to the consumer’s expectations
about their own internal experiences. While attitudes represent evaluations of a
30
brand, entertainment goods are designed to produce affective responses in the
consumer (Tannenbaum 1980, Zillmann 1988). Thus, an exploration of
entertainment experiences differs from attitude research in two ways. First, the
nature of the judgment in attitude research is an evaluation (i.e., liking), versus affect
(i.e., feelings) in entertainment research. Because attitudes and affect have been
shown to be distinct concepts (Richard et al 1996), it is important to examine the
proper construct in a research context. Second, the focus of the judgment in attitude
research is on an external object (e.g., a brand), versus the self (i.e., the consumers’
experienced affect) in entertainment research. This study of entertainment brand
extensions seeks to address these issues by exploring the effects of brand extensions
on consumers’ affective expectations (versus attitudinal evaluations) of their own
future affective states (versus an external object).
As supported by the results of the first study, affective expectations are
influenced by the interaction of affective forecasts and forecast confidence. In
examining the effects of brand extensions on affective expectations, then, the impact
of brand extensions on each of these factors must be studied.
Brand Extension Effects on Affective Forecasts and Expectations: When
considering the consumption of an entertainment brand extension, a consumer
develops an affective expectation for the extension. As suggested in Chapter 1,
affective expectations are significantly influenced by the affective forecast, which is
the consumer’s prediction of a future affective state (Richard et al 1996). This may
include the valence (i.e., positive, neutral, or negative) of how they will feel, the
intensity or strength of this valenced feeling, and the duration or how long the
31
affective state will last (Gilbert et al 1998). (Note: although the theoretical model
can accommodate all three of these affective components, only the valence and
intensity are examined in this study.) The key issue addressed in this chapter is how
marketers can influence how consumers create their forecasts.
Consumers may create an affective forecast from marketing communications,
word-of-mouth referrals, and other sources of information. For instance, by
watching a movie trailer or reading a book jacket, a consumer can form some idea of
how they might feel after watching the movie or reading the book. However, this
may be true of any entertainment experience, regardless of whether it is a new brand
or a brand extension. The difference between brand extensions and new brands is
that, when the experience is a brand extension, consumers have additional
information available for forecasting their experience. That additional information is
their recalled experience of consuming the extension’s parent brand.
In existing brand extension research conducted in traditional product
categories, consumers are found to transfer parent brand attitude and affect to the
brand extension (Aaker and Keller 1990, Boush et al 1987). For entertainment brand
extensions, the relevant memory associations are not only the overall attitude or
affect toward the parent brand, but also the affective experience that the parent brand
evoked in the consumer, since entertainment experiences are primarily consumed for
their affective benefits (Tannenbaum 1980, Zillmann 1988). For instance, imagine a
consumer facing the scenario described in the introduction. When forecasting their
affective experience for Confederate Gold, the consumer may draw upon their
memory of the affective experience evoked by National Treasure, and use this
32
recalled affective experience as a benchmark for creating an affective forecast for the
sequel (Phillips and Baumgartner 2002, Sujan et al 1993). Thus, it is proposed here
that the consumer recalls from memory how they felt as a result of consuming the
parent brand, and transfers this association to the brand extension. This recalled
affective experience, then, influences the consumer’s affective forecast for the brand
extension.
Furthermore, this effect is proposed to be moderated by the similarity of the
brand extension to the parent brand. As described previously, similarity has
repeatedly been found to have a significant moderating effect on brand extension
evaluations. Specifically, the transfer of associations from a parent brand to a brand
extension is enhanced by greater similarity between the two. For instance, referring
back to the scenario described in the introduction, if the consumer perceives that
Confederate Gold will be very similar to National Treasure, the affective experience
recalled for the parent movie will have a significant effect on the affective forecast
for the sequel. However, in the case of an apparently dissimilar sequel, this may not
be the case. If the consumer perceives that Confederate Gold is rather dissimilar to
National Treasure, the affective experience recalled for the parent movie may have
little or no impact on the consumer’s affective forecast for the sequel. The proposed
model shown in Figure 2 suggests that consumers form an affective forecast for a
brand extension as an interaction between their recalled affective experience with the
parent brand and the perceived similarity between the parent brand and the brand
extension. This interaction also has an impact on affective expectations for the brand
extension, as mediated by affective forecast for the extension.
33
FIGURE 2: STUDY 2 CONCEPTUAL MODEL
The proposed effects illustrated by Figure 2 are specified by the following
hypotheses:
H2: A consumer’s affective forecast for an entertainment brand extension
becomes more positive as the affective experience that the consumer recalls
for the parent brand becomes more positive. This effect becomes
significantly stronger as the similarity between the parent brand and brand
extension increases.
H3: A consumer’s affective expectation for an entertainment brand extension
becomes more positive as the affective experience that the consumer recalls
for the parent brand becomes more positive. This effect becomes
significantly stronger as the similarity between the parent brand and brand
extension increases. This effect is mediated by the consumer’s affective
forecast for the brand extension.
Brand Extension Effects on Forecast Confidence: As discussed in the review
of brand extensions literature, when evaluating a brand extension, consumers have
repeatedly been found to take similarity into account. Transfers of associations from
Parent Brand
Affective
Experience
Parent-Extension
Similarity
H2 +
Brand Extension
Forecast Confidence
H4 +
Brand Extension
Affective
Forecast
Brand Extension
Affective
Experience
H3 +
34
parent brand to brand extension are facilitated by greater similarity between the two
items (Aaker and Keller 1990, Boush et al 1987). Furthermore, similarity has been
decomposed into different kinds of similarity, including product feature similarity
(Park et al 1991), brand concept consistency (Park et al 1991), and goal congruency
(Martin and Stewart 2001), any of which can facilitate affect transfer. The same is
proposed to be true for consumers’ recalled affective experiences for parent brands,
which are suggested (in H2 and H3) to influence their affective forecasts and
expectations for brand extensions more as the similarity between the two increases.
In addition to this effect, however, it is suggested that similarity increases
consumers’ confidence in their affective forecasts. When two goods are perceived to
be similar, consumers may categorize them closely in memory, and transfer more
recalled affective experiences from the parent brand to the brand extension (Boush et
al 1987). For instance, a consumer may identify a movie sequel (i.e., brand
extension) as starring the same main characters, having a very similar plot, and
starring the same lead actors as the original movie (i.e., parent brand). These
character, plot, and actor attributes may be perceived as product features, and cause
the consumer to perceive strong product feature similarity between the two movies.
Additionally, the consumer may see a trailer for the movie sequel that makes it
appear to be in the same genre and designed to evoke the same emotional response as
the original movie. This affective characteristic of the two movies may be
considered to be its brand concept, resulting in high brand concept consistency
between the films. Finally, the consumer might recall that the original movie really
succeeded in elevating his mood when he was having a bad day, and perceive that
35
the sequel will have the same mood-improving effects. This can be considered an
example of goal congruency between the two movies. By perceiving the brand
extension as being very similar to the parent brand on multiple dimensions of
similarity, the consumer has transferred more recalled affective associations from the
parent to the extension. Having more information available for evaluation increases
consumers’ confidence in their evaluations (Fazio and Zanna 1981, Wright and
Lynch 1995). Therefore, the more similar a brand extension is to its parent brand,
the more information the consumer has on which to evaluate the brand extension,
and the more confident the consumer will be in their affective forecast for the
extension. This gives rise to the following hypothesis, also displayed in Figure 2:
H4: A consumer’s affective forecast confidence for a brand extension
increases as the perceived similarity of the extension to its parent brand
increases.
Methodology: Study 2
Hypotheses 2-4 were tested using a 2x2 between-subjects experimental
design. Four conditions were created to represent two levels of parent brand
affective experience (positive vs. neutral) and two levels of parent-extension
similarity (high vs. low).
Data were collected from 84 subjects drawn from a population of
undergraduate students enrolled in an introductory marketing course at a large,
private university in the Western U.S. Subjects were randomly assigned to one of
the four conditions. The manipulations for each condition are shown in Appendix 3.
36
All subjects were asked to imagine a scenario in which they were at a
bookstore shopping for a book to read for pleasure. They were told to imagine that
they saw a new book by an author whose books they had read before. Subjects in the
positive (neutral) parent brand affective experience condition were told that they had
enjoyed the author’s previous books very much (somewhat, but not very much), and
that the author’s previous books had made the subject feel very happy (not made
them feel very happy). Subjects in the high (low) similarity condition were told that
the new book seems to be very similar to (very different from) the books that the
authors has written before, because it is in the same (different) genre and stars the
same (different) main character as the author’s previous books.
Subjects were then asked to respond to nineteen closed-ended items (see
Appendix 4), which measured (a) affective expectations for the book, (b) affective
forecasts for the book, (c) forecast confidence for the book, (d) perceived similarity
of the new book to the author’s previous books, and (e) affective experiences for the
author’s previous books.
Affective Expectations were measured with four 7-point items ranging from
strongly disagree to strongly agree in response to the following statements: (i) I
expect that this book will make me very happy, (ii) This book probably won’t put me
in a positive mood, (iii) There is a very good chance that this book will make me feel
good, and (iv) It is very likely that this book will make me laugh a lot.
Affective Forecast was measured with three 7-point items ranging from (i)
neutral to very happy, (ii) neutral to very good, and (iii) neutral to very positive.
37
Forecast Confidence was measured with three 7-point items ranging from
strongly disagree to strongly agree in response to the following statements: (i) I am
very confident about how this book will make me feel, (ii) I am very uncertain about
how this book will affect my mood (reverse-scored), and (iii) I can accurately predict
how this book will affect my emotions.
Parent-Extension Similarity was measured on five 7-point items ranging from
strongly disagree to strongly agree in response to the following items: (i) This book
is very similar to the author’s previous books, (ii) The characters in this book are
very similar to characters in the author’s previous books, (iii) The plot or storyline in
this book is very similar to the plot of the author’s previous books, (iv) The setting
(time, place, etc.) of this book is very similar to the setting of the author’s previous
books, and (v) This book seems a lot like the author’s previous books.
Parent Brand Affective Experience was measured on four 7-point items
ranging from not much to very much in response to the following: (i) How happy did
the author’s previous books make you feel?, (ii) How satisfied did the author’s
previous books make you feel?, (iii) How positive did the author’s previous books
make you feel?, and (iv) How good did the author’s previous books make you feel?
Confirmatory Factor Analyses: Table 5 displays the confirmatory factor
analysis (CFA) for the nineteen items measured in Study 2. The first step of this
analysis was to test the hypothesized factor model against a model in which all items
loaded onto a single construct. Table 5 displays the model fit statistics for a single-
factor solution and for the hypothesized factor solution. The hypothesized factor
model has a better model fit according to the AIC index and other parameters.
38
Therefore, this model can be accepted over the single-factor model. This indicates
that there is discriminant validity between the four constructs of interest.
39
TABLE 5: STUDY 2 CONFIRMATORY FACTOR ANALYSIS MODEL
COMPARISON
Model Fit Parameter Single-factor Model Hypothesized Model
Chi-square 7.092, df=209, p<0.000 2.515, df=209, p<0.000
RMSEA 0.271 0.135
AIC (saturated) 1614.265 (550.000) 657.736 (550.000)
Note: Single-factor model: all items loading onto one latent factor
Hypothesized model: items loading onto their hypothesized latent construct
The second step of the CFA is to examine the factor loadings of each item
onto its hypothesized construct. These loadings are shown in Table 6. As indicated,
each item loads onto its hypothesized construct with a significance of p<0.000. This
analysis suggests that there is convergent validity between the items that measure
each construct. Given the discriminant validity between constructs and the
convergent validity within each construct, the CFA indicates that the fifteen items
are valid measures for this analysis.
40
TABLE 6: STUDY 2 CONFIRMATORY FACTOR ANALYSIS FACTOR LOADINGS
Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
1. I expect that this book will make me very happy. 0.952
2. This book probably won’t put me in a positive mood. (RS) 0.739
3. There is a very good chance that this book will make me feel good. 0.937
4. It is very likely that this book will make me laugh a lot. 0.814
5. I forecast that after watching this movie, I would feel: Neutral –
Very Happy
0.957
6. I forecast that after watching this movie, I would feel: Neutral –
Very Good
0.978
7. I forecast that after watching this movie, I would feel: Neutral –
Very Positive
0.940
8. I am very confident about how this book will make me feel. 0.928
9. I am very uncertain about how this book will affect my mood. (RS) 0.728
10. I can accurately predict how this book will affect my emotions. 0.769
11. This book is very similar to the author’s previous books. 0.947
12. The characters in this book are very similar to the characters in the
author’s previous books.
0.959
13. The plot or storyline in this book are very similar to the plot in the
author’s previous books.
0.828
14. The setting (time, place, etc.) of this book is very similar to the
setting of the author’s previous books.
0.774
15. This book seems a lot like the author’s previous books. 0.972
41
TABLE 6: STUDY 2 CONFIRMATORY FACTOR ANALYSIS FACTOR LOADINGS, CONTINUED
Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
16. How happy did the author’s previous books make you feel? 0.964
17. How satisfied did the author’s previous books make you feel? 0.958
18. How positive did the author’s previous books make you feel? 0.940
19. How good did the author’s previous books make you feel? 0.969
Note: Factor 1: Affective Expectations
Factor 2: Affective Forecast
Factor 3: Forecast Confidence
Factor 4: Parent-Extension Similarity
Factor 5: Parent Brand Affective Experience
Factor loadings indicate standardized regression weights
All loadings are significant at p<0.000
RS indicates reverse-scored items
42
Manipulation Checks: Table 7 displays the manipulation checks for Study 2. The
manipulation check for recalled affective experience indicates that that manipulation
was successful. Parent brand affective experience was significantly greater in the
positive condition (M = 6.10) than in the neutral condition (M = 3.08; F = 163.074,
p<0.000). There was no significant effect of main of parent-extension similarity (F =
0.040, p<0.841) or their interaction (F = 0.912, p<0.342). Thus, the parent brand
affective experience manipulation was successful and confounding effects can be
rejected.
TABLE 7: STUDY 2 MANIPULATION CHECKS
Neutral Parent Brand
Affective Experience
Positive Parent Brand
Affective Experience
Measure
Low
Parent-
Extension
Similarity
High
Parent-
Extension
Similarity
Low
Parent-
Extension
Similarity
High
Parent-
Extension
Similarity
Parent Brand
Affective Experience
3.21
a
(1.43)
2.94
a
(1.36)
6.01
b
(0.57)
6.19
b
(0.69)
Parent-Extension
Similarity
2.60
a
(1.17)
5.63
b
(0.93)
2.42
a
(1.08)
5.74
b
(1.05)
Note: Results are shown as mean (standard deviation)
Means in a given row with the same superscript are not significantly different
at p<0.05.
The manipulation check for parent-extension similarity indicates that that
manipulation was also successful. Parent-extension similarity was significantly
greater in the high (M = 5.69) than the low similarity condition (M = 2.51; F =
188.090, p<0.000). There was no significant effect of the parent brand affective
experience (F = 0.021, p<0.886) or their interaction (F = 0.406, p<0.526). Thus, the
43
parent-extension similarity manipulation was successful and confounding effects can
be rejected.
Hypothesis Testing: Hypothesis 2 suggests that consumers’ affective
forecasts for a brand extension are positively and significantly influenced by their
recalled affective experiences with the parent brand, and that this effect becomes
significantly stronger as the perceived similarity between the parent brand and brand
extension increases. Table 8 displays the results of the ANOVA analysis conducted
to test this hypothesis. The results indicate that consumers’ affective forecasts for
the brand extension were, in fact, influenced by the recalled affective experience for
the parent brand (F = 34.146, p<0.000), but only when the similarity between the
parent brand and extension was high (interaction: F = 37.153, p<0.000).
Specifically, when parent-extension similarity was high, subjects’ affective forecasts
for the brand extension were significantly more positive when the parent brand
affective experience was positive (x = 5.95, s = 0.95) than when the recalled
affective experience was neutral (x = 2.19, s = 1.44). However, when parent-
extension similarity was low, subjects’ affective forecasts for the brand extension
were not significantly different when the parent brand affective experience was
positive (x = 3.54, s = 1.58) than when it was neutral (x = 3.62, s = 1.69). Thus, H2
is supported.
It is interesting to note that greater similarity is not always a positive
attribute. Similarity causes greater transfer of associations from the parent brand to
the brand extension. Thus, in the positive parent brand affective experience
condition, affective forecasts for the brand extension were significantly higher when
44
similarity was high (x = 5.95, s = 0.95) than when it was low (x = 3.54, s = 1.58). In
However, in the neutral parent brand affective experience condition, affective
forecasts for the brand extension were significantly lower when similarity was high
(x = 2.19, s = 1.44) than when it was low (x = 3.62, s = 1.69). Thus, while greater
similarity increases the affective forecast for a brand extension of a positively-
recalled parent brand, it decreases the affective forecast for a brand extension of a
neutrally-recalled parent brand.
TABLE 8: STUDY 2 HYPOTHESIS TEST RESULTS
Neutral Parent Brand
Affective Experience
Positive Parent Brand
Affective Experience
Measure
Low
Parent-
Extension
Similarity
High
Parent-
Extension
Similarity
Low
Parent-
Extension
Similarity
High
Parent-
Extension
Similarity
Brand Extension
Affective Forecast
3.62
a
(1.69)
2.19
b
(1.44)
3.54
a
(1.58)
5.95
c
(0.95)
Brand Extension
Affective Expectation
3.86
a
(1.13)
3.01
b
(1.11)
4.42
a
(1.04)
5.98
c
(0.85)
Brand Extension
Forecast Confidence
3.60
a
(1.63)
4.97
b
(1.44)
3.60
a
(1.35)
5.79
c
(1.02)
Note: Results are shown as mean (standard deviation)
Means in a given row with the same superscript are not significantly different
at p<0.05.
Hypothesis 3 suggests that consumers’ affective expectations for a brand
extension are also positively and significantly influenced by their recalled affective
experiences with the parent brand, and that this effect becomes significantly stronger
as the perceived similarity between the parent brand and brand extension increases.
This effect, however, is expected to be mediated by the affective forecast. Table 8
indicates that consumers’ affective expectations for the brand extension were, in fact,
45
influenced by the recalled affective experience for the parent brand (F = 60.267,
p<0.000), but only when the similarity between the parent brand and extension was
high (interaction: F = 28.067, p<0.000). Specifically, when parent-extension
similarity was high, subjects’ affective expectations for the brand extension were
significantly more positive when the parent brand affective experience was positive
(x = 5.98, s = 0.85) than when the recalled affective experience was neutral (x =
3.01, s = 1.11). However, when parent-extension similarity was low, subjects’
affective forecasts for the brand extension were not significantly different when the
parent brand affective experience was positive (x = 4.42, s = 1.04) than when it was
neutral (x = 3.86, s = 1.13). Thus, H3 is supported.
However, Hypothesis 3 also predicts that the interaction effect of parent
brand affective experience and parent-extension similarity on affective expectation is
mediated by the affective forecast. Table 9 displays the results of the regression
analysis used to test this mediation hypothesis (following the guidelines in Baron and
Kenny 1986). The results indicate that affective expectations are positively and
significantly influenced by the interaction of parent brand affective experience and
parent-extension similarity (Model 1: b = 0.076, p<0.000) and affective forecast
(Model 2: b = 0.657, p<0.000) when each variable is included separately. However,
when both explanatory variables are included in the same model (Model 3), there is a
discernable decrease in the influence of the interaction between parent brand
affective experience and parent-extension similarity on affective expectations.
Although the effect of this interaction is still significant, it is less significant when
the affective forecast is accounted for than when the affective forecast is omitted.
46
Thus, this analysis suggests support for H3. The affective forecast partially mediates
the effect of the interaction of parent brand affective experience and parent-extension
similarity on affective expectations.
TABLE 9: STUDY 2 HYPOTHESIS H3 MEDIATION TEST RESULTS
Model 1 Model 2 Model 3
Parameter
Parent Brand
Affective Experience
x Parent-Extension
Similarity
Affective
Forecast
Parent Brand
Affective Experience
x Parent-Extension
Similarity
Affective
Forecast
b 0.076 0.657 0.020 0.577
0.649 0.866 0.168 0.760
t-statistic 7.715 15.670 2.421 10.956
p-value 0.000 0.000 0.018 0.000
Note: The dependent variable is Affective Expectations for all regressions.
Hypothesis 4 suggests that greater parent-extension similarity increases
consumers’ confidence in their affective forecasts for a brand extension. Table 8
displays the results of the ANOVA analysis conducted to test this hypothesis. The
results indicate that similarity does, in fact, have a positive and significant impact on
forecast confidence. When the recalled affective experience was positive, subjects’
forecast confidence for the brand extension was significantly greater when the
parent-extension similarity was high (x = 5.79, s = 1.02) than when the parent-
extension similarity was low (x = 3.60, s = 1.35). Additionally, when the recalled
affective experience was neutral, subjects’ forecast confidence for the brand
extension was significantly greater when the parent-extension similarity was high (x
= 4.97, s = 1.44) than when the parent-extension similarity was low (x = 3.60, s =
1.63). The main effect of parent-extension similarity on forecast confidence was
therefore significant (F = 34.849, p<0.000).
47
There was no significant effect of parent brand affective experience on
forecast confidence (F = 1.878, p<0.174). The 2x2 ANOVA also found no
significant effect of the interaction (F = 1.878, p<0.174). However, a one-way
comparison between the two high similarity conditions did find greater forecast
confidence when the parent brand experience was positive (x = 5.79, s = 1.02) than
when it was neutral (x = 4.97, s = 1.44), which is a significant effect (F = 4.560,
p<0.039). There was no significant difference when similarity was low. The
difference in confidence when similarity is high is difficult to explain, as there is
little reason to suspect that the valence of the parent brand affective experience
should have an impact on the forecast confidence for the extension, and there was no
confound between the two manipulations. This may be a random effect in the data,
or perhaps future research can explain why there may be such an interaction. It is
also possible that this is the result of a positivity bias. This may explain why the
difference is significant in the positive parent brand affective experience condition,
but absent in the neutral condition. Nevertheless, in both the positive and neutral
parent brand affective experience conditions, greater parent-extension similarity was
associated with greater brand extension forecast confidence. Thus, H4 was
supported.
Discussion
The results of Study 2 support the conceptual model proposed in Figure 2.
As this model suggests, consumers’ affective forecasts and affective expectations for
brand extensions are significantly influenced by their recalled affective experienced
with the parent brand, provided that the brand extension is perceived to be similar to
48
the parent brand. Furthermore, consumers’ confidence in these affective forecasts is
significantly increased by greater perceived similarity between the brand extension
and the parent brand. Thus, in the example at the start of the Introduction, when
considering Confederate Gold, a consumer would recall the affective experience they
had with National Treasure, and determine how similar the sequel seems to be to its
parent movie. If Confederate Gold seems to be very similar to National Treasure,
then the affective experience of the parent movie would heavily shape the affective
forecast for the sequel, and consumers would be very confident about their forecasts.
However, if the sequel seems to be very different from its predecessor, then the
affective experience of National Treasure may have no impact on affective forecasts
for Confederate Gold, and consumers would be less confident about their forecasts.
A limitation to this study is that the manipulations were artificial. Subjects
were told to imagine themselves in a situation, to imagine how an author’s books
made them feel, and to imagine that a new book is very similar or different from
what this author had written before. However, subjects were not presented with the
description of an actual book to evaluate for themselves. In a realistic context,
consumers would probably demonstrate much more individual variation in their
recalled affective experiences and similarity judgments than is observed in this
experimental study. Testing this model with a survey design and having consumers
rate their affective expectations for a real book would provide a more realistic test of
this model. However, reducing this individual variance allows for a better test of the
hypothesized model, and allows the effects of parent brand affective experience and
49
parent-extension similarity to be examined without interference from other,
unexplained factors.
These findings advance the literature on brand extensions by proposing and
testing a model of the effects of brand extensions on affective expectations and its
predictors, affective forecasts and forecast confidence. Although previous studies
have examined the effects of brand extensions on consumers’ attitudinal evaluations
of an external object, prior research has ignored consumers’ self-oriented affective
forecasts of how they will feel as a result of consuming a good. This study addresses
this previously omitted issue of self-referenced affect. Specifically, this study
provides evidence to suggest that consumers develop affective forecasts of how
consuming a brand extension will make them feel based on their recalled affective
experiences with the parent brand. Adding this issue of self-referenced affect to
brand extension literature helps to provide a more complete understanding of the
effects of brand extensions on consumers’ information processing, decision making,
and choice behavior.
These findings also advance the literature on brand extensions by introducing
the concept of forecast confidence to the research domain. Previous studies of brand
extensions have focused on brand extension evaluations as the sole dependent
outcome. This study suggests that brand extension characteristics also impact
consumers’ confidence in those evaluations (or forecasts). This is especially
important for understanding brand extension effects for experience goods, such as
entertainment goods, because these require consumers to develop a forecast of a
future outcome. However, the notion of confidence may also extend to attitude
50
certainty, which can be of importance even for consumer evaluation of brand
extensions (and new brands) in the domain of search goods. Thus, addressing the
issue of forecast confidence contributes to the literature on brand extensions, helping
to provide a more complete understanding of brand extension effects.
These findings also advance the consumer behavior literature on
entertainment. Chapter 1 developed a conceptual model of how consumers create
affective expectations, which is highly relevant to consumer behavior in an
entertainment context. Chapter 2 expanded on this model by identifying the brand
extension effects that act as antecedents to the affective expectations model. By
applying theory developed in another marketing domain (i.e., brand extensions
literature), Chapter 2 has begun to explore the usefulness of the affective
expectations model for studying consumer behavior in an entertainment context, and
made contributions back to that foundational domain.
Another such domain would be to examine individual differences that would
suggest how some consumers might behave differently than others when developing
their affective forecasts and expectations. The next chapter addresses this issue by
exploring the impact that the consumer characteristic of defensive pessimism has on
the affective expectations processes developed in Chapters 1 and 2.
51
CHAPTER 3: EXPECTATIONS OF DEFENSIVE PESSIMISTS
Chapter 1 developed a model of how consumers form affective expectations,
and Chapter 2 identified brand extension characteristics that influence the inputs to
affective expectations. Although these findings may generally hold for most
consumers in most contexts, it can be assumed that there are individual consumer
traits and contextual states that would alter how consumers form their affective
expectations, and how brand extensions impact those expectations. This third study
addresses one such consumer factor, defensive pessimism.
Theoretical Development
Defensive pessimism is a self-esteem protection strategy (Kammen 1989)
that “involves setting unrealistically low expectations in a risky situation in an
attempt to harness anxiety so that performance is unimpaired” (Norem and Cantor
1986, p. 1208). For example, students who are anxious about an impending
examination may adopt a strategy of defensive pessimism. This involves setting low
expectations and using anxiety as a motivator to study harder, thereby improving
performance (Sanna 1996). The motivating influence of defensive pessimism has
been documented in studies regarding educational assessment performance (Martin
et al 2001, Norem 2001, Sanna 1996), and has also been shown to impact the social
behavior of new college students trying to make friends at school (Martin et al 2003).
These studies often view defensive pessimism in an active sense. That is,
people who engage in defensive pessimism take action to improve their performance.
The fact that their actual performance exceeds their expectations would lead to
satisfaction, according to the expectations disconfirmation theory of satisfaction
52
(Oliver 1980). However, in an entertainment context, consumers often are unable to
improve the performance of an entertainment experience. For example, when a
consumer watches a movie or reads a book, the quality of that experience has been
determined when the movie was filmed or the book was published, and the consumer
cannot change the quality of that experience during consumption. Although
consumers cannot increase their satisfaction by actively improving the quality of the
entertainment experience, they can do so by reducing their expectations for the
entertainment experience. This is a more passive form of defensive pessimism,
whereby consumers do not take action to improve performance, but rather reduce
their expectations, thereby increasing their satisfaction with the entertainment
experience.
Defensive pessimism is not a stable personality trait, but rather a contextual
state characteristic (Norem 2001, Sanna 1996). Therefore, the same students who
are defensively pessimistic about an upcoming exam might not be defensively
pessimistic about an entertainment experience, and vice versa. It is important, then,
to question whether consumers would ever be defensively pessimistic about an
entertainment experience. Two factors that influence the use of defensive pessimism
suggest that defensive pessimism could, in fact, apply to entertainment situations.
First, people are more likely to become defensively pessimistic when they are not in
control of an outcome (Martin et al 2001). Because consumers often cannot control
the quality of an entertainment experience, this lack of control would suggest the
potential for defensive pessimism to occur. Second, people are more likely to
become defensively pessimistic when the outcome has self-related implications
53
(Kammen 1989). Entertainment experiences certainly meet this criterion, as people
often consume entertainment to manage their mood (Zillmann 1988), establish their
self-image (Vorderer 2001, Zillmann 2005), and other self-related goals (Bosshart
and Macconi 1998). Therefore, it is reasonable to expect that consumers may engage
in defensive pessimism regarding entertainment experiences.
One specific context in which entertainment consumers could be expected to
be defensively pessimistic is when forming affective forecasts for a brand extension
of an entertainment experience that they enjoyed very much. Imagine a consumer
who very much enjoyed National Treasure and is now forming an affective forecast
for Confederate Gold. If the sequel fails to live up to the consumers recalled parent
brand affective experience, not only will the consumer be dissatisfied with the
sequel, but their brand image for the parent movie may also be diluted. Past research
shows that dissatisfying brand extensions can have negative impacts on their parent
brands (Loken and John 1993). To prevent such dilution, the consumer would
become defensively pessimistic, setting a low affective forecast for Confederate
Gold, which would result in a low affective expectation. This low expectation would
make it easier for the movie to satisfy the consumer, and reduce the risk of dilution
of the parent brand image.
One means through which defensive pessimists justify low expectations is by
generating negative thoughts about the outcome (Norem 2001, Sanna 1996). By
imagining negative outcomes or ways in which a negative outcome might occur,
defensive pessimists give credence to their low expectations, and make them seem
more realistic. Therefore, in the entertainment brand extension context described
54
above, the consumer would generate negative thoughts about Confederate Gold,
thereby justifying their reduced affective forecasts for the sequel. For instance, they
might think about how the sequel might be just a repeat of the first movie, with
nothing new to offer, how the actors might not really be motivated to deliver a good
performance, or other negative thoughts. These thoughts might make the movie
seem less appealing, and encourage the consumer to create a lower affective forecast
than they otherwise would have created if they had not been defensively pessimistic
about the experience.
The proposed model shown in Figure 3 suggests that defensively pessimistic
consumers who evaluate a brand extension generate more negative thoughts, which
reduces their affective forecasts and expectations.
FIGURE 3: STUDY 3 H5 AND H6 CONCEPTUAL MODEL
The proposed effects illustrated by Figure 3 are specified by the following
hypotheses:
H5: Defensively pessimistic consumers who evaluate a brand extension
generate more negative thoughts about that brand extension than (a) non-
H5 +
Negative
Thoughts
Affective
Forecast
Defensive
Pessimism
Brand Extension
Strategy
H6 -
55
defensively pessimistic consumers and (b) defensively pessimistic consumers
who evaluate a new brand.
H6: As the number of negative thoughts generated about an entertainment
experience increases, the affective forecast for that experience decreases.
Another interesting effect that defensive pessimism has is that it makes
consumers more involved in their evaluations (Norem 2001). Consumers have been
found to become more cognitively involved in their evaluations when they are more
motivated to process information, and this motivation is increased when the
evaluation or decision has self-related implications (Petty and Cacioppo 1986).
Therefore, consumers who are defensively pessimistic about their affective forecasts
for entertainment brand extensions are likely to be more cognitively involved in
creating their affective forecasts and affective expectations. This has implications
that relate to the conceptual models developed in Chapters 1 and 2.
In Study 1, forecast confidence was shown to have a moderating influence on
the relationship between affective forecasts and affective expectations. In Study 2,
parent-extension similarity was shown to have a moderating influence on the
relationship between recalled parent brand affective experiences and brand extension
affective forecasts. However, the effect of these two moderators, forecast confidence
and similarity, relies on cognitively-involved consumers who will make the effort to
evaluate forecast confidence and similarity. Because defensive pessimists are more
cognitively involved, they are likely to take similarity and forecast confidence into
account when creating their brand extension affective forecasts and affective
56
expectations, respectively. Consumers who are not defensively pessimistic may be
less likely to do so. These effects are illustrated in Figures 4 and 5.
FIGURE 4: STUDY 3 H7 CONCEPTUAL MODEL
FIGURE 5: STUDY 3 H8 CONCEPTUAL MODEL
The proposed effects illustrated by Figures 4 and 5 are specified by the
following hypotheses:
Parent Brand
Affective
Experience
Parent-Extension
Similarity
Brand Extension
Forecast Confidence
Brand Extension
Affective
Forecast
Brand Extension
Affective
Experience
Defensive
Pessimism
H8 +
Affective
Expectations
Affective
Forecast
Forecast
Confidence
Defensive
Pessimism
H7 +
57
H7: The moderating effect of forecast confidence on the relationship between
an affective forecast and affective expectations is greater for defensive
pessimists than for non-defensive pessimists.
H8: The moderating effect of similarity on the relationship between parent
brand affective experience and brand extension affective forecast is greater
for defensive pessimists than for non-defensive pessimists.
Methodology: Study 3
The hypotheses proposed above suggest a conceptual model of how
defensive pessimism influences consumers’ affective forecasts and affective
expectations. This model was tested using a 2x2 between-subjects quasi-
experimental design. Two conditions were created to represent two levels of brand
strategy (brand extension vs. new brand). Defensive pessimism was measured and
dichotomized into two groups (high vs. low).
Data were collected from 106 subjects drawn from a population of
undergraduate students enrolled in an introductory marketing course at a large,
private university in the Western U.S. Subjects were randomly assigned to one of
the two brand strategy conditions. The manipulations for each condition are
included in Appendix 5.
All subjects were asked to read a description of a fictional movie called
Confederate Gold. Subjects in the new brand condition were only presented with the
description of this movie. Subjects in the brand extension condition were told that
the movie was a sequel to the movie National Treasure, and were given a description
of the parent brand movie.
58
Subjects were then asked to respond to twenty closed-ended items (see
Appendix 6), which measured (a) behavioral intention for Confederate Gold, (b)
affective expectations for Confederate Gold, (c) affective forecasts for Confederate
Gold, (d) forecast confidence for Confederate Gold, (e) consumption of National
Treasure, (f) recalled affective experience for National Treasure, (g) similarity of
Confederate Gold to National Treasure, and (h) defensive pessimism.
Behavioral intention was measured with one 7-point item ranging from very
unlikely to very likely in response to the question “How likely are you to see this
movie?”
Affective Expectations were measured with three 7-point items ranging from
disagree to agree in response to the following statements: (i) I expect that I would
enjoy this movie very much, (ii) This movie probably won’t put me in a positive
mood, and (iii) There is a very good chance that this movie will make me feel good.
Affective Forecast was measured with three 7-point items ranging from
disagree to agree in response to the following statements: I forecast that after
watching this movie, I would feel (i) very satisfied, (ii) very bad, and (iii) very
positive.
Forecast Confidence was measured with three 7-point items ranging from
disagree to agree in response to the following statements: (i) I am very confident
about how this movie will make me feel, (ii) I am very uncertain about how this
movie will affect my mood, and (iii) I can accurately predict how this movie will
affect my emotions.
59
Consumption of National Treasure was measured with one yes/no item
asking “Have you seen National Treasure?”
Subjects who responded in the negative skipped to the similarity items, while
subjects who answered in the affirmative continued to the recalled affective
experience items. These included three 7-point items ranging from disagree to agree
in response to the following statements: After watching National Treasure, I felt (i)
very satisfied, (ii) very bad, and (iii) very positive.
Parent-Extension Similarity was measured on three 7-point items ranging
from disagree to agree in response to the following items: (i) Confederate Gold is
very similar to National Treasure, (ii) The plot or storyline in Confederate Gold is
not very similar to the plot of National Treasure, and (iii) Confederate Gold would
make me feel the same emotions as National Treasure.
Defensive Pessimism was measured on three 7-point items ranging from
disagree to agree in response to the following statements: (i) I go into a movie
expecting the worst, even though I know the movie will probably be OK, (ii) I try to
keep my expectations for a movie fairly low, so that I will be pleasantly surprised
with how much I enjoy the movie, and (iii) I usually set very high expectations for
sequels of movies that I really enjoyed.
After completing all of these closed-ended items, subjects were asked to
record all of their thoughts about Confederate Gold. Finally, subjects were asked to
report whether each thought statement was positive, negative, or neutral.
Confirmatory Factor Analyses: Table 10 displays the confirmatory factor
analysis (CFA) for eighteen of the items measured in Study 3. Items 1 (behavioral
60
intention) and 11 (consumption of National Treasure) were omitted from this
analysis. The first step of this analysis was to test the hypothesized factor model
against a model in which all items loaded onto a single construct. Table 10 displays
the model fit statistics for a single-factor solution and for the hypothesized factor
solution. The hypothesized factor model has a better model fit according to the AIC
index and other parameters. Therefore, this model can be accepted over the single-
factor model. This indicates that there is discriminant validity between the
constructs of interest.
TABLE 10: STUDY 3 CONFIRMATORY FACTOR ANALYSIS MODEL
COMPARISON
Model Fit Parameter Single-factor Model Hypothesized Model
Chi-square 3.376, df=135, p<0.000 3.016, df=135, p<0.000
RMSEA 0.150 0.139
AIC (saturated) 563.700 (378.000) 515.123 (378.000)
Note: Single-factor model: all items loading onto one latent factor
Hypothesized model: items loading onto their hypothesized latent construct
The second step of the CFA is to examine the factor loadings of each item
onto its hypothesized construct. These loadings are shown in Table 11. As
indicated, each item loads onto its hypothesized construct with a significance of
p<0.000. The one exception is item 20, which did not load with the other two items
intended to measure defensive pessimism. As a result, this item was omitted from
further analysis, and only items 18 and 19 were used to measure defensive
pessimism. This analysis suggests that there is convergent validity between the
items that measure each construct. Given the discriminant validity between
constructs and the convergent validity within each construct, the CFA indicates that
the remaining seventeen items are valid measures for this analysis.
61
TABLE 11: STUDY 3 CONFIRMATORY FACTOR ANALYSIS FACTOR LOADINGS
Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6
1. I expect that I would enjoy this movie very much. 0.756
2. This movie probably won’t put me in a positive mood.
(RS)
0.845
3. There is a very good chance that this movie will make
me feel good.
0.903
4. I forecast that after watching this movie, I would feel
very satisfied.
0.849
5. I forecast that after watching this movie, I would feel
very bad. (RS)
0.743
6. I forecast that after watching this movie, I would feel
very positive.
0.803
7. I am very confident about how this movie will make me
feel.
0.778
8. I am very uncertain about how this movie will affect my
mood. (RS)
0.719
9. I can accurately predict how this movie will affect my
emotions.
0.786
10. After watching National Treasure, I felt very satisfied. 1.030
11. After watching National Treasure, I felt very bad. (RS) 0.468
12. After watching National Treasure, I felt very positive. 0.664
62
TABLE 11: STUDY 3 CONFIRMATORY FACTOR ANALYSIS FACTOR LOADINGS, CONTINUED
Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6
13. Confederate Gold is very similar to National Treasure. 1.200
14. The plot or storyline in Confederate Gold is not very
similar to the plot of National Treasure. (RS)
0.612
15. Confederate Gold would make me feel the same
emotions as National Treasure.
0.350
16. I go into a movie expecting the worst, even though I
know the movie will probably be OK
2.199
17. I try to keep my expectations for a movie fairly low, so
that I will be pleasantly surprised with how much I enjoy
the movie.
0.238
18. I usually set very high expectations for sequels of movies
that I really enjoyed. (RS)
0.037
Note: Factor 1: Affective Expectations
Factor 2: Affective Forecast
Factor 3: Forecast Confidence
Factor 4: Parent Brand Affective Experience
Factor 5: Parent-Extension Similarity
Factor 6: Defensive Pessimism
Factor loadings indicate standardized regression weights
All loadings are significant at p<0.05, except Q20
RS indicates reverse-scored items
63
Defensive Pessimism: To create high and low groups for defensive
pessimism, the two items used to measure defensive pessimism were averaged to
create a measure of defensive pessimism. Results for this measure ranged from -3.00
to 3.00, with a mean of -1.26, a median of -1.50, and a standard deviation of 1.34. A
median split was performed on the result, with subjects above the median being
assigned to the “high defensive pessimism” group, and subjects below the median
being assigned to the “low defensive pessimism” group.
Hypothesis Testing: Hypothesis 5a suggests that subjects in the high
defensive pessimism condition will generate more negative thoughts when
evaluating a brand extension than subjects in the low defensive pessimism group.
Table 12 displays the results of the ANOVA analysis conducted to test this
hypothesis. The results indicate that subjects in the high defensive pessimism group
generated more negative thoughts (x = 2.72, s = 1.72) about Confederate Gold when
they thought it was a sequel than subjects in the low defensive pessimism group (x =
1.93, s = 1.60). This difference approached significance (F = 3.041, p<0.087). Thus,
Hypothesis 5a is supported, though not significantly.
64
TABLE 12: STUDY 3 HYPOTHESIS H5 TEST RESULTS
New Brand
(Non-sequel)
Brand Extension
(Sequel)
Measure
Low
Defensive
Pessimism
High
Defensive
Pessimism
Low
Defensive
Pessimism
High
Defensive
Pessimism
Negative Thoughts 2.12
ab
(1.70)
1.65
a
(1.46)
1.93
a
(1.60)
2.72
b
(1.72)
Positive Thoughts 1.38
a
(1.41)
1.30
a
(1.46)
1.69
a
(1.34)
1.40
a
(1.41)
Total Thoughts 3.97
a
(1.99)
3.70
a
(1.46)
4.07
a
(1.71)
4.48
a
(2.00)
Affective Forecast -0.156
a
(1.35)
0.075
ab
(1.18)
0.517
b
(1.11)
-0.220
a
(1.25)
Note: Results are shown as mean (standard deviation)
Means in a given row with the same superscript are not significantly different
at p<0.10.
Hypothesis 5b suggests that subjects in the high defensive pessimism
condition will generate more negative thoughts when evaluating a brand extension
than they will when evaluating a new brand. Table 12 displays the results of the
ANOVA analysis conducted to test this hypothesis. The results indicate that subjects
in the high defensive pessimism group generated more negative thoughts (x = 2.72, s
= 1.72) about Confederate Gold when they thought it was a sequel than when they
thought it was an original movie (x = 1.65, s = 1.46). This difference is significant
(F = 4.902, p<0.032). Thus, Hypothesis 5b is supported.
Hypotheses 5a and 5b suggested implications only for negative thoughts.
However, the mean number of positive thoughts and total thoughts are also reported
in Table 12, for comparison. There were no significant differences between cells for
either positive or total thoughts. The sum of negative plus positive thoughts does not
65
equal the number of total thoughts reported because some thoughts were neutrally-
valenced.
Hypothesis 6 suggests that generating a greater number of negative thoughts
causes subjects to create less positive affective forecasts. Table 13 displays the
results of the regression analysis conducted to test this hypothesis. Model 1 is a
regression of affective forecast on negative thoughts only, and Model 2 includes the
interaction of defensive pessimism and brand extension as a second predictor
variable. In both models, negative thoughts has a significant negative impact on
affective forecasts (Model 1: = -0.391, p<0.000, Model 2: = -0.393, p<0.000).
Thus, Hypothesis 6 is supported. Model 2 also indicates that subjects in the high
defensive pessimism group who evaluated the movie positioned as a brand extension
generated significantly lower affective forecasts than other subjects ( = -0.182,
p<0.043). This result resonates with the data in Table 12, which indicates that
affective forecasts for the sequel were significantly lower for high (x = -0.220, s =
1.25) versus low defensive pessimists (x = 0.517, s = 1.11), and that this difference is
significant (F = 5.286, p<0.026).
TABLE 13: STUDY 3 HYPOTHESIS H6 TEST RESULTS
Model 1 Model 2
Parameter Negative Thoughts Negative Thoughts
Defensive Pessimism
x Brand Extension
b -0.295 -0.289 -0.225
-0.391 -0.393 -0.182
t-statistic -4.339 -4.305 -2.046
p-value 0.000 0.000 0.043
Note: Dependent variable is Affective Forecast
66
Hypothesis 7 suggests that subjects who are more defensively pessimistic
will place greater weight on forecast confidence when creating affective expectations
than will subjects who are less defensively pessimistic. This hypothesis was tested
by comparing the moderating influence of forecast confidence on the relationship
between affective forecasts and affective expectations for subjects who were in the
high versus low defensive pessimism groups. Specifically, a structural equation
model was created, with affective expectations being influenced by the main effect
of affective forecasts and the moderating effect of forecast confidence, as depicted in
Figure 1. Then, the fit of this structural equation model and the regression
coefficient for the moderating effect of forecast confidence were compared between
the high and low defensive pessimism groups. Table 14 displays the results of the
model fit and regression coefficient for each group. Results indicate that the model
has a better fit for the high defensive pessimism group (AIC = 25.973) than for the
low defensive pessimism group (AIC = 37.855). Furthermore, forecast confidence
has a stronger moderating influence in the high defensive pessimism group ( =
0.310, p<0.003) than in the low defensive pessimism group ( = -0.098, p<0.143),
and the difference between these regression weights is significant (F = 4.880,
p<0.029). Thus, Hypothesis 7 is supported.
67
TABLE 14: STUDY 3 HYPOTHESIS H7 TEST RESULTS
Model Fit Parameter Low Defensive
Pessimism
High Defensive
Pessimism
Chi-square 13.855, df=2, p<0.001 1.973, df=2, p<0.373
RMSEA 0.314 0.000
AIC (saturated) 37.855 (28.000) 25.973 (28.000)
Forecast Confidence
Weight
= -0.098, p<0.143
a
= 0.310, p<0.003
b
Affective Forecast = 0.907, p<0.000
a
= 0.654, p<0.000
a
Note: Betas in a given row with the same superscript are not significantly
different at p<0.05.
Hypothesis 8 suggests that subjects who are more defensively pessimistic
will place greater weight on parent-extension similarity when creating their affective
forecasts for brand extensions than will subjects who are less defensively
pessimistic. This hypothesis was tested by comparing the moderating influence of
similarity on the relationship between parent brand affective experience and brand
extension affective forecasts for subjects who were in the high versus low defensive
pessimism groups. Specifically, a structural equation model was created, with
affective forecasts being influenced by the main effect of parent brand affective
experience and the moderating effect of similarity, as depicted in Figure 2. Then, the
fit of this structural equation model and the regression coefficient for the moderating
effect of similarity were compared between the high and low defensive pessimism
groups. Table 15 displays the results of the model fit and regression coefficient for
each group. Results indicate that the model has a better fit for the high defensive
pessimism group (AIC = 31.074) than for the low defensive pessimism group (AIC =
32.901). Furthermore, similarity has a stronger moderating influence in the high
68
defensive pessimism group ( = 0.476, p<0.008) than in the low defensive
pessimism group ( = 0.236, p<0.212), and the difference between these regression
weights is moderately significant (F = 3.774, p<0.057). Thus, Hypothesis 8 is
moderately supported.
TABLE 15: STUDY 3 HYPOTHESIS H8 TEST RESULTS
Model Fit Parameter Low Defensive
Pessimism
High Defensive
Pessimism
Chi-square 16.901, df=1, p<0.000 15.074, df=1, p<0.000
RMSEA 0.515 0.566
AIC (saturated) 32.901 (18.000) 31.074 (18.000)
Similarity Weight = 0.236, p<0.212
a
= 0.476, p<0.008
b
Parent Brand
Affective Experience
Weight
= 0.518, p<0.009
a
= 0.414, p<0.019
a
Note: Betas in a given row with the same superscript are not significantly different at
p<0.10.
Discussion
The results of Study 3 support the theoretical model proposed in Figure 3.
Consumers who are defensively pessimistic generate more negative thoughts and
less positive affective forecasts when they are evaluating a brand extension than
when they are evaluating a new brand. Furthermore, defensively pessimistic
consumers generate more negative thoughts and less positive affective forecasts for a
brand extension than do consumers who are not defensively pessimistic. Thus, when
evaluating the fictional movie Confederate Gold, a defensively pessimistic consumer
will generate more negative thoughts and lower affective forecasts for the movie if
they think it is a sequel to National Treasure than if they think it is an original
69
movie, and also more negative thoughts and lower affective forecasts than
consumers who are not defensively pessimistic.
One other interesting finding in this study is that subjects in the low defensive
pessimism group had very high affective forecasts for Confederate Gold when they
thought it was a sequel. So, for those consumers who wonder why Hollywood
churns out so many disappointing sequels, the answer may be that consumers don’t
know they are bad movies until after they pay for the ticket; at the time of purchase,
however, the non-defensively pessimistic consumer actually has a rather positive
affective forecast for the sequel.
Another interesting finding is that although defensive pessimists generated
lower affective forecasts for the sequel than did other subjects, there were no
significant differences in behavioral intentions (item 1) between any of the groups.
Therefore, although defensive pessimists think more negatively about sequels and
have lower affective forecasts for them, they do intend to see them with the same
likelihood as non-defensive pessimists. Thus, the lower affective forecasts observed
in this study are not simply the result of subjects who have no interest in seeing the
movie, but are, in fact, the “unrealistically low expectations” (Norem and Cantor
1986) that defensive pessimists set to defend themselves in a risky situation.
One limitation to this study is that defensive pessimism was not
experimentally manipulated. Instead, high and low defensive pessimism groups
were created by classifying the subjects according to their responses to the defensive
pessimism measures. Although this does represent a relative difference between the
two groups, it does not necessarily mean that subjects in the high defensive
70
pessimism group are, in fact, defensively pessimistic with respect to their affective
forecasts for movies, or that subjects in the low defensive pessimism group are no
defensively pessimistic. Instead of being absolute classifications, these are simply
relative differences between the subjects who participated in this study. However,
since the ANOVA analyses were conducted to compare differences between the
groups, this limitation does not appear to impact the findings of this study.
A second limitation to this study is that subjects were presented with a text
description of a fictional movie. Although this method was chosen as a more
realistic alternative to the more artificial manipulations of the first two studies, it is
still somewhat unrealistic. Marketing communications for motion pictures are often
audio-visual in nature, are provide consumers with a multimedia sample of scenes
from the movie. Reading a text description of a movie does not give the same
experience as watching a movie preview, and perhaps makes the method slightly less
realistic. However, as all subjects were presented with a text description of
Confederate Gold, differences between groups can still be validly analyzed.
A third limitation to this study is the potential confound between brand
strategy and information availability. Subjects in the sequel condition were
presented with information about the parent brand movie, which increases the
amount of information available for creating an affective forecast for the sequel.
However, subjects in the new movie condition did not have this information
available for evaluation. Because there was no experimental control for information
availability, it is not possible to eliminate that construct as an alternative explanation
for the effects found in the study. However, the fact that subjects in the new movie
71
condition were not explicitly reminded of the movie National Treasure does not
mean that they did not have this information available for creating their affective
forecasts of Confederate Gold. It merely means that these subjects were not
reminded of this information. Therefore, while the potential confound cannot be
ruled out, this does not mean that the brand strategy effect does not exist – merely
that it is not the only possible explanation for the findings.
Another important limitation to this study is that the defensive pessimism
measures were too closely related to the dependent outcome rather than the causal
psychological state. The defensive pessimism measures included in this study were
adapted from other research on the construct (Sanna 1996), in order to preserve
continuity with existing research. However, the wording of the items focuses on
expectations, rather than the anxiety and the desire for self-protection that impacts
expectations. As a result, the relationships between defensive pessimism and
subjects’ affective forecasts and expectations may be artifacts of the research
methodology, rather than ecologically valid findings.
Similarly, the lack of a distractor task between the dependent and
independent measures may have impacted the thought statements that subjects
reported. Subjects were first asked about their expectations, and then asked to report
their thoughts about the movie. As a result, in order to appear consistent, subjects
may have selectively reported thoughts with the same valence as their previously
reported expectations. If so, then the finding that subjects who reported more
negative thoughts also had more negative affective forecasts may be an artifact of the
research methodology, rather than an ecologically valid finding.
72
These findings are intended to advance the literature on defensive pessimism
by proposing and testing a model of the effects of defensive pessimism on affective
forecasts. The findings also suggest that defensive pessimism has more wide-
ranging impacts on affective forecasts and affective expectations, by influencing the
strength of the moderating influences of forecast confidence and similarity in the
affective forecast and affective expectations models, respectively. Defensive
pessimism is a construct that has received some little attention in the affective
forecasting literature, but has not received detailed exploration. The findings
presented here suggest that this construct is worthy of much more exploration to
identify its other potential impacts.
These findings are also intended to advance consumer behavior literature by
incorporating brand extensions with defensive pessimism. Although many
marketing studies have examined brand extension issues, only a handful of studies in
psychology and education have explored defensive pessimism. The findings of this
third study suggest that academic research may be able to shed light on defensive
pessimism, and that marketers can benefit from examining the impact of defensive
pessimism on other consumer behavior issues. Rather than simply studying
defensive pessimism in a consumption context instead of an education context, this
study has actually interacted defensive pessimism with brand extensions to
determine how the two combine to impact consumers’ affective forecasts.
Hopefully, this will motivate further exploration of how defensive pessimism
impacts consumer behavior, both within and beyond the entertainment context.
73
DISCUSSION
Contributions
This dissertation proposes to make several contributions to marketing theory.
One is to examine how consumers create affective expectations for entertainment
experiences. Relatively little academic research in marketing has studied the
entertainment industry, and much of that work focuses on quantitative modeling of
motion picture box office revenues using market-based data (e.g., Eliashberg and
Shugan 1997, Neelamegham and Jain 1999, Sawhney and Eliashberg 1996). While
these methods can be very useful for observing certain behaviors, they do not explain
the psychological influences and processes that cause individual-level consumer
behavior. This dissertation tries to address this gap by examining these
psychological influences. It also helps to explain the effects that may be found in
those econometric studies. For instance, one study of box office revenues found that
movie sequels earn more revenue than original movies (Sawhney and Eliashberg
1996), yet the authors provide no psychological rationale to explain why consumers
would prefer to watch a sequel than an original movie. This dissertation suggests an
explanation for this effect: consumers are more confident in their affective forecasts
for sequels, thereby raising their affective expectations and choice intentions for
sequels, compared to original movies. Thus, by suggesting a psychological
explanation for consumers’ entertainment choices, this dissertation suggests that the
concept of affective expectations and its antecedents can provide a foundation for
studying the psychological influences that drive consumers’ choices for
entertainment experiences.
74
A second contribution is identifying the importance of forecast confidence on
consumers’ affective expectations. As Study 1 suggests, affective expectations for
hedonic experiences are significantly influenced by consumers’ confidence in their
affective forecasts of how those experiences might make them feel. Specifically,
forecast confidence acts as a weighting mechanism that gives rise to the probabilistic
affective expectations. Much of the previous research on affective forecasting has
examined only the forecast, and ignored the forecast confidence. This study suggests
that this omission yields an incomplete understanding of affective expectations.
Another contribution of this dissertation is to identify the impact of brand
extension similarity on consumers’ affective expectations. Prior work on brand
extensions has focused on consumers’ evaluations of a brand extension, not their
expectations about how a brand extension will make them feel. The study also
contributes to the literature by examining the role of similarity on forecast
confidence. Not only does the similarity of a brand extension to its parent brand
moderate the transfer of associations from parent to extension, this similarity also
directly impacts consumers’ forecast confidence for these extensions. Thus, in
addition to influencing consumers’ development of their affective forecasts, brand
extension similarity also enhances consumers’ forecast confidence, strengthening
affective expectations in the direction of the affective forecast. This identifies a new
outcome of brand extensions, which had not been previously explored in the
literature.
Another theoretical contribution of this dissertation is the finding that
defensive pessimism impacts affective forecasts and expectations. Defensive
75
pessimism alters how consumers think about new brands and brand extensions,
thereby impacting their affective forecasts. Interestingly, defensive pessimism also
moderates the influence of forecast confidence and similarity in the affective
expectations and forecast models proposed in the previous studies. These findings
indicate that defensive pessimism has powerful impacts on consumers’ affective
forecasts and expectations, and suggests that this construct may have much to offer
in contributing to scholarly knowledge of consumer behavior.
This dissertation also provides implications for marketing practice. One such
implication is that firms should try to increase consumers’ confidence in their
affective forecasts for the firm’s brands. By doing so, firms can expect to increase
the intensity of consumers’ affective expectations, which will improve them,
assuming that consumers have a positive affective forecast. One example of how
firms can try to increase consumers’ forecast confidence is by providing samples of
the actual experience, such as making clips from a movie or passages from a book
available to customers prior to purchase. Thus, instead of solely trying to convince
consumers to create a positive affective forecast, firms can also try to increase
consumers’ forecast confidence.
A second managerial implication is to exploit brands that provide positive
experiences for consumers, by introducing brand extensions that are similar to these
parent brands. This strategy will generate positive affective forecasts through the
transfer of positive associations from the parent brand to the brand extension, which
will be facilitated by the similarity of the two goods. Furthermore, the high
similarity will increase consumers’ forecast confidence. The combination of a
76
positive affective forecast with strong forecast confidence will yield very positive
affective expectations, which can be expected to increase consumers’ purchase
intentions for the firm’s entertainment experience.
On a related note, this research also indicates that a disappointing parent
brand can be salvaged by introducing a dissimilar brand extension. For example, an
author whose first detective novel fails to satisfy consumers might try writing a
romance novel next. By changing the genre, character, setting, and other story
elements, the new extension of the author’s brand name (i.e., the new novel) will be
perceived as very different from the author’s previous book. It will give the author a
fresh start in the minds of consumers by inhibiting the transfer of negative
associations from the first book to the second. In this way, brands can be reinvented
through the use of dissimilar extensions.
Limitations
One limitation of this dissertation is that the empirical studies are entirely
laboratory-based. While the studies used the different empirical contexts of motion
pictures and literature, they were somewhat artificial due to the setting. Especially in
Studies 1 and 2, the entertainment experiences described in the manipulations were
hypothetical, and may not provide as good a test of real choice decisions as
descriptions of actual entertainment experiences. Although Study 3 included a more
realistic methodology, it still included a fictional movie. However, the choice to use
hypothetical experiences was made to avoid contamination with any real-world
information that subjects may have come across prior to the studies. Now that these
77
effects have been demonstrated experimentally, survey research might examine these
effects in larger, more ecological, consumer samples.
Another limitation of this dissertation is that the empirical work explores only
movies and books. Other entertainment categories, including television, plays, video
games, music, board games, etc., are omitted. The theoretical arguments proposed in
this dissertation should apply to these other entertainment categories as well, but the
empirical studies in this dissertation do not examine these categories. Unlike many
marketing studies that explore the entertainment industry, this dissertation used
empirical measures that can be easily adapted to any entertainment category.
However, the applicability of these measures to categories other than movies and
books is left for future research.
Another limitation of this dissertation is that the data analyses may be
influenced by non-normality or heteroscedasticity. Variables were examined
visually, using Q-Q plots, and judged to be acceptable. However, this technique
provides only a qualitative assessment of the data, and does not preclude the analyses
from being biased by non-normality or heteroscedasticity. Using smoothers and
other advanced techniques may improve the statistical power of these analyses. This
may also improve statistical power, which could improve the significance of
findings, and make marginally significant findings (0.05 < p < 0.10) become
significant (p < 0.05).
Future Directions
This dissertation inspires several possible ideas for future exploration. One
option is to explore the generalizability of this affective expectations framework to
78
other entertainment categories. Much of the entertainment research in marketing
focuses on movies, and uses very movie-specific variables, such as MPAA rating or
star power. This dissertation has expanded beyond movie-specific variables by
examining psychological constructs, such as forecast confidence and affective
expectations. However, most of the empirical studies still focus on the movie
category. While it is proposed that these psychological processes apply to other
entertainment categories, literature was the only other entertainment category tested.
Future research should explore the applicability of these findings to other
entertainment categories, such as television, plays, video games, music, board
games, and others.
Another opportunity for future research is to explore the applicability of these
findings to categories outside the entertainment realm. While entertainment
provided the inspiration for this dissertation, the conceptual models developed herein
should apply to any category, to the extent that it can be characterized as an
experience category that provides some degree of affective benefits. Entertainment
experiences are an excellent example of these goods, because they are experience
goods that provide primarily affective benefits. However, other categories may
exhibit these characteristics, as well. For example, although it may be primarily
consumed for its nutritional value, soup has also been found to provide affective
benefits, as well. Furthermore, like any food, soup is an experience good, so
consumers cannot determine the quality of the soup until they eat it. Therefore, soup
consumers may forecast the affective experience that each brand of soup will
provide, and determine how confident they are in that forecast, which will influence
79
their affective expectations, and ultimately, their choice. It is possible that these
affective expectations will have less impact on soup choice than entertainment
choice, because there are other important attributes that impact soup choice, such as
nutritional content. However, this dissertation proposes that the same basic
processes would impact affective expectations for soup as for entertainment. Thus,
the findings of this dissertation are proposed to have very broad applicability, the
extent of which should be explored in future research.
Another interesting idea is to examine other consumer characteristics that
would impact the effects of forecast confidence on affective expectations, or
similarity on brand extension affective forecasts. For instance, one such effect might
be the consumer’s desire to avoid feeling regret. Anticipated regret has been shown
to significantly influence consumers’ choice intentions (Simonson 1992, Zeelenberg
et al 1996). It is possible that when a consumer desires to avoid regret, they may
place greater emphasis on their forecast confidence, thereby choosing entertainment
options that are extremely similar to their past experiences. In contrast, when
consumers are very willing to accept the possibility of feeling regret, then they may
place less emphasis on their forecast confidence, and exhibit greater variety-seeking
in their entertainment choices. This regret-avoidance characteristic may be an
enduring consumer trait, or a temporary state. Future research may explore the
impact of this and other consumer characteristics on affective expectations.
Another area for future research is to examine the influence of consumers’
forecast confidence on post-consumption satisfaction evaluations. Several studies
have explored the concept of affective misforecasting (i.e., creating inaccurate
80
affective forecasts), and find that consumers tend to misforecast their future affect in
terms of valence (Gilbert et al 2004, Woodzicka and LaFrance 2001), intensity
(Buehler and McFarland 2001, Mitchell et al 1996), and duration (Gilbert et al 1998,
Wilson et al 2000). Furthermore, this misforecasting has been shown to have
implications for post-consumption satisfaction evaluations (MacInnis et al 2005,
Patrick et al 2007). The concepts of forecast confidence and affective expectations
introduced in this dissertation may be explored in that research area, as well, to
examine their impact on satisfaction judgments. In this dissertation, the empirical
work focused on decision-making. In reality, consumers would then make a choice
and undergo a consumption experience. After consumption, consumers would be
able to evaluate the experience to determine how well their actual affective
experience matched their affective expectations. It would be interesting to examine
the influence of forecast confidence on these satisfaction judgments. For instance,
perhaps greater forecast confidence would lead to more intense disconfirmation
reactions. That is, a consumer whose affective experience was more positive than
forecast might be much more satisfied if they were very confident of the lesser
affective forecast than if they were uncertain about what to expect. That is, they
might be much more pleasantly surprised by a superb experience if they were
extremely confident of a mediocre experience than if they were uncertain about what
to expect. Exploring the effects of forecast confidence on satisfaction judgments
could make an important contribution to both the affective forecasting and
satisfaction domains.
81
Another future research issue would be to explore the relative effects of
affective expectations and attitude on consumer choice. In study 1, affective
expectations were demonstrated to be discriminant from attitude, but the two
constructs have several similarities. Attitudes are formed by the interaction of
attribute beliefs and evaluations (Fishbein and Ajzen 1972, Wilkie and Pessemier
1973). Affective expectations are formed by the interaction of affective forecasts
and forecast confidence (per study 1). Furthermore, attitude strength or certainty can
also impact the influence of attitudes on choice, just as forecast confidence can
impact the influence of affective forecasts or affective expectations. However, aside
from demonstrating the differences between these related constructs, this dissertation
has not examined their relative importance in consumer choice. It would be
interesting and valuable to explore which construct has greater implications for
consumer choice, and in which situations one dominates the other. For instance, it
may be that attitude has stronger effects on choice than affective expectations in
utilitarian categories. Alternatively, it may be that affective expectations are better
predictors of choice than attitude in hedonic categories. Understanding whether the
cognitive or affective construct is more influential in any particular consumption
category would benefit marketers seeking to influence consumers’ choices.
Another future research area would be to examine the effects of brand
extensions of family brands. Many entertainment (and other) brand extensions are
derived from family brands, which include multiple exemplars. For instance, when
Rocky Balboa was released in 2006, it had five previous exemplars in the Rocky
brand family. In these situations, it would interesting to explore how consumers
82
identify a referent against which to compare the new extension, especially when
there is considerable variance between the existing brand exemplars. For instance,
did consumers form their affective forecasts for Rocky Balboa by transferring their
recalled associations from the original Rocky, Rocky V, some average of the first five
movies, the exemplar deemed to be most similar to the new extension, or some other
referent? Given the prevalence of brand extensions as a branding strategy and the
ensuing growth in the size of brand families, this is an important question that
remains to be addressed in the brand extensions literature.
On a related note, another area for future exploration is to examine how
consumers form affective forecasts for new brands. In this dissertation, it has been
suggested that consumers transfer affect and other associations from a parent brand
to a brand extension. However, consumers are also able to form affective forecasts
for new brands. Do they do so completely from scratch, or do they draw upon some
referent experience(s) from which they transfer associations, even if the referent is
not a parent brand of the new good? If the latter, how do consumers identify a
referent – do they use similar, previously-consumed experiences as a basis for
comparison, and if so, how do they judge this similarity and identify valid referents?
It may be the case that consumers form affective forecasts by comparing the focal
experience to previously-consumed experiences, based on its similarity to those
referents. Perhaps brand extension effects are so commonly observed because
consumers readily identify the similarity between a brand extension and its parent
brand. However, perhaps differently-branded referents can have the same effect,
when the consumer is prompted to consider them as referents. One branding strategy
83
is to draw reference to the creator(s) of a good to establish them as a referent. For
instance, promotions for 2006’s V for Vendetta stated that the movie was from the
creators of The Matrix. This may establish The Matrix as a referent, and encourage
consumers who enjoyed that movie to form positive affective forecasts for the new
movie. Even without such active branding, consumers may use ingredient brands as
a clue to identifying a referent. For instance, a consumer trying to identify a referent
experience for the 2006 movie Blood Diamond, starring Leonardo DiCaprio, might
recall their experiences with other movies starring that actor. Future research might
explore the processes and heuristics that consumers use to identify referent
experiences when creating affective forecasts for new goods.
A final suggestion for future research is to examine the effects of cross-
category entertainment extensions. In this dissertation, the brand extensions studied
were in the same category as their parent brands (e.g., a new book by an author, a
sequel of an original movie). It would be interesting to explore the effects of cross-
category entertainment brand extensions, such as movies extended from books or
video games based on movies. In addition, it would be interesting to explore the
moderating effects of the level of involvement required of each medium. For
instance, one study found that reading a book requires more cognitive involvement
than watching a movie (Salomon 1984). Assuming that greater cognitive
involvement requirements reduce the number of interested consumers, then it may be
hypothesized that extending to a less-involving medium (e.g., a movie) may generate
greater audience appeal for the extension, while extending to a more-involving
medium (e.g., a book) is more appealing to consumers who are highly involved with
84
the brand. Thus, the direction of extension into more- or less-involving media may
have implications for increasing the breadth of the brand’s customer base versus
increasing the depth of the brand experience for fewer consumers.
85
REFERENCES
Aaker, David A. and Kevin Lane Keller (1990), “Consumer Evaluations of Brand
Extensions,” Journal of Marketing, v54 (January), 27-41
Anderson, Justin (2003), “Return of the Brand Extensions: A Strategic Rationale for
Movie Sequels and Other Brand Extensions,” Unpublished Manuscript
Bagozzi, Richard P., Mahesh Gopinath, and Prashanth U. Nyer (1999), “The Role of
Emotions in Marketing,” Journal of the Academy of Marketing Science, 27
(2), 184-206
Baron, Reuben M. and David A. Kenny (1986), “The Moderator-Mediator Variable
Distinction in Social Psychological Research: Conceptual, Strategic, and
Statistical Considerations,” Journal of Personality and Social Psychology, 51
(6), 1173-1182
Bell, David E. (1982), “Regret in Decision Making Under Uncertainty,” Operations
Research, 30 (5), 961-981
Berger, Ida E. and Andrew A. Mitchell (1989), “The Effect of Advertising on
Attitude Accessibility, Attitude Confidence, and the Attitude-Behavior
Relationship,” Journal of Consumer Research, 16 (3), 269-279
Bhat, Subodh and Srinivas K. Reddy (2001), “The Impact of Parent Brand Attribute
Associations and Affect on Brand Extension Evaluation,” Journal of
Business Research, 53, 111-122
Bosshart, Louis and Ilaria Macconi (1998), “Media Entertainment,” Communication
Research Trends, v18 n3, 3-6

Bottomley, Paul A. and John R. Doyle (1996), “The Formation of Attitudes Towards
Brand Extensions: Testing and Generalising Aaker & Keller’s Model,”
International Journal of Research in Marketing, 13, 365-377
Bottomley, Paul A. and Stephen J.S. Holden (2001), “Do We Really Know How
Consumers Evaluate Brand Extensions? Empirical Generalizations Based on
Secondary Analysis of Eight Studies,” Journal of Marketing Research, 38
(November), 494-500
Boush, David and Barbara Loken (1991), (A Process-Tracing Study of Brand
Extensions Evaluation,” Journal of Marketing Research, 28 (February), 16-
28
86
Boush, David, Shannon Shipp, Barbara Loken, Esra Gencturk, Susan Crockett, Ellen
Kennedy, Bettie Minshall, Dennis Misurell, Linda Rochford, and Jon Strobel
(1987), “Affect Generalizations to Similar and Dissimilar Brand Extensions,”
Psychology and Marketing, v4 n3, 225-237
Broniarczyk, Susan M. and Joseph W. Alba (1994), “The Importance of the Brand in
Brand Extension,” Journal of Marketing Research, 31 (May), 214-228
Buehler, Roger and Cathy McFarland (2001), “Intensity Bias in Affective
Forecasting” The Role of Temporal Focus,” Personality and Social
Psychology Bulletin, 27 (11), 1480-1493
CIA (2005), “The World Factbook,”
http://www.cia.gov/cia/publications/factbook/geos/us.html (accessed June 26,
2006)
Census Bureau (2002), “2002 Economic Census,”
http://www.census.gov/econ/census02 (accessed June 26, 2006)
Cobb, Charles W. and Paul H. Douglas (1928), “A Theory of Production,” American
Economic Review, 18 (1), Supplement (March), 139-172
Czellar, Sandor (2003), “Consumer Attitude Toward Brand Extensions: An
Integrative Model and Research Propositions,” International Journal of
Research in Marketing, 20, 97-115
DelVecchio, Devon (2000), “Moving Beyond Fit: The Role of Brand Portfolio
Characteristics in Consumer Evaluations of Brand Reliability,” Journal of
Product and Brand Management, 9 (7), 457-471
Eliashberg, Jehoshua and Steven M. Shugan (1997), “Film Critics: Influencers or
Predictors,” Journal of Marketing, v61 (April), 68-78
Fazio, Russell H. and Mark P. Zanna (1978), “Attitudinal Qualities Relating to the
Strength of the Attitude-Behavior Relationship,” Journal of Experimental
Social Psychology, 14 (4), 398-408
Fishbein, Martin and Icek Ajzen (1972), “Attitudes and Opinions,” Annual Review of
Psychology, 23, 487-544
Gilbert, Daniel, Elizabeth C. Pinel, Timothy D. Wilson, Stephen J. Blumberg, and
Thalia P. Wheatley (1998), “Immune Neglect: A Source of Durability Bias in
Affective Forecasting,” Journal of Personality and Social Psychology, 75 (3),
617-638
87
Gilbert, Daniel T., Carey K. Morewedge, Jane L. Risen, and Timothy D. Watson
(2004), “Looking Forward to Looking Backward: The Misprediction of
Regret,” Psychological Science, 15 (5), 346
Green, Melanie C., Timothy C. Brock, and Geoff F. Kaufman (2004),
“Understanding Media Enjoyment: The Role of Transportation into Narrative
Worlds,” Communication Theory, v14 n4 (November), 311-327
Gurhan-Canli, Zeynep (2003), “The Effect of Expected Variability of Product
Quality and Attribute Uniqueness on Family Brand Evaluations,” Journal of
Consumer Research, v30 (June), 105-114
Gurhan-Canli, Zeynep and Durairaj Maheswaran (1998), “The Effects of Extensions
on Brand Name Dilution and Enhancement,” Journal of Marketing Research,
v35 n4 (November), 464-473
Hansen, Christine H. and Ranald D. Hansen (2000), “Music and Music Videos,” In:
Zillmann, Dolf and Peter Vorderer (Eds.), Media Entertainment: The
Psychology of Its Appeal. Lawrence Erlbaum Associates: Mahwah, NJ.
Havlena, William J. and Morris B. Holbrook (1986), “The Varieties of Consumption
Experience: Comparing Two Typologies of Emotion in Consumer Behavior,”
Journal of Consumer Research, 13 (December), 394-404
Hirschman, Elizabeth C. and Morris B. Holbrook (1982), “Hedonic Consumption:
Emerging Concepts, Methods, and Propositions,” Journal of Marketing, 48
(Summer), 92-101
Holbrook, Morris B. (1999), “Popular Appeal Versus Expert Judgments of Motion
Pictures,” Journal of Consumer Research, 26 (September), 144-155
Holbrook, Morris B., Robert W. Chestnut, Terence A. Oliva, and Eric A. Greenleaf
(1984), “Play as a Consumption Experience: The Roles of Emotions,
Performance, and Personality in the Enjoyment of Games,” Journal of
Consumer Research, v11 n2, 728-739
Holbrook, Morris B. and Elizabeth C. Hirschman (1982), “The Experiential Aspects
of Consumption: Consumer Fantasies, Feelings, and Fun,” Journal of
Consumer Research, 9 (September), 132-140
Holbrook, Morris B. and Donald R. Lehmann (1981), “Allocating Discretionary
Time: Complementarity Among Activities,” Journal of Consumer Research,
7 (4), 395-406
88
John, Deborah Roedder, Barbara Loken, and Christopher Joiner (1998), “The
Negative Impact of Extensions: Can Flagship Brands Be Diluted?” Journal of
Marketing, v62 (January), 19-32
Kammen, Robert H. (1989), “An Analysis of Causal Attributions and State Anxiety
of Optimists and Defensive Pessimists,” PhD Dissertation, Indiana University
of Pennsylvania
Krider, Robert E. and Charles B. Weinberg (1998), “Competitive Dynamics and the
Introduction of New Products: The Motion Picture Timing Game,” Journal of
Marketing Research, v35 n1, 1-15
Lakoff, George (1987), “From Wittgenstein to Rosch,” In: Lakoff, George, Women,
Fire, and Dangerous Things. Chicago: University of Chicago Press, Chapter 2
Loken, Barbara and Deborah Roedder John (1993), “Diluting Brand Beliefs: When
Do Brand Extensions Have a Negative Impact?” Journal of Marketing, v57
(July), 71-84
Lopes, L L. (1984), “Risk and Distributional Inequality,” Journal of Experimental
Psychology: Human Perception and Performance, 10 (4), 465-485
Lopes, L. L. (1987), “Between Hope and Fear: The Psychology Risk,” Advances in
Experimental Social Psychology, 20, 255-95
Lopes, Lola L. (1990), “Re-modeling Risk Aversion: A Comparison of Bernoullian
and Rank Dependent Value Approaches,” In: von Furstenberg (Ed.) Acting
Under Uncertainty: Multidisciplinary Conceptions. Boston: Klewer. 267-99
MPAA (2004), “2004 Movie Attendance Study,” Motion Picture Association of
America. http://www.mpaa.org, (accessed March 17, 2005)
MacInnis, Deborah J., Vanessa M. Patrick, and C. Whan Park (2005), “Looking
Through the Crystal Ball: Affective Forecasting and Misforecasting in
Consumer Behavior,” Review of Marketing Research, 2
Martin, Andrew J., Herbert W. Marsh, and Raymond L. Debus (2001), “Self-
Handicapping and Defensive Pessimism: Exploring a Model of Predictors
and Outcomes From a Self-Protection Perspective,” Journal of Educational
Psychology, 93 (1), 87-102
Martin, Andrew J., Herbert W. Marsh, Alan Williamson, and Raymond L. Debus
(2003), “Self-Handicapping, Defensive Pessimism, and Goal-Orientation: A
Qualitative Study of University Students,” Journal of Educational
Psychology, 95 (3), 617-628
89
Martin, Ingrid M. and David W. Stewart (2001), “The Differential Impact of Goal
Congruency on Attitudes, Intentions, and the Transfer of Brand Equity,”
Journal of Marketing Research, v39 (November), 471-484
Mellers, Barbara A., Alan Schwartz, K. Ho, and Ilana Ritov (1997), “Decision Affect
Theory: Emotional Reactions to the Outcomes of Risky Options,”
Psychological Science, 8, 423-429
Mellers, Barbara A., Alan Schwartz, K. Ho, and Ilana Ritov (1999), “Emotion-Based
Choice,” Journal of Experimental Psychology, 128 (3), 332-345
Milberg, Sandra J., C. Whan Park, and Michael S. McCarthy (1997), “Managing
Negative Feedback Effects Associated With Brand Extensions: The Impact of
Alternative Branding Strategies,” Journal of Consumer Psychology, v6 n2,
119-140
Mitchell, Terence R., Leigh Thompson, Erika Peterson, and Randy Cronk (1997),
“Temporal Adjustments in the Evaluation of Events: The “Rosy View”,”
Journal of Experimental Social Science, 33, 421-448
Mundorf, Norbert and Kenneth R. Laird (2002), “Social and Psychological Effects of
Information Technologies and Other Interactive Media,” In: Bryant, Jennings
and Dolf Zillmann (Eds.), Media Effects: Advances in Theory and Research,
2
nd
Edition. Lawrence Erlbaum Associates: Mahwah, NJ
Neelamegham, Ramya and Dipak Jain (1999), “Consumer Choice Process for
Experience Goods: An Econometric Model and Analysis,” Journal of
Marketing Research, v36 (August), 373-386
Nelson, Phillip (1970), “Information and Consumer Behavior,” Journal of Political
Economy, 78 (March/April), 311-329
New York Times (2006), www.nytimes.com
Norem, Julie K. (2001), “Defensive Pessimism, Optimism, and Pessimism,” In:
Edward C. Chang (Ed.), Optimism and Pessimism: Implications for Theory,
Research, and Practice. American Psychological Association: Washington,
D.C., 77-100
Norem, Julie K. and Nancy Cantor (1986), “Defensive Pessimism: Harnessing
Anxiety as Motivation,” Journal of Personality and Social Psychology, 51
(6), 1208-1217
90
Oliver, Richard L. (1980), “A Cognitive Model of the Antecedents and
Consequences of Satisfaction Decisions,” Journal of Marketing Research, 17
(4), 460-469
Park, C. Whan, Sandra Milberg, and Robert Lawson (1991), “Evaluation of Brand
Extensions: The Role of Product Feature Similarity and Brand Concept
Consistency,” Journal of Consumer Research, v18 (September), 185-193
Patrick, Vanessa M., Deborah J. MacInnis, and C. Whan Park (2007), “Not as Happy
as I Thought I’d Be? Affective Misforecasting and Product Evaluations,”
Journal of Consumer Research, 33 (4), 479-489
Petty, Richard E. and John T. Cacioppo (1986), Communication and Persuasion.
Springer-Verlag: New York, NY
Phillips, Diane M. and Hans Baumgartner (2002), “The Role of Consumption
Emotions in the Satisfaction Response,” Journal of Consumer Psychology, 12
(3), 243-252
Richard, Rene, Joop van der Pligt, and Nanne de Vries (1996), “Anticipated Affect
and Behavioral Choice,” Basic and Applies Social Psychology, 18 (2), 111-
129
Sanna, Lawrence J. (1996), “Defensive Pessimism, Optimism, and Simulating
Alternatives: Some Ups and Downs of Prefactual and Counterfactual
Thinking,” Journal of Personality and Social Psychology, 71 (5), 1020-1036
Salomon, Gavriel (1984), “Television is ‘Easy’ and Print is ‘Tough’: The
Differential Investment of Mental Effort in Learning as a Function of
Perceptions and Attributions,” Journal of Educational Psychology, 76 (4),
647-658
Sawhney, Mohanbir S. and Jehoshua Eliashberg (1996), “A Parsimonious Model for
Forecasting Gross Box-Office Revenues of Motion Pictures,” Marketing
Science, v15 n2, 113-131
Simonson, Itamar (1992), “The Influence of Anticipating Regret and Responsibility
on Purchase Decisions,” Journal of Consumer Research, 19 (1), 105-118
Sujan, Mita, James R. Bettman, and Hans Baumgartner (1993), “Influencing
Consumer Judgments Using Autobiographical Memories: A Self-Referencing
Perspective,” Journal of Marketing Research, 30 (4), 422-436
Sullivan, Mary W. (1992), “Brand Extensions: When to Use Them,” Management
Science, v38 n6 (June), 793-805
91
Tannenbaum, Percy H. (1980), “Entertainment as Vicarious Emotional Experience,”
In: Tannenbaum, Percy H. (Ed.), The Entertainment Functions of Television.
Lawrence Erlbaum Associates: Hillsdale, NJ
Vorderer, Peter (2001), “It’s All Entertainment – Sure. But What Exactly Is
Entertainment? Communication Research, Media Psychology, and the
Explanation of Entertainment Experiences,” Poetics, 29, 247-261
Wernerfelt, Birger (1988), “Umbrella Branding as a Signal of New Product Quality:
An Example of Signalling By Posting a Bond,” RAND Journal of Economics,
v19 n3 (Autumn), 458-466
Wilkie, William L. and Edgar A. Pessemier (1973), “Issues in Marketing’s Use of
Multi-Attribute Attitude Models,” Journal of Marketing Research, X
(November), 428-441
Wilson, Timothy D., Thalia Wheatley, Jonathan M. Meyers, Daniel T. Gilbert, and
Danny Axsom (2000), Focalism: A Source of Durability Bias in Affective
Forecasting,” Journal of Personality and Social Psychology, 78 (5) 821-836
Woodzicka, Julie A. and Marianne LaFrance (2001), “Real Versus Imagined Gender
Harassment,” Journal of Social Issues, 57 (1), 15-30
Wright, Alice A. and John G. Lynch, Jr. (1995), “Communication Effects of
Advertising Versus Direct Experience When Both Search and Experience
Attributes Are Present,” Journal of Consumer Research, 21 (4), 708-718
Zeelenberg, M., J. Beattie, J. van der Plight and N. K. de Vries (1996),
“Consequences of Regret Aversion: Effects of Expected Feedback on Risky
Decision Making,” Organizational Behavior and Human Decision Processes,
65 (2), 148-158
Zillmann, Dolf (1988), “Mood Management: Using Entertainment to Full
Advantage,” In: Donohew, Lewis, Howard E. Sypher, and E. Tory Higgings
(Eds.) Communication, Social Cognition, and Affect. Lawrence Erlbaum
Associates: Hillsdale, NJ
Zillmann, Dolf (1994), “Mechanisms of Emotional Involvement in Drama,” Poetics,
v23, 33-51
Zillmann, Dolf (2005), “Empathy: Affective Reactions to Others’ Emotional
Experiences,” In: Bryant, Jennings and Peter Vorderer (Eds.), Psychology of
Entertainment. Mahwah, NJ: Erlbaum
92
APPENDIX A: STUDY 1 MANIPULATIONS
Condition Affective Forecast Forecast Confidence
1 Positive High
2 Positive Low
3 Neutral High
4 Neutral Low
Condition 1: Imagine that you have decided to see a comedy movie. Your local
movie theater is playing a comedy starring an actor who you think is extremely
funny. This actor’s other movies have made you laugh very much, and this movie
should be able to make you feel extremely happy. You have seen all of this actor’s
movies, and they always make you feel the same way. Also, you have read several
reviews of this movie, and all of the critics agree about how funny this movie is.
Therefore, you feel very confident about how happy this movie will make you feel.
Condition 2: Imagine that you have decided to see a comedy movie. Your local
movie theater is playing a comedy starring an actor who you think is extremely
funny. This actor’s other movies have made you laugh very much, and this movie
should be able to make you feel extremely happy. You have seen only a few of
this actor’s movies, and they never make you feel the same way. Also, you have
read several reviews of this movie, and none of the critics agree about how funny
this movie is. Therefore, you are not at all confident about how happy this movie
will make you feel.
Condition 3: Imagine that you have decided to see a comedy movie. Your local
movie theater is playing a comedy starring an actor who you think is somewhat
funny, but not extremely funny. This actor’s other movies have made you laugh, but
only a little bit, and this movie should be able to make you feel somewhat happy,
but not extremely happy. You have seen all of this actor’s movies, and they always
make you feel the same way. Also, you have read several reviews of this movie, and
all of the critics agree about how funny this movie is. Therefore, you feel very
confident about how happy this movie will make you feel.
Condition 4: Imagine that you have decided to see a comedy movie. Your local
movie theater is playing a comedy starring an actor who you think is somewhat
funny, but not extremely funny. This actor’s other movies have made you laugh, but
only a little bit, and this movie should be able to make you feel somewhat happy,
but not extremely happy. You have seen only a few of this actor’s movies, and
they never make you feel the same way. Also, you have read several reviews of this
movie, and none of the critics agree about how funny this movie is. Therefore, you
are not at all confident about how happy this movie will make you feel.
93
APPENDIX B: STUDY 1 MEASURES
Items 1-3 measure Affective Expectations on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
1. I expect that watching this movie will make me very happy.
2. Watching this movie probably won’t put me in a positive mood. (RS)
3. There is a very good chance that watching this movie will make me laugh a lot.
Items 4-6 measure Affective Forecast:
I forecast that after watching this movie, I would feel …
4. Very Sad 1 2 3 4 5 6 7 Very Happy
5. Very Negative 1 2 3 4 5 6 7 Very Positive
6. Very Bad 1 2 3 4 5 6 7 Very Good
Items 7-9 measure Forecast Confidence on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
7. I am very confident about how this movie will make me feel.
8. I am very uncertain about how this movie will affect my mood. (RS)
9. I can accurately predict how this movie will affect my emotions.
Items 10-11 measure overall attitude toward the movie:
10. My overall opinion of this movie is …
Very Bad 1 2 3 4 5 6 7 Very Good
11. In general, I would (like/dislike) this movie.
Dislike 1 2 3 4 5 6 7 Like
Items 12-13 measure Risk Probability on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
12. There is a very good chance that this movie will not satisfy me.
13. I am likely to be disappointed by this movie.
Items 14-15 measure Risk Consequence on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
14. If I see this movie and don’t like it, I will feel very bad.
15. If I see this movie and don’t like it, I will be very unhappy.
RS: reverse-scored
94
APPENDIX C: STUDY 2 MANIPULATIONS
Condition Parent Affective Experience Parent-Extension Similarity
1 Positive High
2 Positive Low
3 Neutral High
4 Neutral Low
Condition 1: Imagine that you are at a bookstore shopping for a book to read for
pleasure. As you are browsing the New Releases section, you see a book by an
author whose books you really enjoy very much. You have read several of this
author’s previous books, and every one has put you in a very positive mood, made
you laugh a lot, and made you feel very happy. You see that this book seems to be
very similar to the books that the author has written in the past. It is in the same
genre (or style) as the author’s previous books, and even stars the same main
character that the author often writes about. The back cover states, “If you loved
(this author’s) previous books, then you’ll love this one even more!” Therefore, you
expect that reading this book would make you feel very similar to the way you felt
after reading the author’s previous books.
Condition 2: Imagine that you are at a bookstore shopping for a book to read for
pleasure. As you are browsing the New Releases section, you see a book by an
author whose books you really enjoy very much. You have read several of this
author’s previous books, and every one has put you in a very positive mood, made
you laugh a lot, and made you feel very happy. You see that this book seems to be
very different from the books that the author has written in the past. It is in a
different genre (or style) than the author’s previous books, and does not star the same
main character that the author often writes about. The back cover states, “Forget
everything you thought you knew about (this author)!” Therefore, you expect that
reading this book would make you feel very different from the way you felt after
reading the author’s previous books.
Condition 3: Imagine that you are at a bookstore shopping for a book to read for
pleasure. As you are browsing the New Releases section, you see a book by an
author whose books you enjoy somewhat, but not very much. You have read
several of this author’s previous books, and none of them has really affected your
mood, made you laugh a lot, or made you feel very happy. You see that this book
seems to be very similar to the books that the author has written in the past. It is in
the same genre (or style) as the author’s previous books, and even stars the same
main character that the author often writes about. The back cover states, “If you
loved (this author’s) previous books, then you’ll love this one even more!”
Therefore, you expect that reading this book would make you feel very similar to the
way you felt after reading the author’s previous books.
95
Condition 4: Imagine that you are at a bookstore shopping for a book to read for
pleasure. As you are browsing the New Releases section, you see a book by an
author whose books you enjoy somewhat, but not very much. You have read
several of this author’s previous books, and none of them has really affected your
mood, made you laugh a lot, or made you feel very happy. You see that this book
seems to be very different from the books that the author has written in the past. It
is in a different genre (or style) than the author’s previous books, and does not star
the same main character that the author often writes about. The back cover states,
“Forget everything you thought you knew about (this author)!” Therefore, you
expect that reading this book would make you feel very different from the way you
felt after reading the author’s previous books.
96
APPENDIX D: STUDY 2 MEASURES
Items 1-4 measure Affective Expectation on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
20. I expect that this book will make me very happy.
21. This book probably won’t put me in a positive mood.
22. There is a very good chance that this book will make me feel good.
23. It is very likely that this book will make me laugh a lot.
Items 4-7 measure Affective Forecast:
I forecast that after watching this movie, I would feel …
24. Neutral 1 2 3 4 5 6 7 Very Happy
25. Neutral 1 2 3 4 5 6 7 Very Good
26. Neutral 1 2 3 4 5 6 7 Very Positive
Items 8-10 measure Forecast Confidence on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
27. I am very confident about how this book will make me feel.
28. I am very uncertain about how this book will affect my mood. (RS)
29. I can accurately predict how this book will affect my emotions.
Items 11-15 measure Parent-Extension Similarity on the following scale:
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
30. This book is very similar to the author’s previous books.
31. The characters in this book are very similar to the characters in the author’s
previous books.
32. The plot or storyline in this book are very similar to the plot in the author’s
previous books.
33. The setting (time, place, etc.) of this book is very similar to the setting of the
author’s previous books.
34. This book seems a lot like the author’s previous books.
Items 16-19 measure Parent Brand Affective Experience on the following scale:
Not Much 1 2 3 4 5 6 7 Very Much
35. How happy did the author’s previous books make you feel?
36. How satisfied did the author’s previous books make you feel?
37. How positive did the author’s previous books make you feel?
38. How good did the author’s previous books make you feel?
RS: reverse-scored
97
APPENDIX E: STUDY 3 MANIPULATIONS
Condition Brand Strategy
1 Sequel
2 Nonsequel
Condition 1: Confederate Gold: After decoding several clues hidden in artifacts on
display in the Smithsonian museum, an archaeologist and his colleagues embark on
an adventure to find a treasure that was hidden long ago by the leaders of the
Confederacy during the American Civil War. However, another team of treasure
hunters is also searching for the same treasure. After his friends are kidnapped by
the bad guys, the archaeologist must choose between salvaging the lost treasure and
saving his friends lives. This movie is a sequel to the movie National Treasure, in
which an archaeologist decodes clues hidden in the Declaration of Independence and
other historic artifacts to find a treasure chest rumored to have been hidden by the
leaders of the American Revolution.
Condition 2: Confederate Gold: After decoding several clues hidden in artifacts on
display in the Smithsonian museum, an archaeologist and his colleagues embark on
an adventure to find a treasure that was hidden long ago by the leaders of the
Confederacy during the American Civil War. However, another team of treasure
hunters is also searching for the same treasure. After his friends are kidnapped by
the bad guys, the archaeologist must choose between salvaging the lost treasure and
saving his friends lives.
98
APPENDIX F: STUDY 3 MEASURES
Item 1 measures behavioral intention on the following scale:
Very Unlikely -3 -2 -1 0 +1 +2 +3 VeryLikely
19.Howlikelyareyoutoseethismovie?
Items2-4measureAffectiveExpectationsonthefollowingscale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
20. I expect that I would enjoy this movie very much.
21. This movie probably won’t put me in a positive mood. (RS)
22. There is a very good chance that this movie will make me feel good.
Items 5-7 measure Affective Forecast on the following scale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
23.Iforecastthatafterwatchingthismovie,Iwouldfeelverysatisfied.
24.Iforecastthatafterwatchingthismovie,Iwouldfeelverybad.(RS)
25.Iforecastthatafterwatchingthismovie,Iwouldfeelverypositive.
Items8-10measureForecastConfidenceonthefollowingscale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
26.Iamveryconfidentabouthowthismoviewillmakemefeel.
27.Iamveryuncertainabouthowthismoviewillaffectmymood.(RS)
28.Icanaccuratelypredicthowthismoviewillaffectmyemotions.
Next,thefollowingdescriptionofthemovieNationalTreasurewasgiven:
NationalTreasureisa2004movieinwhichanarchaeologist(playedbyNicolas
Cage)decodesclueshiddenintheDeclarationofIndependenceandotherhistoric
artifactstofindatreasurechestrumoredtohavebeenhiddenbytheleadersofthe
AmericanRevolution.
Item11measuresconsumptionoftheparentbrand:
29.HaveyouseenNationalTreasure?(circleone:)
a. Yes(continuetoquestion12)
b. No(skiptoquestion15)
99
Items 12-14 measure Parent Brand Affective Experience on the following scale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
30.AfterwatchingNationalTreasure,Ifeltverysatisfied.
31.AfterwatchingNationalTreasure,Ifeltverybad.(RS)
32.AfterwatchingNationalTreasure,Ifeltverypositive.
Items15-17measureParent-ExtensionSimilarityonthefollowingscale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
33.ConfederateGoldisverysimilartoNationalTreasure.
34.TheplotorstorylineinConfederateGoldisnotverysimilartotheplotof
NationalTreasure.(RS)
35.ConfederateGoldwouldmakemefeelthesameemotionsasNationalTreasure.
Items18-20measureDefensivePessimismonthefollowingscale:
Disagree -3 -2 -1 0 +1 +2 +3 Agree
36.Igointoamovieexpectingtheworst,eventhoughIknowthemoviewill
probablybeOK.
37.Itrytokeepmyexpectationsforamoviefairlylow,sothatIwillbepleasantly
surprisedwithhowmuchIenjoythemovie.
38.IusuallysetveryhighexpectationsforsequelsofmoviesthatIreallyenjoyed.
(RS)
Items21-22wereopen-endeditemstoelicitsubjects’thoughtsaboutConfederate
Gold,andtohavesubjectsratethevalenceofeachthoughtgenerated.
39.Inthespaceremaining,pleasewritedownasmanythoughtsaboutConfederate
Goldasyoucanthinkof.Youmayusethebackofthispage,ifnecessary.
40.AfterwritingdownallofyourthoughtsaboutConfederateGold,pleasegoback
andplaceaplus(+)signnexttoallofyourpositivethoughtsandaminus(-)sign
nexttoallofyournegativethoughtsaboutthemovie.Ifathoughtisneutral
(neitherpositivenornegative),pleasewriteazero(0)sign.
RS:reverse-scored

Abstract (if available)

Abstract Entertainment goods have two important characteristics: 1) they primarily provide affective benefits and 2) their quality cannot be determined prior to consumption. Therefore, when evaluating future entertainment experiences, consumers form expectations about the emotional gratification that they will receive, which depends upon the entertainment brand being evaluated and the consumer's characteristics, among other factors. This dissertation explores the process by which consumers form affective expectations and examines the impacts of brand strategy and consumer characteristics upon this process. The first study explores the process by which consumers form affective expectations, finding that consumers form affective expectations by weighting their affective forecast by their confidence in that forecast. The second study examines the influence of brand extensions on this process, finding that the similarity of a brand extension, such as a movie sequel or new book by a serial author, to its parent brand facilitates the transfer of associations from the parent brand to the brand extension, and that these associations form the basis for the affective forecast for the extension, which, in turn, impact affective expectations for the extension. Findings also suggest that similarity increases consumers' confidence in their affective forecasts for brand extensions, which impacts affective expectations for the extensions. The third study addresses a consumer characteristic, finding that consumers who are defensively pessimistic generate more negative thoughts about brand extensions than new brands, resulting in lower affective forecasts and expectations. Furthermore, this study finds that defensive pessimism strengthens the role of similarity on the transfer of associations from parent brands to brand extensions, and also strengthens the role of forecast confidence in the formation of consumers' affective expectations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

How product designs and presentations influence consumers’ post-acquisition decisions

PDF

The motivational power of beauty: how aesthetically appealing products drive purchase effort in consumers

PDF

Savoring future experiences: antecedents and effects on evaluations of consumption experiences

PDF

Forgiveness: elucidating the underlying psychological processes that foster brand forgiveness and interpersonal forgiveness

PDF

Essays on the economics of digital entertainment

PDF

Essays on understanding consumer contribution behaviors in the context of crowdfunding

PDF

Statistical algorithms for examining gene and environmental influences on human aging

PDF

Towards health-conscious spaces: building for human well-being and performance

Asset Metadata

Creator Anderson, Justin R. (author)

Core Title How sequels seduce: consumers' affective expectations for entertainment experiences

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Publication Date 07/26/2007

Defense Date 06/18/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag affective expectations,affective forecasting,brand extensions,defensive pessimism,Entertainment,movies,OAI-PMH Harvest

Language English

Advisor MacInnis, Deborah J. (committee chair), Folkes, Valerie (committee member), Park, C. Whan (committee member), Wilcox, Rand R. (committee member)

Creator Email jrander2@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m682

Unique identifier UC1204772

Identifier etd-Anderson-20070726 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-521839 (legacy record id),usctheses-m682 (legacy record id)

Legacy Identifier etd-Anderson-20070726.pdf

Dmrecord 521839

Document Type Dissertation

Rights Anderson, Justin R.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu