Master Thesis MSc Business Studies
UNIVERSITEIT VAN AMSTERDAM
“I can't understand why people are frightened of new
ideas. I'm frightened of the old ones.” – John Cage
An empirical study on the effects of innovative programming by large opera houses onthe appeal to its potential audience.
Name: Margot Stam, BA Student number: 5744008
Supervisor: Prof. Mr. Dr. N.M. Wijnberg
1
ABSTRACT
For opera companies balancing the need to innovate and satisfying the audience demand is
becoming more and more difficult. To support the decisions that need to be made, a greater
understanding of the audience, and how they react to innovation in programming, is
necessary. This study intends to do so by conducting an empirical study among potential
opera attendees and relating this to an existing measure of innovation in programming and
motivational goals of the audience. The results show a strong negative correlation between
innovation and audience appeal. The operas can be categorized into two groups. First,
innovative operas, which generate low appeal with the audience. Second, considerably
conform operas which generate high appeal to the audience. The sample shows three groups
into which the respondents could be clustered. Although non-significant, the clusters show
similar reactions to an increase or decrease of innovation. They do differ in overall score. The
difference between appeal ratings of cluster 1 is constant when altering the degree of
conformity of the presented operas. This implies a difference in basic appeal towards the
opera in general. Due to reliability issues of the core constructs of the motivational goals, the
correlation between these groups and the motivational goals as constructed by Cuadrado and
Mollà (2000) could not be tested. Despite this, theoretical interpretation of the results of
hypothesis 3 supports the division of the performing arts audience into two clusters according
2
TABLE OF CONTENT
ABSTRACT 1
TABLE OF CONTENTS 2
1. INTRODUCTION 4
2. LITERATURE REVIEW AND HYPOTHESES
2.1 Innovation and the performing arts 7
2.2 Audience and innovation 12
2.2.1 Attendance goals 14
2.2.2 Types of audience 19
3. DATA AND METHODOLOGY
3.1 Research method and design 22
3.2 Independent variables 23
3.2.1 DiMaggio-Stenberg Index of Conformity 23
3.3. Dependent variables 27
3.3.1 Audience appeal 27
3.4 Innovation according to the audience 28
4. ANALYSIS AND RESULTS
4.1 Descriptive statistics and preliminary data analysis 30
4.1.1 Socio-demographic profile 30
4.1.2 Audience appeal 30
4.1.3 Attendance goals 31
4.2 Hypothesis 1: Linear Regression Analysis 34
4.3 Hypothesis 2: Principle Factor Analysis 36
4.4 Hypothesis 3: Cluster Analysis 38
4.5 Hypothesis 4: Linear Regression Analysis (untested) 41
4.6 Audience and interpreting innovation in the opera 41
3
5. DISCUSSION 45
6. CONCLUSION 51
REFERENCES 52
4
1. INTRODUCTION
In the past, researchers have found that when arts organizations adopt an innovative work into
their repertoire, they are taking a risk in terms of whether the general audience is going to
appreciate it (Pierce, 2000; Liddle, Lupo and Borden, 2002; DiMaggio and Stenberg, 1985).
This is partly caused by the fact that artistic innovation in the composition of highbrow music
has not — in the past decades— been sufficiently attractive and exciting to the public to build
and hold audiences. In the eighteenth and nineteenth century, audiences listened to the music
of their own time, and were, for the most part, charmed and captivated by the innovations of
contemporary composers. In the twentieth century, contemporary concert music has largely
failed to capture the public’s commitment. Contemporary, and also modern, opera seems to
have become a world of purists. Unfortunately, these purists alone do not have the capacity to
fully support the opera in its existence. I have experienced the different appreciations of
artistic innovations by audiences myself. While studying musicology, my fellow students and
I were delighted to discover that Die Soldaten by Bernd Alois Zimmerman would be
performed by De Nederlandse Opera (DNO) in the Muziektheater in Amsterdam. Die
Soldaten is a 20th century opera with unconventional composition techniques. Even though we
were blown away by the quality of the performance, I also noticed a significant part of the
audience leaving the concert hall, long before the last note was played.
The diversity of program appreciation significantly increases the revenue risk that
artistic departments need to consider when programming new, or innovatively interpreting,
old operas (DiMaggio and Stenberg, 1985). Although this is part of the considerations made
by opera companies when programming for the coming seasons, artistic departments within
often heavily guard the artistic freedom of interpretation. The sales and marketing
5
reach as many potential consumers as possible. Unfortunately, actual empirical evidence on
the influence of artistic innovations on the size and characteristics of the audience, which
could support sales and marketing manager, is either absent or limited.
Attending to this research gap, this study will relate audience appeal in the highbrow
performance arts, more specifically the opera, to a measure of artistic innovation directed at
repertoire and programming. Scoring operas according to measures of innovation and relating
those ratings to the audience appeal will create insight into the effects of artistic innovation;
such as: the size of box office sales and the type of audience it attracts. Previous research on
innovation is abundant. Additionally, the field of artistic innovation also has a substantial
body of literature to build upon (Castañer and Campos, 2002; DiMaggio and Stenberg, 1985;
Heilbrun, 1993, Heilbrun, 2001; Kim and Jensen, 2011; Pierce, 2000; Martorella, 1977).
Mainly these studies have focused on innovation in repertoire choices, ‘form’ and ‘content’
innovation by arts organizations. This excludes innovations within the performance, in the
sense of directory and dramaturgy. This means that the influence on the potential audience
regarding their perception of how an innovative piece looks is overlooked. I will also rank the
operas according their innovation in form, but then relate these rankings to the audience
appeal and type, focusing on implications for marketing departments regarding the
appreciation of the piece by the audience prior to attendance. By relating these measures to
the audience, this study addresses a gap in the existing literature that can bridge the theoretical
knowledge on artistic innovation, and concrete practical implications for managers in
marketing and sales. The results could support managers in their decisions on marketing
strategies, their directional focus and help allocate their resources in a more effective manner.
In addition to this I will explore the interpretation of the term innovation, also
addressing innovation within the performance. Taking canonical works, rethinking these and
6
mis-en-scène. This particular point has been underrepresented in previous research. Probably
due to the ‘eventness’ of a live performance. Also, since opera staging seems to be embedded
in the historical setting and the social context of the times, it is difficult to control for all of
these influences. Quite similar to the problem of naming on a referent when deciding which
repertoire choices are innovative and which are not (Levin, 2008).
During the literature review I will elaborate on the key constructs and the existing
literature regarding those constructs — creating a view of what is known about the subject.
Throughout this section I will also state the hypotheses. I have divided this literary review
into two parts: artistic innovation, innovation and audience, respectively the independent and
7
2. LITERATURE REVIEW AND HYPOTHESES
2.1 Innovation and the performing arts
Artistic innovation in the production of highbrow music is often associated with the
production of contemporary work (Castañer and Campos, 2001; Heilbrun, 2001; O’Hagan
and Neligan, 2005). However, in reality there is a lack of consensus regarding the definition
of innovation in opera production. There exists a large gap between what is innovation to
potential audience, and what is innovation to the artistic department of a theater. Focusing on
repertoire, you often see that certain operas are considered innovative by the average potential
customer of the opera, whilst in reality these operas have been part of the canon for some
time. I refer to many operas written in the twentieth century like operas by Berg, Stockhausen,
Adams and Britten. These are composers of operas that are performed often and would, in the
context of present day, not be considered particularly innovative by experts. However,
instinctively you can imagine the average audience to classify these operas as innovative. This
discrepancy might illustrate what actually entails innovation. Taking the definition of
innovation as something ‘new’ to a field, innovation is determined in relation to what has
gone before (Schumpeter, 1942). More specifically, what has gone before to the best
knowledge of the referent. Due to the boundaries of knowledge of and exposure to past
compositions, the average audience member will have a different notion of what is innovative
as compared to the artistic departments of opera companies, which mainly consist of experts.
Moreover, both managers and audiences tend to take a more self-referential approach. This
negatively contributes to the consensus among researchers regarding the definition of
innovation in high-art organization such as opera companies (Castañer and Campos, 2011). It
illustrates a gap between the interpretations of innovation, and implies that managers tend to
8
programming, on what attracts what type of audience, and thus where they need to direct their
marketing efforts towards.
Traditionally, artistic innovation is defined as the introduction in a field or market of
something new (Schumpeter, 1942). In addition, Castañer and Campos (2001) propose that
the process of determining artistic innovation is entirely dependent on who or what is
identified as the referent. Like mentioned earlier, in the field of theater, managers tend to take
a self-referential approach. This approach is not appropriate when dealing with innovation,
since you cannot determine how ‘new’ a production is. You must be able to determine if it is
new to a group of people, to a field, or maybe even a complete departure from conventions
(Becker, 1982). Particularly when using this information in the light of marketing and sales,
your decisions relate to the audience. This should reflect in naming your referent. A referent
outside of the focal theater environment is essential for research, as the participant’s views
will be contributed towards the tabulation of empirical data used in this study. Especially in
the world of highbrow performing arts like opera, this is a major problem. Within the arts
organizations, there are two different orientations. The art professionals, like art directors,
tend to have an art-centered approach and have difficulty including marketing oriented
considerations in their programming. In their orientation their work is in the benefit of the
organization, whilst actually they are contributing to putting the organization in jeopardy
(Kotler, 1997). Robert Kelly (1993) says: “marketing is feared and even hated for what it
might do to the arts; on the other hand, there is an unquestioning conviction that marketing
can work miracles for the arts. To compound matters, these contradictory beliefs are often
held by the same persons. The arts professionals must become more heavily involved in the
marketing decision-making process, since avoidance will lead to ineffective marketing, not to
an absence of marketing.” Kotler (1997) writes that the arts organizations, like opera
9
product focus, to a focus that balances the artistic decision-making process with the needs and
preferences of the audience, something the marketing department can facilitate. Performing
arts do not exist in a societal vacuum and should not be treated as such whilst expecting to
survive from its contributions. The degree and way in which an opera house chooses to
innovate and its effect on different types of audience is of great significance to this goal.
According to Castañer and Campos (2001) there are two ways in which arts
organizations can innovate: in content and in form. Note that this is different from the way art
can innovate. The general notion of innovation in form is the way a piece is performed in
relation to its style, techniques and general design. Examples are décor, costumes,
choreography, including other art forms and interaction with the audience. Content refers to
the essence, or meaning, of the piece which can be carried out through the elements belonging
to the form of the piece. Castañer and Campos do not discuss the way the artist can innovate,
but how the art organization can innovate. Form innovation in the organizational sense can
refer to the programming of an opera that is new to the audience. By researchers like
Kimberly and Evansko (1981) and Damanpour (1995) this is criticized. They do not find this
an act of innovation, but an ‘adoption of external innovation’. However, since the referent in
this study is the potential audience of the opera, the programming of these operas has
important implications toward the appeal of the audience and should not be excluded but
magnified and investigated.
Innovation in content by arts organizations is defined as strategies considering opera
selection, sequence and timing of the entire program. In this study, innovation in form as
proposed by Castañer and Campos is the form of innovation under investigation.
During previous research, different approaches have been developed to measure
artistic innovation in theaters. The majority focuses on programming and repertoire; for
10
conventionality of certain pieces of music (Campos and Castañer, 1998; DiMaggio and
Stenberg, 1985; Heilbrun 2001; Martorella, 1977; Pierce, 2000).
The DiMaggio-Stenberg Index of Conformity is an index on the mean number of times
a composition, which is performed by the focal theater company, is also performed by other
theaters within the scope of the research. The higher the conformity score, the less innovative
an opera is considered. DiMaggio and Stenberg (1985) chose this strategy in measuring
innovation by theater companies, and therefore define innovation by arts organizations as
non-conformity in repertoire. By using this method they include old works that are brought
back after a long period of time, instead of only focusing on recently composed works. This
method can be considered incomplete since it only addresses an organization’s innovation in
the sense of form, and not the content of the production (opera selection, sequence and timing
of entire repertoire). However, it does imply something of significance. It implies that
something innovative does not have to be new as in its years of existence, but merely needs to
be new ‘relative to the existing state of the relevant art form’ (Castañer and Campos, 2001).
Where DiMaggio has provided a useful method of measuring innovation by arts
organizations, innovation at the level of the art performance has not been blessed with such a
concrete method. It remains a somewhat abstract concept with multiple parameters that are
yet to be defined in a standardized manner. Besides the programming of the organization, the
staging, operatic mis-en-scène and how this carries out the meaning of the piece has an effect
on how innovative a performance can be seen. When looking at innovation as something
‘new’, it has a lot to do with expectations and past experiences. In order to be able to make
useful recommendations on how to rate these aspects of opera, you need to have a strong base
of knowledge as to what is already out there.
According to Aronson (1991) the contemporary staging is not just aimed at staging the
11
one whole. This creates a scenography that is stylistically part of the whole and not just a
context of the other aspects of the opera. However, a decade has passed since the publication
of the article by Aronson, and the term ‘contemporary’ might not apply to the operas he was
referring to at the time. Also, taking the Schumpeter’s definition of innovation in account, it
would be impossible to concretely pin-point what is innovative before it has prevailed itself. It
seems that the only strategy to define innovation in the performing arts is to identify an
innovation that has already been created and possibly could create followers and become the
dominant design. In the opera this may be of moderate proportions since it is a field that
attracts purists that support conventionality on principle. They wish their opera to be ‘pure’
and unadulterated. However, tampering with the design and context has resulted in drawing a
larger and more diverse audience (Bolstein, 1994). Generally speaking the opera is (in)famous
for its conservative programming. Opera companies often have about two dozen titles in their
repertoire that are expected to sell out (Miller, 1995). These are performed to tend to the
demand and deal with the financial constrains the companies face. For instance the
programming of world premieres, is a high-risk endeavor. The audience will be unfamiliar
with the opera and not attend (Pierce, 2000). To determine the strength of this effect I propose
the following hypotheses:
Hypothesis 1: Operas that score low on form innovation, create a stronger appeal with
potential audiences than operas that score high on form innovation.
Hypothesis 2: The audience appeal can be categorized into two groups that are in
12
2.2 Audience and innovation1
At the core of this study lies the appeal of the potential audience. This section will discuss
methods that have been used by other researchers to group the audience according to different
parameters. This will provide an introduction into the grouping methods that will be used in
this research.
Grouping audiences has been of interest to researchers in the past (Belk, Semanik and
Andreasen, 1980; Dimaggio, Useem and Brown, 1977). However, most of them had the
tendency to solely focus on demographics (education, income, occupational status and racial
minorities), which has resulted in the grouping of certain types of audiences among different
art related leisure activities. More relevant to this research are the studies that have
investigated the characteristics within these demographic groups (Boorsma, 2006; Cuadrado
and Mollà, 2000).
Boorsma (2006) elaborated on this by focusing on the consumer’s value consideration
when consuming art. She builds from the conceptualization of hedonistic and experiential
consumption developed by Holbrook and Hirschman (Hirschman and Holbrook, 1982;
Holbrook and Hirschman, 1982). Analyzing the hedonistic perspective requires to ignore the
utilitarian values involved in the decision-making of a consumer, and to focus on the pleasure,
hedonistic fulfillment, emotional arousal, amusement, and imaginary and sensory stimulation
experienced by the consumer. The performing arts, such as the opera, are a typical example of
experiential products. The consideration of the consumer is mainly directed at its intrinsic
value, which refers to the experiences sought for its own sake that will inherently give
pleasure (Holbrook et al., 1984). All possible extrinsic benefits that might arise are considered
1 Throughout the literature review theory on audience that was originally set in context of symphonic
performances will be applied to the opera. To ensure the applicability I tested the correlation between opera audiences and symphony audiences based on basic demographics. The data was retrieved from the report of a large scale research among American performing arts attendees (Nichols, 2002). The results showed a
13
less important. This information lays a foundation for researchers to conceptualize the
emotional responses to artistic stimuli and, most relevant to this study, the motives for seeking
these experiences. According to Woods (1987) this conceptualization must incorporate the
differences for each individual. Audience members differ regarding their cognitive capacities.
For instance, because states of discomfort motivate the consumer more than external
attractiveness. People seek challenges for their capacities and are encouraged by the mental
anticipation of the expected experiences. To further understand the consumers, researches
needed to specify these hedonistic experiences and also have a way to group audiences
according to these experiences.
Research by Woods (1987), Mannel and Iso-Ahola (1987), Csikzenmihalyi (1996) and
Colbert et al. (2001) constructed two key classes of hedonistic experiences. The first class
consists of stimulating, exciting, surprising and/or challenging experiences. The second class
entails relaxing, entertaining and/or comfortable experiences. The first class is motivated by
novelty, challenge and stimulation, whereas the second is motivated by an escape from
stressful daily life. Holbrook and Zirlin (1985) have conceptualized this towards the
performing arts, such as the opera, and made the division between profound aesthetic
experiences and simple hedonistic pleasure.
Building upon this, researchers like Cuadrado and Mollà (2000) have taken similar
attendance goals and developed them into four general clusters of motivations after
conducting a survey among Spanish theater attendants. This resulted in four different
audience types: beginners, enthusiast, theater buffs and indifferents. The four constructed
attendance motivations are social hedonism, emotions, cultural fulfillment and interest. This
division by Cuadrado and Mollà (2000) will play a key part in the research section of this
14
Grouping the audience in separate clusters provides the organization with information
of value, especially since these clusters are based upon the reason they attend the opera.
However, the first step is to identify these groups and relate them to something the opera
companies can control, referring to the operas they program. Taking this into account I
propose the following hypothesis:
Hypothesis 3: Within potential opera audience there exist distinct groups based on
their appeal towards innovation in form.
2.2.1 Attendance goals
The investigation of audience attendance has been of interest to researchers for some decades
now (Bergdàa & Nyeck, 1995; Cooper & Tower, 1992; Colbert, 1991; Cooper-Martin, 1991;
Steinberg, Miaoulis & Lloyd, 1982; Hawes, 1978). It has also been of interests to marketing
departments. The aim is to identify the drivers of consumer behavior. Whether it is called
goals, needs, values or motivations, they all aim to do the same (Hawes 1978). By looking
into the psychology of potential audience you add an extra dimension to the determination of
what strategy will optimize the size of the audience. The use of this dimension comes forth
out of the notion that people have different needs and perceive different benefits when it
comes to their leisure time (Hawes, 1979). Experiential goods like the opera, include
intangible, hedonic and affective contemplations. These contemplations are more important
than the utilitarian considerations and differ from person to person, based upon their personal
and social features (Hawes, 1979). According to Cooper-Martin (1991) the most important
consideration is hedonism, in several different forms. The motives are directed at oneself;
15
another picture was painted by Cooper and Tower in 1992 who set up a collection of needs to
illustrate the different types of motives people have for attending the performing arts (table 1).
Ideal needs Emotional
needs Educational needs Social needs Personal needs
Esthetics Stimulation Eduaction Interaction New experiences
Beauty Fantasy Growth Sharing Escape from boredom and routine
Transformation Imagination Spiritual nourishment' Contact Entertainment
Transcendence Relaxation Social play
Greater awareness
Table 1. Coop and Tower (1992). Needs satisfied by the performing arts.
Bouder-Pailler (1999) took a slightly different approach and investigated the
underlying dimensions of the goals stated by researchers like Cooper and Tower (1992),
Cooper-Martin (1991) and Bergdàa & Nyeck (1995). They state that the distinction between
cognitive and affective is only part of the intrinsic motive. By including an extrinsic motive
they also include a social component, social hedonism. This consists of a need to ‘be with
others’ and ‘to belong’, including to improve social status.
16
Building upon this heritage, Cuadrado and Mollà (2000) proposed taxonomy for
performing arts attendees, not only by creating a set of motivational goals, but also to
characterize several different types of attendees. Later on in this literature review I will
elaborate on the different types. First, I will discuss the four different attendance motives as
constructed by Cuadrado and Mollà (2000): social hedonism, emotions, cultural fulfillment
and interest.
Social hedonism
One definition of hedonism is the pleasure of sensuous gratification for oneself (Bilsky &
Schwartz, 1994). It stands in close relationship to an openness to change and
self-enhancement. Directed at the pursuit of emphasizing someone’s dominance over others, it is
in psychology often associated with self-centeredness and selfishness. In this case, however,
we are focusing on social hedonism which also includes a value that seems contradictory, e.g.
love, kindness, sympathy (Bilsky & Swartz, 1994). Including utilitarianism helps us to
understand social hedonism better.
Utilitarianism is an ethical philosophy in which the greatest good is to achieve the
happiness of the largest number of people possible. Robert Perloff (1987) explained that
utilitarianism “promotes the greatest happiness principle and advocates an existence that is
exempt as far as possible from pain and contains as much enjoyment as possible”. However,
Corlett and Angelo (1988) criticized this statement because it implies the goal is solely
directed at personal gain, instead of the ‘greater good’. They state that utilitarianism is a social
hedonism that requires moral agents to reach the optimal enjoyment as possible. Not just for
the individual, but for the maximum amount of people. According to Cuadrado and Mollà
(2000), in the case of performing arts audience’s social hedonism expresses itself through
17
social prestige and signaling one’s own social prestige by attending the performance or just by
having the opportunity to dress to the occasion.
Emotions
The effect of the performing arts on the human emotions has been a point of interest to people
for as long as we can remember. Whereas Frijda (1986) stated it can evoke a high degree of
excitement and admiration, Goldman (2001) confirms that are often experiences felt that go
beyond the experience of pleasure only. The emotional attendance motivation is based on
pleasure, hedonistic fulfillment, arousal, relaxation, amusement, and sensory and imagery
stimulation of the individual. It refers to a more general need to experience emotion through
the performance, based on dynamic interactions between the consumer and the product, in this
case the audience member and the opera performance (Boorsma, 2006).
Cultural fulfillment
Cultural fulfillment is an attendance goal that refers to having the motivation to share an
experience, gain educational development and self-fulfillment through the arts in general
(Cuadrado and Mollà, 2000). The educational motive is not provided with the essential
interpretation of what this entails. Past literature can help us interpret cultural fulfillment. It
can be interpreted by having to do with the cohesion of a social class. For centuries the
high-arts have provided the upper social classes with cultural capital. According to DiMaggio and
Useem (1978) the homogeneous audiences of the high performing arts are the result of a
tradition of cultural socialization in the upbringing of generations of the higher social class.
Exposure to the high arts at home as well as in school provides them with the skills to decode
the arts. DiMaggio and Useem (1978) explain this as follows: “Artistic meaning is encoded in
18
paintings of the impressionists or the musical compositions of Debussy and Stravinsky, may
seem natural and harmonious to a generation that has assimilated the conventions implicit in
them. Individuals must learn to "read" a painting or a piece of music just as they must learn to
read the printed word.”
Taking this into account, the educational value of the opera towards the audience
could be interpreted as the need to develop these high-culture decoding skills. Nonetheless, it
does raise questions regarding the significance of reinforcing the cohesion of a social class
opposed to the cultural enrichment of the individual (DiMaggio and Useem, 1978).
Cultural fulfillment is also characterized by the motive to share an experience and seek
self-fulfillment. One could interpret this in combination with the educational motive. Through
the arts they try to enrich themselves culturally as a result of their . Although speculative,
self-fulfillment may be the consequence of recognizing their capacity to decoding the arts and
share the experience within their social class.
Interest
Although cultural fulfillment also includes a great deal of interests, it is a general interest in
the arts. Interest is directed at a more specific interest. This motive is driven by the attendant’s
interest to see particular artists or the work of a specific director (Cuadrado and Mollà, 2000).
Cuadrado and Mollà (2000) mention that this attendance goal is a departure from the theory
by Bouder-Pailler (1999) upon which they based their hypotheses. They do not provide an
interpretation of the origin of this attendance goal, but there is room for interpretation. This
goal is associated with two audience types, theater buffs and enthusiast, that will be discussed
in the next section. These groups have the highest frequency of attendance, which helps with
the interpretation of this motive. The prevalence of this factor in these groups could exist due
19
their familiarity with artists and directors, they have the ability to make this part of their
consideration to attend. Potential audience with a lower attendance frequency is not aware of
the difference between artist and directors and cannot create a preference in that aspect.
Having a preference, or at least an awareness of the artists and directors, is consequently
accompanied by a sense of familiarity. Familiarity could account for the benefits of having an
interest in particular director or artists, because it is known to influence positive emotional
reactions to art (Pierce, 2000; Scherer, 2001; Silva, 2005).
2.2.2. Types of audience
The characteristic of opera attendees have been discussed for a long time. It has painted a
clear picture, with little surprises regarding the demographics of the audience. The typical
opera attendee is non-Hispanic white, between the age of 45 and 54, is a college graduate and
has an income of more than $75,000 dollar a year (Nichols, 2003). While this information has
certainly has been useful, especially to researchers interested in social class, it does not grand
us much understanding regarding the diversity of attendance goals among the audience
(DiMaggio & Unseem, 1978; DMaggio, 1987). When looking at the attendance goals as the
antecedents of consumer appeal, you can imagine that they differ among this seemingly
homogeneous audience. It are the differences within the audience regarding the relative
significance of these motivations that are most interesting. People in the audience seem to
differ in a variety of characteristics that can pinpoint the foundation of their appeal towards a
product or service (Cooil et al., 2007; Anderson et al., 2008). Only until recently, researchers
shifted their view from solely basic demographics to motivational characteristics (Cuadrado &
Mollà, 2000; Bouder-Pailler, 1999; Cova, 2003.) By identifying different types of audiences
by their attendance goals, marketing departments of opera houses can allocate their efforts
20
into frequent attendants, results will be greater than solely focusing on the groups that already
attend regularly. Also, through market research opera companies can gain insight into the
composition of their audiences and identify which type of audience is underrepresented and
needs extra attention. The following types that will be discussed are distinguished by
Cuadrado and Mollà in 2000.
Beginners
Beginners is an audience group that is young and highly educated. They attend performing
arts like the opera for emotional and cultural fulfillment. They do not have much interest in
the social factors of a theater visit. An interesting statement is made by Cuadrado and Mollà
(2000). They state that beginners have the potential of becoming theater buffs because of their
interest in the emotional and educational value of attending the theater. Theater buffs have a
higher frequency of visits; therefore this is an especially interesting group for the marketing
department to target.
Theater buffs
Theater buffs are on average of an older age, but also highly educated. They attend the opera
for cultural enrichment and to see particular artists, directors and pieces. This type of audience
finds all attendance motives important, but in a clear order of significance. Most important is
the educational value, followed by the interest in performing arts in general and the emotional
aspect. The social aspects are barely considered.
Enthusiasts
The group that is considered enthusiasts consist on average of the same age groups as the
21
the enthusiasts all factors are important, however, most essential are the emotional motives,
followed by the educational motives. Theater buffs are the only ones that consider the social
factors in their decision to attend.
Indifferents
The naming of this audience group comes from the fact that they are not overly focused on
any of the attendance goals. They are mainly interested in having a good time. On average,
they are young of age and less educated than for instance the beginners and the theater buffs.
Indifferents tend to go to venues and performances that are convenient and familiar.
The goal of this study is to bridge the theory around that relates innovation with
audience appeal and the theory that relates motivational goals and subgroups within the
audience. The theory of Cuadrado and Mollà (2000) goes into the reasons why people attend
the opera. Relating this to the repertoire choice of the opera companies will provide
information on how the program can motivate different types of attendees within the potential
audience. The final hypothesis attempts to relate the identification of audience types (based
upon their attendance motivations) to an aspect that the opera companies can control
(programming). Taking this and the discussed theory into account I propose the following
hypothesis.
Hypothesis 4: The clustered groups within potential opera audience based on their
appeal towards innovation in form correlate to the attendance goals and
22
3.
DATA AND METHODOLOGY
During this chapter, the research method, different variables and analysis methods will be
described.
3.1 Research method and design
To achieve the objectives set in this study, I included an established measure of innovation
(DiMaggio-Steinberg Index of Conformity) and conducted an empirical study in the form of
an online survey among potential audience members of the opera. The conformity variable
will provide data that indicate the degree of innovation of particular operas. This will function
as an independent variable. The empirical study will provide the dependent variables in the
form of the appeal of the specific pre-selected operas and an independent variable that
indicates the relative importance of the different attendance goals.
An online survey was chosen to be most appropriate for this study due to the low
amount of costs involved and the degree of convenience it brings when dealing with the short
amount of time to conduct the research. Also, its explanatory nature, rather than just
descriptive, is particularly suitable. The survey was held among potential opera attendees who
were reached through personal networks and other platforms like forums and newsletters.
This strategy was chosen to optimize the reach of the survey during the limited time
(Saunders et al., 2009). In the survey the respondents were asked to answer questions that
addressed their appeal to visit specific opera pieces, their attendance goals when visiting the
opera and some basic demographics (age group, sex, highest level of education, employment
status, and household composition). Before the start of the survey the respondents were
informed that the results will stay anonymous and confidential. Also, as an incentive to
23
winning one of two theater gift cards worth 25 euro. The first two questions served as a
control to make sure the respondent was interested in visiting the opera, which would
determine whether they are a potential opera attendee or not. Respondents with no interest in
visiting the opera were excluded from later analysis.
The survey consisted of three parts with a total of 40 questions. De Nederlandse Opera
(DNO) served as a referent in the study, because of its unique position in the Netherlands and
their stable and considerably consistent productions (Castañer and Campos, 2011). DNO does
not have regularly varying artistic innovations that can interfere with what was intended to be
measured and hurt the reliability of the study. I refer to innovations that might distract the
consumer from the program, like an alternative performance location, special effects or other
artistic innovation in content. Audience members are therefore able to form an expectation
when considering attending the opera. The survey was spread among Dutch residents. For a
sample consisting of (potential) opera attendees is required, the data will not solely be
collected from the audience from DNO, but from potential audience within the reach of the
Muziektheater, the theater that hosts DNO.
A total of 120 questionnaires was started. With a dropout rate of 31,67 % and 13
respondents not being potential opera attendees this led to 79responses that were suitable for
analysis.
3.2 Independent variables
3.2.1 DiMaggio-Stenberg Index of Conformity
As discussed in the literature review, this index measures the degree of conformity of an
opera when programming it for production. In other words, the DiMaggio–Stenberg index of
conformity measures the commonality of repertoire. The higher the conformity score, the less
24
allowed me to rate opera’s according to DiMaggio and Stenberg (DiMaggio and Stenberg,
1985). The database includes 23 large opera theaters in Europe. It consists of the
programming of major opera houses in Europe during the seasons 2008/09 until 2012/13,
listed below. The listings of the included opera houses were retrieved from an online opera
archive, Opera Critic, including information on productions throughout the world (The Opera
Critic, 2013). To ensure the reliability of this source random checks were made in reference to
the official program archives of the opera companies included. This resulted in an one on one
correspondence. The database contains 2484 performances of 496 operas by 237 different
composers. The average amount of performances per opera per company will determine the
conformity index for the operas (DiMaggio & Stenberg, 1985). The following opera houses
were included:
1. Barcelona - Liceu 2. Berlin - Staatsoper 3. Berlin - Deutsche Oper 4. Bruxelles - La Monnaie
5. Copenhagen - Royal Danish Opera 6. Dresden - Semperoper
7. Frankfurt - Oper 8. Göteborg - Opera 9. Köln - Oper Köln
10. London - English National Opera 11. London - Royal Opera
12. Hamburg - Staatsoper
13. Madrid - Teatro Real 14. Milano - Teatro alla Scala 15. Munich - Bayerische Staatsoper 16. Oslo - Norwegian Opera House 17. Paris - Opera National de Paris 18. Stockholm - Royal Opera 19. Stuttgart - Staatstheater 20. Venice - La Fenice 21. Vienna - State Opera
22. Vienna - Theater an der Wien 23. Zurich - Opera House
A selection of 15 operas was made to be used for the test-survey. The operas were
selected to represent the spread of conformity scores of the database. This resulted in an
overrepresentation of unfamiliar operas, which led the respondent to lose interest during the
survey. Possibly due to a lack of familiarity. Therefore, I included more extremes in the
survey by adding titles with a high degree of conformity. The following operas were selected
25
Title Composer Index of Conformity
Le nozze di Figaro Mozart 2,260869565
La boheme Puccini 2,217391304
Il barbiere Rossini 2,173913043
Carmen Bizet 1,695652174
Lulu Berg 0,434782609
Agrippina Handel 0,217391304
The gambler Prokofiev 0,130434783
Retable de maese Pedro Falla 0,086956522
I puritani Bellini 0,086956522
Die feen Wagner, R. 0,043478261
Gisei - Das Opfer Orff 0,043478261
Les Mamelles de Tiresias Poulenc 0,043478261
Poliuto Donizetti 0,043478261
Zémire et Azor Gétry 0,043478261
Aschemond Oehring 0,043478261
Table 2. DiMaggio-Stenberg Index of Conformity. Operas included in the survey.
3.2.2 Attendance goals
Part of the survey consists of a question where respondents are asked to rank their attendance
goals for when they visit the opera to be able to identifying the type of audience member the
respondent is; beginner, theater buff, enthusiast or indifferent according to Cuadrado and
Mollà (2000) with slight modifications to apply the method to opera and to the target sample
of this study. This method was founded by preliminary theoretical research on methods to
group the audience (Cuadrado and Mollà, 2000).
1. A way to spend an evening 2. Educational development 3. Entertainment
4. Interest in the arts in general 5. Relaxation
6. Self-fulfillment
7. Social interaction with fellow opera enthusiasts
8. Social prestige
9. To be able to participate in future conversations about the opera
10. A way to spend time with friends 11. To dress up
12. To feel emotion
13. To see particular artists 14. To share an experience 15. To see a director’s work
26
The way these attendance goals will be ordered will show the relative importance of
the four main goals social hedonism, emotions, cultural fulfillment and interests for each
respondent. The ranking question asks for a careful interpretation since it is a question that
includes probabilities without replacement. However, it is exactly the relativity of the
attendance goals, and later the motivational patterns, that will show what type of audience
member the respondent is. Table 4 shows the relative importance of the attendance goals to
the specific types of audience. Table 3 shows the motivations presented to the respondents
and with which attendance goals they are associated.
Table 3. Motivations grouped by attendance goals (Cuadrado & Mollà, 2000).
Social hedonism Emotions Cultural Fulfillment Interests
Beginners ** **
Theater buffs ** *** **
Enthusiasts * *** *** **
Indifferents ** * *
Table 4. Relative importance of the attendance goals per audience type (Cuadrado & Mollà, 2000)
Social hedonism Emotions Cultural fulfillment Interests
Social interaction with fellow opera enthusiasts
Entertainment Educational development Interest in a particular artist A way to spend time with friends Relaxation Self-fulfillment Interest in a director's work A way to spend an evening To feel
emotion
To share an experience To dress up Interest in the arts in
general Acquire social prestige
To be able to participate in future conversations about the opera
27
Since the grouping of audiences by Cuadrado and Mollà2000) was strictly conducted
under symphony attendees it is important to test the applicability of its constructs on to opera
attendees.Also, their sample consisted of Spanish audiences, which do not necessarily lead to
results corresponding to the Dutch public. To test if the audience types that Cuadrado and
Mollà (2000) have constructed are applicable to this sample a correlation matrix will be
constructed and a reliability analysis will be conducted on the constructs through principle
factor analysis. The correlation matrix will show the correlations between the items and
provide the first information on the possibility for computing mean variables for the items that
are associated with the same motivational goal. After the construction of these mean variable
the sample can be grouped into groups based on the relative importance of the motivational
goals as shown in table 4.
2.3 Dependent variable
2.3.1 Audience appeal
During the first part of the survey respondents were asked questions regarding the probability
for them to positively consider attending the pre-selected operas based on the title and
composer. More specifically, they were asked to indicate the likelihood for them to positively
consider attending the proposed opera if it were performed by DNO in the Amsterdam
Muziektheater. This block consisted of 15 questions, all including one opera title. The
presented operas had been ranked beforehand according to innovations regarding
programming. To do this I applied the DiMaggio-Stenberg conformity index (DiMaggio and
Stenberg, 1985). During the survey respondents were asked to rate the opera on a 5-point
Likert-scale on how likely he or she is to attend were it to be performed the Muziektheater in
Amsterdam (Likert, 1932). Using this method will require the respondent to give an answer,
28
room for nuances and more reliable results. The scale was posed as follows: (1) very unlikely,
(2) unlikely, (3) neutral, (4) likely, (5) very likely. These measuring methods were used
because they play into the familiarity with a particular opera. It reads into the innovation of
the repertoire only and not into the innovation of the performance. The audience will decide
on an opera before knowing what exactly will be presented to them. This is typical for the
performing arts and makes the consideration of the potential audience essential. The proposed
operas will not be accompanied by texts or images, as seen in the advertisement of opera
performances, to exclude possible intrinsic messaging and rhetoric signaling (Ehses, 84;
Gardner 2000; Scott, 2010).
After data collection, I will deal with missing data according to the most appropriate
methods, based on the amount and type of missing data. Then, a descriptive analysis will be
conducted and the first hypotheses will be tested through a linear regression analysis of the
means of the appeal per opera and the degree of innovation according to the
DiMaggio-Stenberg index of conformity of the operas included in the online survey.
Hypothesis 2 requires the identification of separate themes in the ratings of the operas.
These trends will be uncovered through a dimension reduction in the form of a principle
factor analysis. This statistical method will decrease the number of variables (15) and show
which operas are rated in the same manner. The results will show if there are indeed distinct
ways in which specific operas are rated. At the same time, it will identify which of the operas
belong to which factor. The next step investigates if the different factors in the appeal ratings
are correlated with, and due to the difference in conformity. This will be determined through a
linear regression analysis.
Coming towards grouping the audience, hypothesis 3 is aimed at finding different
clusters within the sample. By conducting a cluster analysis of the appeal ratings of all the 79
29
run three separate linear regression analyses between the mean appeal of the cluster of each
opera and the conformity index scores of that same opera. This will determine if, and to what
degree, the difference between the clusters is based upon innovation in form.
The ability to test hypothesis 4 depends on the results of the reliability analysis aiming
to identify the motivational pattern factors initially identified by Cuadrado and Mollà (2000).
Positive identification will lead to the grouping of the sample according to the 4 clusters
within the audience through a cluster analysis and will validate running a linear regression
analysis of the means of appeal of the separate clusters and the conformity scores of the
operas. Negative identification will lead us to conclude the clusters are not applicable to this
sample and to a discussion in the limitation section of this study. Also, for exploratory
reasons, a factor analysis will be conducted to show possible other groups of variables that
reduce the amount of variables into just a few combined mean variables.
2.4 Innovation according to the audience
At the end of the survey respondents were asked to describe what they would expect to
encounter if they were to visit an ‘innovative’ opera. In a way it is an answer which cannot be
answered since innovation is mostly characterized by being something new, and thus
unexpected. However, it will give an insight into the expectancy of potential audience when
there is advertised with innovation. It regards an open question and the respondents are free to
answer in any way. The responses will be analyzed and reduced to a set of key themes. This
will lead to a list of themes with the frequency that they were mentioned. I do not expect to be
able to create standardized results; this part of the research is conducted for exploratory
30
4. ANALYSIS AND RESULTS2
During this chapter the results of data analysis will be presented. Analysis was conducted
using the statistical software program IBM SPSS Statistics 20. Throughout the analysis a
confidence level of 95% is used.
4.1 Descriptive statistics and preliminary data analysis 4.1.1 Socio-demographic profile
Women comprised 69% of the sample. 60% of the sample consisted of respondents of the age
group 18 to 30 years old. The second largest group was the age group between 51 and 70
(30%). Nearly the entire sample (97%) holds a professional degree or higher of which 44%
hold a master’s degree or higher. Regarding the employment status the sample was mostly
comprised of people who are either employed for wages (41%) or student (34%). Considering
the age groups, this spread was to be expected. Also the household composition (unmarried
63%) and the household income (below average 50%, average 13%, above average 29%)
correspond with this. This profile of the sample will have consequences for the validity of the
results which will be discussed in the final part of this chapter, discussing the limitations.
4.1.2 Audience appeal
The data representing the audience appeal towards the specific operas presented in the
questionnaire consisted of 15 variables in the form of a 5-point Likert scale. After running a
frequency analysis 5 missing cases became apparent. Since the missing data were less than
10% of the entire sample, the data are discrete and there were related variables to be used as
possible ‘donors’ I decided on hot-deck imputation as the best method to deal with the
31
missing data. Listwise deletion has the disadvantage of decreasing the sample size, pairwise
can bias estimates and lead to mathematically inconsistent results. Also, mean substitution can
flatten the distribution of the variables. Hot-deck imputation deals with missing values by
replacing them with the values of a ‘donor’ in researcher-determined categories. In this case
the missing cases were located in Q3 and Q4. In respect to the other questions, they both have
the strongest correlation (0.682 and 0.650, sig. ,001) with Q7. Also, Q7 has no missing data
and is a related variable. Therefore Q7 was used as deck. After running the hot-deck
imputation, another frequency analysis confirmed there were no missing data left.
4.1.3 Attendance goals
To be able to test hypothesis 4 the constructs proposed by Cuadrado and Mollà (2000) needed
to be tested for reliability. Beforehand, since the data consist of rankings of 15 items, I
recoded these values by inverting the score (15=1, 14=2, 13=3 etc.). To get a view of the
correlations within the variables a correlation matrix was constructed of the means, standard
deviation and the Kendall’s tau correlations. Kendall’s tau was used due to the question
format, which resulted in rank variables (table 5). The correlation matrix shows few
significant positive correlations and a handful of significant negative correlations, which
indicate the reliability analysis of the constructs will most likely turn out to be non-significant.
However, since the reliability of the constructs will determine if hypothesis 4 is testable, a
reliability analysis was conducted to determine to what degree the constructs are applicable to
32 Variables M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. To dress up 3,9494 3,29685 - 2. A way to spend an evening 8,7595 3,66975 ,174* - 3. Social prestige 3,7089 2,51739 ,125 ,158 - 4. A way to spend
time with friends
8,1899 2,69889 -,036 ,062 -,022 - 5. Social interaction
with fellow opera enthusiasts
3,6709 2,31894 -,130 -,134 -,133 ,086 - 6. To be able to
participate in future conversations about the opera
3,7595 2,94486 -,258** -,068 -,116 -,137 -,214* - 7. To feel emotion 11,6456 3,12983 -,015 -,184* ,043 -,110 ,152 -,214* - 8. Relaxation 9,7089 3,03464 -,115 -,110 -,107 -,099 -,130 -,064 ,013 - 9. Entertainment 9,7468 3,87122 -,006 ,229** -,054 -,139 -,075 ,004 -,060 ,222** - 10. Educational development 12,0633 2,68109 ,069 -,077 ,035 ,063 ,036 -,003 -,084 -,346** -,260** - 11. Self-fulfillment 10,8228 3,11226 -,036 -,208* ,054 ,095 -,128 ,141 -,065 -,135 -,267** ,311** - 12. Interest in the arts in general 11,8354 2,76160 ,002 ,080 -,069 -,031 -,025 ,093 -,197* ,010 -,062 -,051 -,031 - 13. To share an experience 7,6203 3,46523 -,021 -,200* -,057 -,016 -,115 ,196* -,077 -,043 -,062 -,063 ,015 ,008 - 14. To see particular artists 8,0506 3,71034 -,236** -,285** -,137 -,078 ,142 -,120 ,090 ,021 -,301** -,021 -,040 -,097 -,105 - 15. To see a director's work 6,4684 3,53675 -,203* -,139 -,135 -,144 ,032 -,105 ,083 ,054 -,165* -,029 -,151 -,194* -,117 ,491** - Correlation is significant at the 0.05 level (2-tailed).
Correlation is significant at the 0.01 level (2-tailed).
Table 5. Correlation matrix. Attendance motivations (Cuadrado & Mollà, 2000).
The first construct, social hedonism (SH), consists of the items social interaction with
fellow opera enthusiasts (SH_Fellow), a way to spend time with friends (SH_Friends), a way to spend an evening (SH_Evening), to dress up (SH_Dress), social prestige (SH_Prestige)
and to be able to participate in future conversations about the opera (SH_Future). A
reliability analysis shows the construct to be unreliable (cronbach’s Alpha -.159) due to
33
items its shows several significant negative correlations which leads us to conclude the
construct, as it is now, is not applicable to the sample. Cronbach’s Alpha was improved to
.378 by excluding SH_Future, SH_Friends and SH_Fellow. However, also the exclusion of
these items did not lead to an acceptable reliability value.
The second construct, emotions (Emo), consists of the items entertainment
(EMO_Enter), relaxation (EMO_Relax) and to feel emotion (EMO_Emo). A reliability
analysis showed the construct not to be reliable (cronbach’s alpha 0.243). Excluding
EMO_Emo could improve the value to 0.496, which is still insufficient to consider the
construct to be applicable to this sample.
The third construct, cultural fulfillment (CF), consist of the items educational
development (CF_Edu), self-fulfillment (CF_Self), to share an experience (CF_Share) and interest in the arts in general (CF_Interest). A reliability analysis resulted in a cronbach’s
alpha of 0,068 which is insufficient to consider the construct to be reliable. Excluding
CF_Share and CF_Interest led to a value of 0.497, which is still considered too low.
The final construct, interest (I), consists of the items to see particular artists (I_Artist)
and to see a director’s work (I_Director). A reliability analysis gave a cronbach’s alpha of
0.749 which indicates good reliability of this construct.
These reliability analyses will have significant consequences for the testability of
hypothesis 4, which will be further addressed in the paragraph addressing hypothesis 4 and in
34
4.2 Hypothesis 1: Linear Regression Analysis
Hypothesis 1: Operas that score low on form innovation, create a stronger appeal with
potential audiences than operas that score high on form innovation.
To detect if there is a relation between audience appeal and the degree of conformity over the
entire sample a linear regression analysis was conducted. The level of conformity
(DiMaggio-Stenberg Index of Conformity) functions as the independent variable, the audience appeal (Q3
to Q17) as the dependent variable. The independent variable consists of the mean values of
the audience appeal per opera.
The SPSS output (table 6) shows us that there exists a strong positive correlation
between audience appeal and degree of conformity (0.793, sig. 0.01). The model summary
confirms the goodness of fit with a coefficient of determination (R2) of 0.629 and an adjusted
R2 value of 0.600. This means that 60% of the variance in audience appeal is explained by the
model. An analysis of variance shows no problems with heterogeneity of variance (F =
21.995, sig. 0.001). The coefficients show (table 7), as the correlation analysis also computed,
a standardized β of 0,793 with a significance level of 0.01 based on the t-value (4.690). The unstandardized β is 0.251. These values show that when the independent value increases with one unit, the dependent value will increase by 0.251, which is an increase of 5.02%
(0.251/5*100).
a. Predictors: (Constant), INDEX_Title b. Dependent Variable: Mean_Appeal
Table 6: Linear regression analysis. Hypothesis 1.
Model Summaryb
Model R R Square Adjusted R Square
Std. Error of the Estimate 1 ,793a ,629 ,600 ,18402
35
Figure 2. Linear regression analysis. Hypothesis 1 Figure 3. Linear regression analysis. Hypothesis 1
The histogram (table 2) shows a strong normal distribution, especially considering the
small sample size. As show in the P-P Plot (figure 3), the values are shouldering the
regression line. The residual statistics hold values of 0.000 for the residual and standardized
residuals. The scatter plot of the regression standardized residual against the regression
standardized predicted value show one possible outlier. However, further investigation
through plotting the unstandardized residual against the outcome variable shows this is not the
case. Therefore the results can be considered significant and hypothesis 1 confirmed.
Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 2,462 ,059 42,085 ,000 INDEX_Title ,251 ,054 ,793 4,690 ,000
36
4.3 Hypothesis 2: Principle Factor Analysis
Hypothesis 2: The audience appeal can be categorized into two groups that are in
direct correlation with the degree of conformity of the operas.
Hypothesis 2 will be tested through a principle factor analysis and confirmatory statistics for
the applicability of this method (Kaiser-Meyer measure of sampling adequacy and Bartlett’s
test for sphericity). There are separate groups uncovered in the sample based upon the
variables Q3 to Q17. Though dimension reduction, by looking for factors with significantly
different outcomes regarding the appeal towards specific opera’s, I uncovered two distinct
groups of operas which are rated significantly different from one another (table 8). After
running the factor analysis the KMO and Bartlett’s test for sphericity showed sufficient
support for this method. The Barlett’s test is significant (0.000), which was to be expected
since the correlation matrix already showed strong correlations between several variables. The
KMO is also high at 0.887, which is positive for the applicability of the principle factor
analysis. The outputs also show two uncovered groups that evoke different tendencies which
explain 70,539% of the total variance.
The pattern matrix provides information on the division of the variables among these
two groups. It shows one group of 4 items, opposing another group of 14 items. To confirm
the reliability of these constructs a reliability analysis was conducted which led to a
Cronbach’s alpha of 0.949 for the first factor and a value of 0.863 for the second factor. These
values show a strong degree of reliability. The next step is to prove that the factors are
significantly different regarding conformity scores through a one-sample t-test. At first glance
37
results show the same. The results of this t-test (table 9-10) show significant (p<0.05)
difference based on the score of the grouped operas on the DiMaggio-Stenberg index of
conformity. Based upon these results, hypothesis 2 is confirmed.
Table 8. Principle factor analysis. Hypothesis 2.
One-Sample Statistics
N Mean Std. Deviation Std. Error Mean
INDEX_Factor_1 11 ,1107 ,12043 ,03631
INDEX_Factor_2 4 2,0870 ,26327 ,13164
Table 9. One-Sample T-Test. Hypothesis 2.
One-Sample Test
Test Value = 0
t df Sig. (2-tailed) Mean Difference 95% Confidence Interval of the Difference
Lower Upper
INDEX_Factor_1 3,048 10 ,012 ,11067 ,0298 ,1916
INDEX_Factor_2 15,854 3 ,001 2,08696 1,6680 2,5059
Table 10. One-Sample T-Test. Hypothesis 2.
Factor 1 (Cronbach’s alpha: 0.949)
Title Index
Factor 2 (Cronbach’s alpha: 0.863)
Title Index Q5_Lulu 0,434782609 Q3_Boheme 2,217391304 Q6_Agrippina 0,217391304 Q4_Figaro 2,260869565 Q8_Gambler 0,130434783 Q7_Barbiere 2,173913043 Q9_Puritani 0,086956522 Q16_Carmen 1,695652174 Q10_Retable 0,086956522 Q11_Aschemond 0,043478261 Q12_Fee 0,043478261 Q13_Gesei 0,043478261 Q14_Mamelles 0,043478261 Q15_Poliuto 0,043478261 Q17_Zemire 0,043478261
38
4.4 Hypothesis 3: Cluster analysis
Hypothesis 3: Within potential opera audience there exist distinct groups based on
their appeal towards innovation in form.
To uncover clusters or ‘clicks’ of respondents that
respond similar to the same operas, a cluster
analysis was performed on Q3 to Q17 for all the 79
cases. The analysis uncovered three different
clusters of fair cohesion among the respondents.
The first cluster comprised 45,6% (n=36) of the
total sample. The second cluster comprised 45,6%
(n=28) and the third of 19,0% (n=15) of the total
sample. Table 11 shows the relative influence the
answers of Q3 to Q17 have on the clustering of the
sample. This table shows that the operas that are
considered innovative to the measure of conformity
in repertoire by DiMaggio and Stenberg are of
greater influence than the operas with a high degree
of conformity. This suggests distinct correlations
between the three groups and the
DiMaggio-Stenberg scores. To identify the specific cases that
belong to the three different clusters, during the
analysis of the clusters a cluster membership variable Table 11. Relative influenceof items.
39
After separating the three clusters, a linear regression analysis was conducted three
times, once for each one of the three clusters as dependent variable and the transposed mean
DiMaggio-Stenberg scores as an independent variable (the mean scores of each cluster per
opera). The results for the first cluster (n = 28) are non-significant (R = 0.779, sig. 0.06) with
an adjusted R2 of 0.475. The results for the second cluster (n = 36) show a positive and
significant correlation (R = 0.877, sig. 0.025) with an adjusted R2 of 0.692, but a
non-significant standardized β of 0.877 (F = 10.000, sig. 0.051). The results for the third cluster (n = 15) also show a positive and significant correlation (R = 0.861, sig. 0.030) with an adjusted
R2 of 0.655, but also a non-significant standardized β of 0.861 (F = 8.603, sig. 0.61). The residual statistics for all three clusters hold values of 0.000 for the residual and standardized
residual. The correlation between the conformity scores and cluster 2 and 3 are therefore
strong, but the regression analysis is non-significant. Comparing the significant correlation
results, cluster 2 and 3 do not differ much as to their coefficient towards the independent
variable. However, looking at the intercept of the two models (table 13-14) we find a value of
3.428 for the second cluster and a value of 1.346 for the third cluster. The first cluster,
although non-significant, has an intercept of 2.528. Also the descriptive statistics of the mean
variables of the separated clusters show difference in mean value (table 12). This indicates the
differences among the clusters is possibly based on the intercept, rather than on the
coefficient.
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation Mean_Cluster_1 28 2,00 3,07 2,3592 ,34793 Mean_Cluster_2 36 2,87 5,00 3,3398 ,44628 Mean_Cluster_3 15 1,00 2,27 1,3930 ,43095 Valid N (listwise) 15
