Brand Loyalty: A Study of the Prevalent

Usage of Celebrity Endorsement in

Cosmetics Advertising.

Authors:

Emma Lfgren
Juan Li

Supervisor:

Margareta Paulsson

Student
Ume School of Business
Spring semester 2010
Master thesis, one-year, 15 hp

ACKNOWLEDGEMENTS

Conducting this thesis has been a great learning experience.

Over the past ten weeks, we have received great support
from many people at Ume School of Business, Sweden.
We would like to thank everyone who has helped us to
make and complete this interesting journey. Family, friends
and primarily the respondents have contributed with
valuable additions and made it possible for us to complete
this study feeling proud of ourselves.
We would like to express our gratitude to our supervisor,
Margareta Paulsson, whose encouragement, guidance and
important feedback from the initial to the final step enabled
us to present this final work.

Thank You,
Emma Lfgren
Juan Li
Ume, May 26th 2010

SUMMARY
The usage of celebrity endorsements has been confirmed to result in more favorable
advertisement ratings and positive product evolutions (Dean and Biswas, 2001). It has even
become one of the most popular forms of retail advertising (Choi and Rifon, 2007). Marketers
have heavily relied on celebrity endorsement, because they believe in its positive impact of
assisting in improvement of brand awareness, brand equity, and even financial returns.
Celebrities are easily chosen by marketers to peddle their products. Especially in cosmetics
industry, this tactic is commonplace and often seen in TV advertisements, magazines and other
types of media. If you flip through the television channels, you can easily find a number of
different celebrity endorsers; from Eva Longoria for LOreal to Julia Roberts for Lancme and
even Nicole Kidman for Schweppes. What do marketers expect from enormous amounts spent
on celebrities appearing in their advertisements? Besides returns on the financial aspect and
brand awareness, using celebrities in advertisements can be a shortcut to assisting brands in
gaining recognitions. All these efforts may be contributions in creating brand loyalty, which is
where the marketing battle really begins. The brand loyalty concept has been somewhat
overlooked, in the sense that many studies have only been focused on the behavioral aspect.
Recent studies actively propose that brand loyalty shall take both behavioral and attitudinal
aspects into consideration in order to make it comprehensive. David Aaker (1991) suggested,
The brand loyalty of the customer base is often the core of a brands equity. Because of the
existence of brand loyalty, it makes the brand choice and brand equity possible. For these
particular reasons and arguments, the topic was chosen regarding both celebrity endorsement and
brand loyalty in cosmetics advertising, with the aim to probe the relationship between them. This
study concerns celebrity endorsement in cosmetics advertisements and the creation of brand
loyalty, thus investigating the use of celebrities in advertising and the effect it has and if and how
it contributes to brand loyalty. It is based on customers perception of this type of advertising
technique in relation to the cosmetics industry, and two different types of brand loyalty,
attitudinal loyalty and behavioral loyalty.
The literature and theories used for investigating and supporting this area mainly focused on the
two most important concepts for this study, celebrity endorsement and brand loyalty. In order to
review these two theories and support the research question, additional concepts concerning
endorsement, attitudes, behavior, brands and products were investigated and evaluated. These in
turn enabled the creation of a questionnaire, covering the area to be scrutinized.
For the purpose of this quantitative study, a self completion web based questionnaire was carried
out. The survey was sent by email to possible respondents, with a students email account at
Ume School of Business. The sampling method used when sending out the web survey was of a
non-probability nature.
The findings led to a conclusion confirming that neither brand loyalty nor attitudinal loyalty is
proved to be created by famous endorsers. Although, not bearing brand loyalty in mind, it can be
seen that the endorsers help customers to get to know a brand better. The data findings also
suggest that the costs of using celebrity endorsement are not reasonable in relation to gaining
respondents attitudinal loyalty. It is positively confirmed that both behavioural and attitudinal
loyalty should be included when measuring actual brand loyalty concerning cosmetics brands.

Table of Contents
CHAPTER 1 INTRODUCTION .......................................................................................................... 1
1.1 Background and Research Problem ................................................................................................ 1
1.2 Knowledge Gap ............................................................................................................................. 4
1.3 Research Question ......................................................................................................................... 5
1.4 Purpose.......................................................................................................................................... 5
1.5 Delimitations ................................................................................................................................. 5
CHAPTER 2 THEORETICAL METHODOLOGY ............................................................................ 7
2.1 Choice of Subject........................................................................................................................... 7
2.2 Perspectives ................................................................................................................................... 7
2.3 Scientific Approaches .................................................................................................................... 8
2.3.1 Epistemology .......................................................................................................................... 8
2.3.2 Ontology ................................................................................................................................. 8
2.4 Research Strategies ........................................................................................................................ 9
2.4.1 Research Reasoning Strategy................................................................................................... 9
2.4.2 Research Design Strategy ...................................................................................................... 10
2.4.3 Data Collection Strategy........................................................................................................ 10
2.4.4 Critique Discussion of Data Collection .................................................................................. 11
2.5 Preconception .............................................................................................................................. 12
2.6 Choice of Theories ....................................................................................................................... 12
2.7 Literature Search.......................................................................................................................... 13
CHAPTER 3 THEORETICAL FRAMEWORK ............................................................................... 15
3.1 Brief of the Literature Review ...................................................................................................... 15
3.2 Celebrity Endorsement................................................................................................................. 15
3.2.1 Meaning Transfer Process ..................................................................................................... 15
3.2.2 Reference Group Theory ....................................................................................................... 18
3.2.3 Endorsement by Product Types ............................................................................................. 19
3.2.4 Attitude Change Through Visual Imagery ............................................................................. 19
3.2.5 The Social Adaptation Theory ............................................................................................... 21
3.2.6 The Elaboration Likelihood Model Theory ............................................................................ 22
3.3 Brand Loyalty .............................................................................................................................. 23
3.3.1 Behavioral Loyalty ................................................................................................................ 25
3.3.2 Attitudinal Loyalty ................................................................................................................ 25
3.3.3 Brand Affect and Brand Trust ............................................................................................... 27

3.3.4 The Relationship Between Two Loyalties .............................................................................. 28

3.3.5 The Attribute Frequency Concept .......................................................................................... 28
3.3.6 The ABC Model of Attitudes................................................................................................. 29
3.4 Conclusion................................................................................................................................... 30
CHAPTER 4 PRACTICAL METHODOLOGY ................................................................................ 33
4.1 Research Survey .......................................................................................................................... 33
4.2 The Survey Design ...................................................................................................................... 33
4.3 Sampling ..................................................................................................................................... 35
4.4 Reliability .................................................................................................................................... 37
4.5 Validity ....................................................................................................................................... 37
4.6 Statistical Tools ........................................................................................................................... 38
4.7 Limitations .................................................................................................................................. 38
CHAPTER 5 RESEARCH FINDING AND DATA ANALYSIS ....................................................... 40
5.1 Demographical Data Analysis ...................................................................................................... 40
5.1.1 Age Groups ........................................................................................................................... 40
5.1.2 Types of Cosmetics Purchase ................................................................................................ 41
5.1.3 Purchase Frequency .............................................................................................................. 41
5.1.4 Spending ............................................................................................................................... 42
5.1.5 Sources ................................................................................................................................. 43
5.2 Reliability Test ............................................................................................................................ 43
5.3 ANOVA Analysis ........................................................................................................................ 44
5.4 Correlation Analysis .................................................................................................................... 47
5.5 Descriptive Analysis .................................................................................................................... 49
5.6 Conclusion................................................................................................................................... 52
CHAPTER 6 DISCUSSION AND FURTHER RESEARCH............................................................. 54
6.1 Discussion ................................................................................................................................... 54
6.2 Managerial Recommendations ..................................................................................................... 56
6.3 Contribution ................................................................................................................................ 57
6.4 Limitations and Further Research ................................................................................................. 57
REFERENCES .................................................................................................................................... 59
APPENDICES ..................................................................................................................................... 62
Appendix 1-Survey in English ........................................................................................................... 62
Appendix 2-Survey in Swedish .......................................................................................................... 65

Appendix 3-Survey Coding ............................................................................................................... 68

Appendix 4- The email of respondent recruitment .............................................................................. 70
Appendix 5-Raw Data Set ................................................................................................................. 71

List of Tables
Table 1 Cronbachs Alpha Test.............................................................................................................. 44
Table 2 Dependent Variable: Behavioral Loyalty ................................................................................... 44
Table 3 Dependent Variable: Behavioral Loyalty ................................................................................... 45
Table 4 Dependent Variable: Attitudinal Loyalty ................................................................................... 46
Table 5 Dependent Variable: Attitudinal Loyalty ................................................................................... 46
Table 6 Brand Endorsers and Behavioral Loyalty .................................................................................. 47
Table 7 Famous Endorsers and Behavioral Loyalty................................................................................ 48
Table 8 Brand Endorsers and Attitudinal Loyalty................................................................................... 48
Table 9 Famous Endorsers and Attitudinal Loyalty ................................................................................ 49

List of Figures
Figure 1 Meaning Transfer Process........................................................................................................ 16
Figure 2 Visual and Verbal Loop Possibilities ....................................................................................... 20
Figure 3 The Elaboration Likelihood Model (ELM) of Persuasion ......................................................... 22
Figure 4 Consumer Brand Loyalty Classification ................................................................................... 24
Figure 5 Brand Loyalty Model............................................................................................................... 26
Figure 6 ABC Attitude Model ............................................................................................................... 30
Figure 7 Theoretical Framework ............................................................................................................ 31
Figure 8 Age Groups of the Respondents ............................................................................................... 40
Figure 9 Types of Cosmetics Often Purchased by Respondents .............................................................. 41
Figure 10 Respondents Purchase Frequency ......................................................................................... 41
Figure 11 Spending on Each Purchase ................................................................................................... 42
Figure 12 Sources of Getting to Know Cosmetic Brands ........................................................................ 43
Figure 13 The Means of Attitudes about Cosmetics Advertisements ...................................................... 49
Figure 14 The Means of Important Factors Influence Cosmetics Purchasing Decisions .......................... 50
Figure 15 The Means of Behavioral Loyalty Variables .......................................................................... 51
Figure 16 The Means of Attitudinal Loyalty Varaiables ......................................................................... 52

Shoot for the moon. Even if

you miss you will land among
the stars.
- Les Brown Sr.

CHAPTER 1 INTRODUCTION
The purpose of this chapter is to present the background to the research problem and the aim of
the study. It shall be the first step for the entire research. Furthermore, the intention is to provide
readers with a clear understanding of why the authors have decided to investigate this specific
area and how the topic is related to previous research.

1.1 Background and Research Problem

In recent years, the plethora of use of celebrities in advertisements has become more prevalent
than ever. Marketers have the perception that this technique of persuasion is a winning formula
to build up brand image, to increase sales revenue, and to gain strong brand loyalty. We are
bombarded by a variety of different advertisements in our everyday life without having a choice
not to. Celebrity endorsement is very popular and widely used in advertising. From magazines to
TV advertisements, from toothpaste to luxury goods, celebrity-endorsed products have
penetrated almost every aspect in our life. Famous endorsers appear in approximately 25 percent
of all kinds of television commercials to promote different products and brands, and roughly ten
percent of advertisers budgets are spent on the use of celebrities (White, Goddard, & Wilbur,
2009). This phenomenon displays the popularity and importance of the practice on
communicating brand images, where the aim is to communicate the message to consumers in a
suitable manner in order to achieve best results possible.
According to Grants definition, a celebrity endorser can be described as follows:
Any individual who enjoys public recognition and who uses this recognition on behalf of a
consumer good by appearing with it in an advertisement (McCracken, 1989, p. 310).
Strictly defining a celebrity, it refers to a movie, TV star, or a singer. But in broader meanings,
any individual who is successful in the world of sports, business, art, politics, and the military
can be attributed into the group of celebrities. (McCracken, 1989, p. 310) The celebrities
themselves transfer their own personality and image to the brand and product they are
representing, generating a related image between themselves and the brand or product. In order
for the endorsement to be successful, there are three different attributes the celebrity needs to
communicate to the consumers. These are; credibility, attractiveness and power, where
credibility is represented by knowledge, skills and expertise concerning the product/s,
attractiveness includes likeability, familiarity and similarity, and power enables the endorser to
affect the consumer (Byrne, Whitehead & Breen, 2003 p.291).
The most important aspect when practicing celebrity endorsement is the choice of celebrity, to
see that the image and personality projected on the brand and its existing image is suitable and
enhancing. The impact on and the outcome of the endorsement have to be well evaluated before,
during and after it takes place, with the intention of generating customers and creating brand
loyalty. In addition to likeability and recognition, celebrity endorsement can create new segments
of customers with a connection of some sort to the celebrity but with no previous attitude or
preference towards the brand (Byrne et al., 2003 pp.292-293).

Page | 1

Celebrity endorsement is frequently used in the promotion of cosmetic products. By using

celebrities as endorsers in advertising for this specific type of products, the company in question
communicates the talent, beauty and perfection that are often associated with famous
actresses/actors, artists, etc. This is something consumers can relate to in the sense that they can
buy the particular product and thereby achieve a connection to and find similarities with the
endorsing celebrity. Customers attitudes regarding brands and products are well enhanced by
celebrities, but whether it generates repeat purchase intention and brand loyalty is not so clear
(Byrne & Whitehead 2003 p.292).
The use of celebrity endorsement is not a new innovation in the advertising industry, an industry
involving massive money investments. Marketers do certainly expect returns from their
investments, not only by increasing sales revenues but by something even more important in the
long run- brand loyalty. Till (1998) mentioned that celebrity endorsement can provide a great
value in building strong brand equity, but also improving a brands market position if the
endorsers are used wisely. This marketing strategy is a swift and valuable way to build up instant
brand recognition. It saves marketers a lot of time when trying to attain consumers awareness
and draw attention to the brand. This results in the marketers ultimate objective from all efforts
put into the campaign, to be attaining brand loyalty.
Why is brand loyalty so important to marketers? A variety of instruments are used in order to
capture consumers attention, interests and loyalty. As Jack Trout (2006) stated, Marketing
battles takes place in the mind of a consumer or prospect. That is where you win. That is where
you lose. Take Coca-Cola as an example. On June 10th, 1985, the Coca-Cola Company
announced they were changing their 99-year-old formula. This abrupt behavior made it one of
the biggest risks within the consumer goods area. The company did not foresee the firestorm
coming, which ended up with returning to their original formula.
The events of 1985 changed forever the dynamics of the soft-drink industry and the success of The Coca-Cola
Company, as the Coca-Cola brand soared to new heights and consumers continued to remember the love they have
for Coca-Cola (The Coca-Cola Company).
This historical event makes marketers and companies wonder; how can companies generate
brand loyalty and how to keep it? Brand loyalty has the mystery power of attaining consumers
over and over again. At the beginning, consumers might buy a specific brand because of
objective reasons. However, after a long time of being advertised and exposed on the market, the
brand might become part of the memory and also part of the consumers life. Consumers might
associate the brand with many different and important occasions throughout time (Fisher, 1985).
Therefore, brand loyalty makes consumers deeply committed to a specific brand. This is
something marketers are constantly searching for by running a vast amount of different
marketing campaigns.
This leads us to the question of what brand loyalty really is and how it is constituted. Brand
loyalty has always been the primary focus of the behavioral aspect. Many researchers have
suggested that brand loyalty is a conscious reflection of a repetitive purchasing behavior of the
same brand. However, a supplement to the definition of brand loyalty has been brought up in
recent years. Jacoby and Kyner (1973) first brought up brand loyalty as consisting of more than
Page | 2

just a physical purchase. In their experimental research, six essential traits were included in order
to accurately define loyalty:
Bias
Behavioral response
Expressed over time
By some decision-making unit, with respect to one or more alternative brands
A function of inertia
A function of a psychological (decision-making, evaluative) process
This supplement to the original thoughts concerning brand loyalty clearly points out that
attitude should be included along with behavior to define loyalty (Bandyopadhyay & Martell,
2007). Dick and Basu (1994) further emphasized that loyalty should consist of a favorable
attitude in addition to a repeat purchase behavior. Bandyopadhyay and Martell (2007) argues in a
recent research paper that they have clearly redefined the meaning of brand loyalty, based on
reviews and summaries from other researchers previous studies. They suggested that true
loyalty should consist of; (1) a favorable repeat purchase (behavioral) pattern, and (2) a
favorable disposition (attitude) towards the brand, thereby including both attitude and behavior
(Bandyopadhyay & Martell, 2007).
There is no doubt celebrity endorsers influence consumers purchasing decision. If there is an
inherent fit or consistency between a celebrity and a product, the endorsement is in fact perfectly
matched with the purpose of celebrity endorsement (Till & Busler, 1998). However, there is no
evidence proving that the usage of celebrity endorsement will achieve stronger brand loyalty in
comparison to a non-use of celebrity endorsement. In fact, consumers might pay more attention
to celebrities in advertisements than the actual product endorsed, which is not the marketers
intention. As in most cosmetics advertisements, physically attractive celebrities are especially
recognized. They might exert a magnetic power to consumers resulting in a behavioral purchase
instead of true brand commitment. But whether or not the purchase behavior will become
repetitive is less clear (Byrne & Whitehead, 2003). Based on the advertisement for Edge
disposable razors, a laboratory study found that the status of celebrity endorsers was, to many
consumers, not related to an assessment of the true value of the product. The celebrity endorsers
still served as a peripheral cue because they were adored more than ordinary people (Kahle &
Homer, 1985, p. 954). However, a simple behavioral response cannot be considered as real brand
loyalty, and the behavioral response might easily be wavered by other factors if consumers do
not have a strong commitment to the brand. Heavy use of celebrities in cosmetics advertisements
might create a paradox, because consumers cannot identify whether they like the celebrity in the
advertisement or if it is the actual product they like. The ultimate goal of marketers is making
consumers assert most of their attention to the products. But whether this goal can be
accomplished by using celebrities is unsure. Attractiveness remains important but the attraction
of the celebrity is not the best way to build a positive benefit for the brand (Till & Busler, 1998, p.
585).
Euromonitor (2009) pointed out in its report, The problem is that the celebrity beauty industry
has become over saturated by the large number of products that have inundated the market
(Prance, 2007). Cosmetics companies are putting themselves at risk as there are too many
celebrity endorsed products available to choose from for the consumers, causing confusion.
Page | 3

Based on the findings in a study, consumers are more willing to see anonymous spokespersons or
experts in advertisements, rather than the use of a celebrity endorser. In the test, 56 per cent of
volunteers were opting to buy a cosmetic product that was marketed by an unknown person
(Prance, 2007). The appropriate usage of celebrity endorsement should be supportive in building
brand loyalty but not function as a purpose of beautification. Moreover, celebrity endorsers
should not be used as an excuse of lazy marketing when marketers cannot think of better ways to
promote the products (Till & Busler, 1998, p. 586).
Celebrity endorsers often appear in cosmetics advertisements but not all of them fully match with
the products or brands. The fact that they are famous might often be the only reason for a
marketer to choose a celebrity as a spokesperson. In Kahl and Homers match-up hypothesis
(1985), they stated that the physical attractiveness of a celebrity endorser may only enhance both
product and ad-based evaluation if the products characteristics match-up with the image
conveyed by the celebrity. Further empirical studies made by Kamins (1990), based on the
match-up hypothesis, suggest that there is no evidence proving a general axiom in advertising
saying what is beautiful is good (Kamins, 1990, p. 4). That is a too simplistic of a judgment
for marketers and with no doubt important for celebrity marketing. Celebrities, as a person or as
a character, do convey a profound meaning in our life, but to many ordinary people, the
celebritys unreal attractiveness has fewer connections with the situation in reality. It is pleasant
to see celebrities appearing in a number of cosmetics advertisements, but perhaps not in every
one. The meaning of many cosmetics advertisements is to serve as a social approval message,
which merely encourages people to respond by adding intangible pressure on them, because the
endorsers look too perfect in advertisements. Is it really necessary for us, as human beings, to
look perfect? The original purpose of using celebrities in advertising is mainly for the support of
attaining brand loyalty.
Many prior researches have been done regarding celebrity endorsement and brand loyalty
separately. However, there is no connection between a celebrity endorser and its effectiveness in
creating brand loyalty as yet. We do not know whether the prevalent usage of celebrities as
endorsers in especial cosmetics advertising has an effect on creating brand loyalty or not. Even if
it is effective, what kind of loyalty can it create? As we have discussed earlier, brand loyalty
should not only be defined by behavioral loyalty, which should also include attitudinal loyalty.
Therefore, this research aims to explore the connection between using celebrities and its
effectiveness in order to create brand loyalty in cosmetics advertisements.

1.2 Knowledge Gap

In spite of a vast amount of previously published research papers dedicated to the field of
celebrity endorsement and brand loyalty separately, there are no studies which specifically focus
on identifying or establishing a relationship between the two thus far. Celebrity endorsement is a
conventional marketing instrument used by marketers to communicate brand images and
information to consumers. Brand loyalty is the main effect marketers expect from consumers as
the final goal of all marketing campaigns, including TV advertisement, magazines, radio, etc.
Many previous studies have provided solid theoretical foundations in the fields that are covered
in the study, which are all essential to the undertaking of this research paper. From an authors
point of view, this research paper might add some contributions into the field of branding and
marketing management. It might also help companies in the cosmetics industry to better grasp
Page | 4

consumers opinions concerning celebrity endorsement. Most importantly, this research can aid
marketers to identify a more efficient way to gain brand loyalty and thereby also to keep it.

1.3 Research Question

In line with the research problem and the knowledge gap, the main research questions are
generated:
How efficient is the use of celebrity endorsement in cosmetics advertisements for the
creation of brand loyalty?
What influence does celebrity endorsement in cosmetics advertisements have on brand
loyalty?
If celebrity endorser in cosmetics advertisement have no influence on brand loyalty, what
factors do consumers take into consideration when making their cosmetics purchase
decisions?

1.4 Purpose
Considering the massive amount of celebrity endorsement practices in cosmetics advertising, this
study is an effort to examine the link between celebrity endorsement and its effectiveness in
attaining consumers brand loyalty. Brand loyalty, as stated earlier, should take both behavior
and attitude into consideration. Therefore, the purpose is also to explore the effectiveness of the
use of celebrity endorsement in advertisements and its influence on consumers behavioral and
attitudinal loyalty. Although there is not a direct way to test the relationship between the
effectiveness of advertisements and consumer behavior, nor the effectiveness of advertisement
and consumers attitudes, this research applies to advertisement influence, behavioral loyalty
attributes and attitudinal loyalty attributes separately. By utilizing all previously gained
knowledge, the results shall be able to serve the research purpose.

1.5 Delimitations
It is necessary to define the boundaries of this research. To narrow down the scope of the
research, the study only focuses upon the phenomenon of celebrity endorsement in a general
view without taking cultural backgrounds into consideration. The foundation theories of celebrity
endorsement were constructed from a more cultural perspective. McCracken (1989) argued that
the necessity of considering cultural background is placed in the point of the symbolic properties
of the celebrity endorser used, unlike in many different cultures.
Furthermore, in this research we do not identify the existence of neither attitudes or behaviors of
consumers. It assumes that consumers already have previous cosmetics purchasing experience,
which may have been influenced by advertisements. However, this point is unknown when
conducting this study. Therefore, there is no evidence of how to identify whether it has been
caused by behavioral loyalty or attitudinal loyalty. This study has chosen to ignore the fact that
one of the different types of loyalty might already exist in consumers minds, or not exist at all.
In addition, brand awareness should be the first thing to be confirmed. This study assumes the
Page | 5

pre-existence of brand awareness, as it is a prerequisite in forming brand loyalty. This is due to

the fact that consumers face all sorts of media exposure in their everyday life. It is very unlikely
that consumers who have purchase experience regarding cosmetics have not previous knowledge
about any brands available on the market at all.
This research study confines the celebrity endorsers as female actresses, singers or models who
are physically attractive, and frequently appearing in many of the cosmetics advertisements.
Consumers of cosmetics products typically consist of female customers. Consequently, the
research survey is therefore restricted to female respondents, even though male consumers are
equally important to marketers. Furthermore, our respondents are mostly limited to students at
Ume University with several non-student respondents. A similar study concerning both males
and females could be interesting to marketers, but the product categories included might need to
be reconsidered.

Page | 6

CHAPTER 2 THEORETICAL METHODOLOGY

The purpose of this chapter is to explain the research design for the undertaking of this study.
Clearly defining and identifying appropriate research methods and data collection procedures
will maximally support the addressing of the research objectives, and also answer the research
questions.

2.1 Choice of Subject

Before moving into the earnest of the research, we would like to briefly present the
considerations of the choices that were made. The subject was chosen based on our interest and
knowledge previous to this study. The research aims to test the most likely relationship between
celebrity endorsement and brand loyalty in combination, so as to also discover the establishment
of the different types of brand loyalty. Furthermore, this research will focus on cosmetics
products, since many celebrity endorsements are associated with this product category. As
cosmetics are categorized in the consumable goods market, it is important for brands to build up
a strong brand loyalty in order to maintain existing customers and approach more potential
consumers. Considering all of the above mentioned, a topic within this field was chosen.

2.2 Perspectives
Different perspectives can very often be the case when conducting a research study. The
perspectives can be viewed from those who use cosmetics products and are aware of certain
cosmetics brands available on the market. From a marketers point of view, brand loyalty is a
commitment given by consumers, which enables the marketers to gain advantages in comparison
to their competitors within the industry. Brand loyalty allows marketers to create a salient
influence on consumers purchasing behavior. This influence might be intangible or tangible. As
time goes by, brand loyalty will become inertia in the consumers purchasing behavior. However,
many marketers might overlook consumers basic needs due to the market becoming
increasingly sophisticated. They rely heavily on celebrity marketing, resulting in the cosmetics
market becoming a celebrity-obsessed platform.
What expectations do consumers have on a brand? From the consumers perspective, an
excellent brand represents a distinction from the rest of the brands available on the market. When
they hold strong loyalty towards a brand or product, it saves their time in searching and
comparing alternatives while shopping, as well as it gives them confidence concerning the
quality of the product. Understanding brand loyalty is not easy. Gaining brand loyalty is even
more difficult as many brands are competing in the market. In this specific matter, this research
is attempting to consider the viewpoint of the customers towards cosmetics advertisements. We
have chosen to place the centre of attention on the customers in this study, since the purpose is to
examine the influence of celebrity endorsement on brand loyalty. This choice is based on the
importance of customers own perspective and preferences regarding cosmetics brands on the
market. Knowing what reasons leads a consumer to purchase a brand, and knowing what
attitudes consumers hold towards cosmetics advertisements might help marketers in
understanding their consumers. It will further support marketers in underpinning the brand
loyalty construction by using more effective marketing strategies and communication
instruments.
Page | 7

2.3 Scientific Approaches

2.3.1 Epistemology

Epistemology consists of different viewpoints within the field of study on what is considered to
be acceptable knowledge, where the positivistic and the hermeneutic (interpretivism) approach
will be investigated (Saunders, Lewis & Thornhill 2007, p.102). The positivistic approach is a
natural science epistemology and relates to research and findings that can only be considered as
facts through its nature. Thus, observable phenomena are the only means that can lead to creating
credible data within this stance. In addition to new data, positivism enables the creation of
hypotheses through the use of already existing theories (Saunders, Lewis & Thornhill 2007,
p.103: Bryman & Bell 2007, p.16). The focal point of the hermeneutic approach is not essentially
the data per se, but the understanding of the data. It stresses the existence of a difference between
natural sciences and social sciences (people and institutions) (Bryman & Bell 2007, p.17). Thus,
the choice of epistemology is based upon the interpretation of information.
Information in this specific study is regarded as the knowledge gained through the collection of
data by conducting questionnaires and experimental interviews. The intention is to be able to
distinguish a pattern within the data collected, which will thereby enable us to generalize the
findings throughout the population integrated in the study. Since the positivistic viewpoint
argues for the inability to change the essence of the data collected, which also is the fact in this
study, it is argued that this approach is the most suitable one. One common assumption exists
regarding this approach; the researcher of a positivistic study is not affected by or affects the
particular subject, but advocates independency. (Saunders, Lewis, & Thornhill, 2007, p. 103).
Based on the previously mentioned sources of knowledge and the fact that all the data will be
coded and interpreted in a statistical manner, generating absolute numbers, the nature of the
study is positivistic.
2.3.2 Ontology

Ontology is a concept related to different approaches on how to view reality (Bryman & Bell
2007 p.22). Two options are at hand in order to evaluate the best alternative for the specific
research problem, and thus result in the ability to interpret the best suitable line of action. These
options involve the matter of two different approaches, objectivism and constructionism. If the
social entities have an objective nature and act externally to social actors, they are considered
within the concept of objectivism, which argues for a social phenomena encounter with facts that
cannot be influenced or changed in any manner. On the other hand, if the social entities are of a
constructive nature, i.e. affected by actors perceptions and actions, they belong to the concept of
constructionism. It stresses the relevance of social actors in the explanation of social phenomena
(Saunders, Lewis, & Thornhill, 2007, p. 108). We have chosen an objectivistic ontological
approach for this study.
The research subject in this study can be considered very definite, in the sense that celebrity
endorsement can or cannot have an effect on attitudinal and behavioral brand loyalty. The
subject is studied without an emphasis on external actors, thus the respondents opinions and the
data gathered are interpreted by their pure nature. The actions leading to the respondents
opinions are of great interest, but the reasons for making a decision regarding it are not intended
Page | 8

to be found through this particular study. We find it very important to be objective in this study
since it is projected to be unbiased, which may not be the case when performing a study with a
subjectivistic approach, where there is room for personal interpretations.
Even though a choice was made regarding the approach within both epistemology as well as
ontology, it can still be argued that it is not strictly one way or the other. It must be
acknowledged that parts of the study can be somewhat influenced by a different viewpoint within
this scientific area, due to its content and nature.

2.4 Research Strategies

The main goal of a research strategy is to identify the best way to approach the research
questions for both the design and the data collection.
2.4.1 Research Reasoning Strategy

In order to comprehend our sophisticated world, we often employ three different types of
reasoning; deductive, inductive and the combined inductive-deductive approach. Though, two of
these types of reasoning, the deductive and inductive approach, are mainly used in most social
science researches.
Deductive reasoning evolved from Aristotles syllogism, which is the cornerstone in contribution
to formal logic. It underlines the syllogism as through a sequence of formal steps of logic, from
the general to the particular, a valid conclusion can be deduced from a valid premise (Cohen,
Manion, & Morrison, 2007, p. 6). In other words, deductive reasoning starts from a general point
of view. At the beginning, researchers collect the most relevant theories regarding a selected
topic. After narrowing down the theories into specific hypotheses, the researchers will further
gather observations or experiments in order to test them. Finally, the results through previous
procedures will enable researchers to draw a conclusion whether to accept or reject the
hypotheses derived from the theories used in the research (Study skills). However, it has a prime
limitation which can only deal with certain types of statements.
Francis Bacon was the first philosopher to criticize the use of a deductive approach, because of
its major premises were often preconceived notions which inevitably bias the conclusions.
(Cohen, Manion, & Morrison, 2007, p. 6) He further proposed inductive reasoning as means of
using a number of individual cases would lead to a hypothesis. This reasoning would ultimately
generate a new theory. Thus, inductive reasoning requires a confirmation based on observing
what actually happens. It begins with phenomenon or observations in a specific area, where
researchers are interested in finding out patterns to answer the reasons behind it. After
identifying patterns, the researchers will formulate possible hypotheses based on these. In the
end, all the previous procedures will help researchers to draw up a theory or a conclusion (Study
skills). The central gravity of inductive reasoning is that people are free to seek significant
relationships and laws with sufficient data, where a preconceived idea is not required (Cohen,
Manion, & Morrison, 2007, p. 6).
The two methods discussed are different concerning the way of conducting research. The
inductive process is more open-ended and exploratory, especially at the beginning. The
Page | 9

deductive processes are, however, more narrow in nature and need testing or hypotheses in order
to confirm the theories in the research. The latter approach applies to the specific research
problem investigated in this study. It is derived from a real life phenomenon, based on the
established research questions, and relevant theories are examined for further experiments or
hypotheses tests. At the end, conclusions will be drawn based on a data analysis of results in
order to confirm the theories selected, as the detailed background and problem have been
discussed in the previous chapter. The foundation of this thesis is based on existing research in
the fields of celebrity endorsement, its functions, behavioral loyalty and attitudinal loyalty. This
study departs from the three main areas and carries out a rather in-depth study beyond them. The
purpose of the study, as stated before, is merely to try to find out how effective celebrity
endorsement can be in approaching consumers brand loyalty. The process of this study, for that
reason, can be positioned as a deductive logic approach. By doing so, the ultimate goal is to
confirm that prevalent usage of celebrity endorsement in cosmetics advertisements might or
might not be the most effective strategy. Proper conclusions and suggestions can be found in
chapter 6.
2.4.2 Research Design Strategy

Generally speaking, social research work consists of two types of fundamental research;
descriptive research, concerning What is going on? and exploratory research asking Why is it
going on? Descriptive research has the characteristics of measuring and estimating the sizes,
quantities or frequencies. It merely presents variables without analyzing their relationships.
However, exploratory research is more popularly employed in qualitative research as its
characteristics are to generate ideas, insights or hypotheses. (Kent, 2007, pp. 17-18)
According to Kent (2007, p. 12), there is a common mixture in practice of exploration,
description, and investigation of the relationships between variables and a causal analysis.
Usually, the researchers will begin a study with an exploratory point, and in the main research
phase, descriptive data will be collected and followed by a continuous analysis in the nature of
the relationship between the variables. Each stage will shade into the next with exploratory
research becoming the basis for description, and description forming the first part of an
investigative analysis and so on (Kent, 2007, p12). Some research is only focused on generating
ideas that can be attributed as exploratory. Other research, which is limited to the analysis of one
variable at a time, can be addressed as descriptive. For most social research, there is always a
combination of exploratory and descriptive research (Kent, 2007, pp.12-13).
Since the aim of descriptive and exploratory research is simply to collect information related to
the research question, the descriptive approach is suitable for this study. However, investigative
research, which goes beyond just collecting data and information, is also applicable to this study.
Since it focuses on analyzing the relationships of variables, and studies the statistical association
or correlation of two or more variables. However, the investigative approach does not examine
the causal relationship between variables (Kent, 2007, p. 18). Hence, the use of a combination
between the two different approaches origins from the main purpose of the study; to get a better
understanding of whether celebrity endorsement in cosmetics advertisements is effective in
gaining brand loyalty.

Page | 10

2.4.3 Data Collection Strategy

Essentially, market research can be divided into two main types: qualitative data and quantitative
data. Qualitative data consists of words, phrases, text or image. This data is constructed by
researchers using a systematic record process. As qualitative data is based on words, phrases, and
images of all kind, researchers cannot predetermine categories and codes to address the result of
the observations and possible responses (Kent, 2007). Qualitative data procedures allow
researchers to develop their own theory from data collected through both deductive and inductive
approaches (Saunders, Lewis, & Thornhill, 2007, p. 480). With a difference from qualitative data,
quantitative data is derived from numerical records. The appropriate situation for constructing
quantitative data is when the object of the research is to measure and analyze relationships or
patterns of concepts, trends and other variables. The main consideration of quantitative data is to
lead a numerical data analysis from relevant theories with a predetermined set of categories or
codes on which to map responses or observations (Kent, 2007, p. 119).
Considering the research problems and objectives of this research, it would be an ideal way to
use both qualitative and quantitative data. However, due to a limited amount of time and
availability issues, we have chosen to use a process of quantitative data collection. Besides, the
research problem is to identify the linkage between celebrity endorsement in cosmetics
advertisements and the different brand loyalty types, and we intend to be able to detect a pattern
from our results. Therefore quantitative data is more appropriate to use in this case. The
quantitative data will be collected through a survey with a reasonable sample size. The
investigative research procedure will be carried out after the data collection is completed. We
consider using quantitative data as a way to approach our primary data sources.
Additionally, secondary sources are used in our study. This is essential to the research, especially
in the review of existing literatures regarding the research background and problems. All the
secondary sources refer to articles published on journals such as Journal of Consumer Research,
Journal of Advertising etc., academic references such as Consumer Behavior (Solomon,
Bamossy, Askegaard, & Hogg, 2006), Market Research, Approaches, Methods and Applications
in Europe (Kent, 2007), and other articles of all kinds found on the internet which are relevant to
this research.
2.4.4 Critique Discussion of Data Collection

Research data can generally be attained through primary sources or the aforementioned
secondary data sources. A primary source of data is items that are original to the problem under
study (Cohen, Manion &Morrison 2005, p.162). The main sources in this study that can be
considered to be primary are the data collected through the questionnaire, and also research
reports and magazines. Many articles contain statements referring to previously published
articles, and we therefore have to go back to this initial source in order to gain as much
trustworthiness as possible, without any misinterpretations. It is important to criticize and
evaluate this type of information, so as to see to its accuracy and credibility. The genuineness of
the data is established, through looking at the documents or results itself instead of its actual
statements (Cohen, Manion & Morrison 2005, p.162). In addition, accuracy of the data, when
referring to an article as a primary source, is mirrored by the credibility of the authors of the
article in question.
Page | 11

Secondary sources are used because of the absences of first-hand data, which are not directly
developed or collected to the undertaken researches. Although a secondary source is a
convenient alternative to compensate for the missing primary sources, it sometimes lacks
credibility and has limited value due to errors caused by information transmission among
researchers. Therefore, researchers should bear the responsibility when using secondary sources.
Even though secondary source is not prevalently advocated in rigorous academic researches, we
shall not minimize the importance of it. Plentiful instances have also been demonstrated the
significant contribution of the use of secondary sources in social science researches (Cohen,
Manion, & Morrison, 2007, p. 194).

2.5 Preconception
We both possess preconceptions of different kinds, which should be briefly mentioned in order
to accentuate knowledge and experiences gained previous to performing this study. Both of us
have knowledge concerning the subject of marketing, due to both bachelor and master level
courses at different universities. It is considered to be important when investigating the subject in
question since one needs to know the relationships of concepts and theories, how to connect
them and what to focus on. This, in relation to other subjects studied throughout several years is
by us perceived to be of utmost importance for the study, since it shapes knowledge and
perceptions. In addition to these educational preconceptions, we consider ourselves to have
sufficient international experience, which can be applied to performing a study within this area,
due to heritage and studies abroad. This aspect is significant in the sense that it is perceived to
provide a broader perspective and a different point of view when investigating the specific
subject.

2.6 Choice of Theories

In spite of numerous theories and sources that also could have contributed to our research, we
have chosen the set of theories and sources presented in the theoretical framework because of the
relevance and coherence to our topic. Not only have we tried to seek well-recognized and highlyrelated literatures to identify our study in a proper position, but also have we sought a best
connection to our topic. The choice we made is to some extent not easy to be understood, as it is
neither fully dedicated to celebrity endorsement nor brand loyalty. We have strived to pursue
theories and prior studies to best fit our research topic in a rigorous context based on our
knowledge and other points of view. Establishing such an association which has not yet been
found is not as easy as a single topic with prior ample studies. Not only curiosity made us choose
such a topic, but also the spirit of need-to-know has encouraged us to embark on this specific
research. A strong philosophy has kept echoing in our minds, that gaining knowledge is not all
about finding the right answer. It is also about the journey of finding the answers. Other than that,
we have tried to keep the chosen literatures and sources to the least contradicting and repetitive.
There are no guidelines telling us what theory sources should be chosen, but only our preunderstanding about the topic. This pre-understanding has been leading us to find appropriate
theories through prior researches to denote our attention, endeavoring to remain only pioneered
studies to be presented in our chosen field.

Page | 12

2.7 Literature Search

Information regarding the subject and previous studies made are an important source of
knowledge when conducting research. This information is collected with the intention of making
sure the study has not been previously performed, neither in a similar nor exact manner. This
particular study consists of information gathered from a number of different credible sources and
channels, including scientific journals and articles, databases, literature and documents. These
sources can be seen to be credible since it is referred to by reliable authors and revised
continuously.
Both secondary and primary sources are used in this study. This is essential to the research,
especially in the review of existing literatures regarding the research background and problems.
All the secondary sources refer to articles published on journals such as Journal of Consumer
Research, Journal of Advertising etc., academic references such as Consumer Behavior
(Solomon, Bamossy, Askegaard, & Hogg, 2006), Market Research, Approaches, Methods and
Applications in Europe (Kent, 2007), and other articles of all kinds found on the internet which
are relevant to this research.
Research data can generally be attained through primary sources or the aforementioned
secondary data sources. A primary source of data is items that are original to the problem under
study (Cohen, Manion & Morrison 2005, p.162). The main sources in this study that can be
considered to be primary are the data collected through the questionnaire, and also research
reports and magazines. Many articles contain statements referring to previously published
articles, and we therefore have to go back to this initial source in order to gain as much
trustworthiness as possible, without any misinterpretations. It is important to criticize and
evaluate this type of information, so as to see to its accuracy and credibility. The genuineness of
the data is established, through looking at the documents or results itself instead of its actual
statements (Cohen, Manion & Morrison 2005, p.162). In addition, accuracy of the data, when
referring to an article as a primary source, is mirrored by the credibility of the authors of the
article in question.
Secondary sources are used because of the absences of first-hand data, which are not directly
developed or collected to the undertaken researches. Although a secondary source is a
convenient alternative to compensate for the missing primary sources, it sometimes lacks
credibility and has limited value due to errors caused by information transmission among
researchers. Therefore, researchers should bear the responsibility when using secondary sources.
Even though secondary source is not prevalently advocated in rigorous academic researches, we
shall not minimize the importance of it. Plentiful instances have also been demonstrated the
significant contribution of the use of secondary sources in social science researches (Cohen,
Manion, & Morrison, 2007, p. 194)
Several different databases at the Ume University library website have been an invaluable
source, where a significant number of scientific articles and other material have been found and
thoroughly evaluated. The ones mainly used for information and article search in this study are
some of the most popularly used within the business context. Business Source Premier (EBSCO),
Emerald Fulltext, Academic Search Elite (EBSCO), Web of Science (ISI) and Elsevier
ScienceDirect all functioned as sources of information regarding the research subject, previously
Page | 13

conducted research and its relevant information, whilst the DIVA-portal also contains scientific
and student theses carried out and published from different universities.
A number of key words were used for attaining information and relevant articles to support the
study. Different keywords for different areas in the field were selected and browsed. The concept
of celebrity endorsement involved keywords like and similar to; celebrity endorsement,
reference groups, visual image, endorsement and attitude. Information regarding brand
loyalty was found through use of the following keywords; brand loyalty, repeat purchase,
behavioral loyalty and attitudinal loyalty. In addition to the mentioned sources of
information, course literature related to methodology and research was used in order to gain
deeper knowledge and understanding regarding these topics. We consider this literature to
possess credibility and trustworthiness since it is up to date and is written in the business
administration archetype.

Page | 14

CHAPTER 3 THEORETICAL FRAMEWORK

Understanding relevant theories regarding the chosen research topic is a must for both the
authors and readers. It is also a prerequisite to carry out an appropriate data collection. In this
chapter, two main topics: celebrity endorsement and brand loyalty will be discussed in detail.

3.1 Brief of the Literature Review

In order to provide the reader with a deeper insight regarding the relationship between the chosen
theories and literature supporting the research area, a brief description follows before proceeding
to the theoretical framework.
Celebrity endorsement is one out of two major factors in this study, leading to the natural
approach of investigating and providing knowledge regarding this area as an introductory stage
in this chapter. With the intention of providing a link to the second major aspect, brand loyalty,
this area will also be covered together with the first. Different existing theories related to the two
concepts will each and every one be studied and considered in order to find support and
explanations regarding the problem faced in the research question. The theories included in this
chapter were chosen due to their applicability to the considered area, in addition to disregarding
those found to be less contributing. Hence, a thorough analysis of different theories at hand was
completed and resulted in the authors choosing a number of different ones which provided a link
to each other.
The first concept, celebrity endorsement, is investigated from a number of different aspects
related to it. The intention is to highlight the core by dealing with the concept itself and its
meaning. From that, the linkage to contributing factors and resulting actions and outcomes is
penetrated. Celebrity endorsement is dependent upon and affects its surroundings by its existence.
Therefore, factors that the concept is influencing, such as celebrity endorsement meaning transfer,
reference groups, product type categories, social adaptation theory etc. is covered in this chapter.
The second concept will be covered in a similar manner, with an emphasis on the foundation of
brand loyalty itself and the two different types, attitudinal loyalty and behavioral loyalty. In
addition to the importance of covering this area, it has to be acknowledged that brand loyalty
does not happen by its pure nature, it is affected by external forces. Therefore, related factors
(brand affect, brand trust, the ABC model etc.) pointing at its influence and upcoming, are of
great importance when looking deeper into this area.

3.2 Celebrity Endorsement

3.2.1 Meaning Transfer Process

McCracken (1989) pointed out the celebrity endorsement process as a process of meaning
transfer. Symbolic properties belonging to an endorser have been transferred to a specific
product or brand, and then from the product to the consumers. This specific type of endorsement
is different from other general types of meaning transfer. Because of celebrities existing in our
society for meanings, it makes them valuable in the process of meaning transfer. This transfer
process consists of three stages, as shown below in Figure 1 (McCracken, 1989).

Page | 15

Figure 1 Meaning Transfer Process (McCracken, 1989)

Stage one: Because of endorsement, advertisements have access to a group of persons from the
culturally constituted world that is different from ordinary people. Endorsement also makes this
group of people in charge of detailed and important meanings. As celebrities enjoy a popular
social recognition and a unique reputation, they are able to deliver more comprehensive and
depth meaning compared to unknown endorsers. Celebrities are powerful in media, though the
meanings delivered are not difficult to find. Because of their special and unique persona,
outstanding from anonymous actors, they exert the meaning with unforgettable impression. Due
to their career specialty, every point of acting in a show, movie, or musical gives them the
possibility in contrast to a different group of objects, persons and context. All of this is also a
process of meaning transfer, that new properties will be resided within the celebrities. Hence,
when they appear in an advertisement, these properties will be transferred to that specific product
(McCracken, 1989).
Stage two: Theoretically, a celebrity endorser is chosen based on a desired meaning decided by
the marketers. The symbolic properties of the meaning and a product must be matched. Ideally,
marketing or advertising companies can decide what type of symbolic properties consumers are
looking for. They can then find a perfect-matched celebrity from a roster to represent this.
However, this in fact, is impossible. Marketers can only rely on a very general rendering of
what meanings are available to them (McCracken, 1989, p. 316).
Advertising agents should make identifications and deliver these meanings to a specific product
when the selection is finished. The advertisement should then comprise all the meanings needed
to be obtained from the celebrity, especially the salient meanings. Only the meaning intended to
be delivered should be captured. Unwanted meanings should be kept out of the evoked set by
adding it into advertisements with people, objects and contents which have the same meanings as
the celebrity. All these are salient message cues that help the target consumers to get what ever
meanings they expect from the celebrity in question. Advertisements mostly only transfer
celebrities to a new role instead of purely transferring meaning of the celebrities.

Page | 16

Important to notice, there must be a connection between the product and the celebrity endorser in
order to have a high similarity. This will ensure consumers can transfer the meaning delivered
into their reality because of existing similarities (McCracken, 1989).
Stage three: Explains the process of how celebrity endorsement makes the properties of a
product become the properties of consumers. It is the most complicated and difficult stage
compared to the previous two. In this material world, individuals are trying to explore and search
for products with practical meanings. Those meanings certainly are useful in helping the
individuals to furnish aspects of the self and the world. As the meanings existing everywhere in
this world, the individuals have constantly been putting effort on searching the meanings they
need. However, it is not enough for consumers to understand the meaning of an object by simply
owning it. And the meaning cannot enter the consumers mind and self automatically. They
must claim, exchange, care for, and use the consumer good in order to possess the meaning.
McCracken (1989, p. 317)mentioned that celebrities play a role in the final stage of meaning
transfer because they have created the self. Through different characteristics and contexts in
their acting career, celebrities have earned a lot of attention and reputations in the first stage of
the meaning transfer. The self created by celebrities is also attractive in the eyes of the
consumers. They have set an exemplary, desirable role model for consumers. Consumers build
up their aspects of the world and the self by moving the symbolic properties of consumer goods
into their lives, because those consumer goods are associated with the celebrities they desire. But
this does not mean that consumers only want to follow what celebrities have done today.
However, celebrities certainly give the consumer goods some sort of personality traits in the
meaning transfer process. Consumers therefore also want to process the personalities attached to
the products, because they appreciate for meaning and are keen to build up their own aspects of
the world and the self by having celebrities as references. In this object world, celebrities are just
a small group compared to everything else. They are outstanding from the rest and therefore are
super consumers of a kind. It is also important to know that individualism and alienation exert a
right to individuals of freedom to define their aspects of the self and the world. Also because of
these, celebrities play a key role of meaning transfer to consumers. However, the meaning
transfer would not happen automatically. It needs cooperation from the celebrity to the
advertisement director, as well as from the other people involved. If the ideas in advertisements
cannot be understood by consumers, the use of celebrities is useless. As consumers are the final
act of the transfer process, they should be able to find the connection between the celebrity
endorser and the product in order to complete the transfer process (McCracken, 1989).
The meaning transfer model of celebrity endorsement depicted how the transformation begins
from a celebrity, to a consumer as the final destination. This model not only explained what
functions a celebrity endorser holds in the endorsement practice, but it also compromises the
deficits of source credibility and source attractiveness models (McCracken, 1989), which is
highly relevant to this study due to the purpose of the study.

Page | 17

3.2.2 Reference Group Theory

A reference group is defined as An actual or imaginary individual or group conceived of

having significant relevance upon an individuals evaluations, aspirations or behavior
(C.Whan & V.Parker, 1977, p. 102).
Associated with the definition, three important motivational influences of a reference group need
to be mentioned, these being; informational, value expressive, and utilitarian.
Informational reference group influence: each individual tends to search products proven to have
positive credibility by independent expertise or professionals, which help to confirm their
existing knowledge or familiar environment. The individual searches product information from
groups or people who have sufficient knowledge about these products, such as people working
with the product or people who use the product. Through observing, the individuals learn from
others whom might confront the same situations or have previous experience (C.Whan &
V.Parker, 1977, p. 103). Furthermore, according to the attribution theory, the purchasing
decision of the individual is influenced by another individuals or reference groups, such as
experts, authorities or celebrities (Jones, David E., Harold H., Nisbett, Valins, & Weiner, 1972).
Utilitarian influence: the individual needs certain conformity in a purchasing situation, otherwise
it is dangerous not to conform (Asch, 1952). Seeking a match to what another individual or
reference group prefers or satisfies is important when making a purchase decision, as social
interactions exist. Utilitarian influence also emphasizes explicit rewards and punishments
because of the individual having a desire to be more affirmed or accepted by social involvement,
such as a community, brand tribe or fan club. Therefore, an attempt to satisfy others
expectations is performed (C.Whan & V.Parker, 1977, p. 103).
Value expressive influence: individual seek consistency between themselves and a reference
group by sharing something in common, like buying products used, or those recommended by
the reference group. A positive self-image is important, since an individual is more willing to be
associated with positive referents instead of negative ones. Liking a person or a reference group
in a product advertisement can lead an individual to purchase it, regardless whether the product
fits the individuals need or not (C.Whan & V.Parker, 1977, p. 103).
To further explain the function of how celebrity endorsement works, reference group theory
looked up the ubiquitous features of the celebrity market from a psychological view. The
reference group theory viewed the function of celebrities influence on consumers from three
difference perspectives. Because of their special social status in our society, they are often
chosen by marketers to speak for the products. Celebrities exert the power of reference group on
consumers. It was therefore chosen to be discussed in the literature study of this research. Further,
the experimental research proved that endorsers should be used according to different product
types. Celebrity endorsers are widely used mainly because of their attractiveness and likable
features. This helps to explain the widely usage of celebrities in cosmetics advertisements.
Friedmans experimental studies showed that costume jewelry endorsed by celebrities
appeared to be the best match. Although the result did not test the match between cosmetics and
celebrities, it highly suggested that if brand name and advertisement recall are most desirable,
advertisers should use a celebrity as an endorser. Besides, the use of appropriate endorsers
Page | 18

might induce attitude change because of trustworthiness, expertise, similarity, attractiveness

and likableness (Friedman & Friedman, 1979). These findings have positive contributions to
this current research.
3.2.3 Endorsement by Product Types

Friedman and Friedman (1979, p. 63)categorized three different types of endorsers: (1) the
celebrity; (2) the professional (or recognized) expert; (3) the typical consumer. Previous
research studies have shown that celebrity-endorsed advertisements have higher effectiveness
than the other two types of endorser advertisements: the professional (or recognized) expert and
the typical consumer. Celebrity endorsements usually create physical attractiveness or liking.
Professional endorsers connect to a product by their expertise. Ordinary consumer endorsers
emphasize a similarity to a product in an advertisement, creating a resonance among the
endorsers and the consumers (Brock, 1965). The use of different types of endorsers aims to
create or influence attitudes or attitude changes. Two social influence processes, identification
and internalization, can interpret the functions of celebrity and expert endorsers (Kelman, 1961).
Identification: if an individual admires or believes in a person, such as a product endorser, they
would more likely agree with an attitude or behavior of the endorser. The individual is satisfied
with believing and liking because of their acts in the process.
Internalization: if an individual agrees with an attitude or behavior of another person, it is
because they agree with the essentiality of the new attitude or behavior. This process assimilates
their personal values to a new belief or attitude.
These two processes explain the function of different types of endorsers. Showing that different
product categories need different types of endorsers, which can be one of the reason why
celebrities are prevalently used in cosmetics advertisements. Identification explains an
advertisement consisting of an attitude or behavior advocated by a celebrity, emphasizing a link
to the attractiveness and likableness of celebrities. However, internalization occurs if the
information of an advertisement transmits an authorized, trustworthy and honest message.
Therefore, it underlines the persuasion power of an authority or expert (Friedman & Friedman,
1979).
3.2.4 Attitude Change Through Visual Imagery

Advertising is a visual and verbal product presentation, with the exception of radio advertising.
The visual content selection and decision of an advertisement is as important as writing an
effective copy in order to match the product. In fact, less people know about how effective the
visual content in convincing purchasing can be. The original function of advertising is to
enhance a product sale or promotion deal. As an alternative, most of the consumer product
advertisements are to persuade consumers to purchase while creating or keeping positive product
attitudes (Rossiter & Percy, 1980). This is different from the multi-attribute attitude model
proposed by Fishbein and Ajzen (Fishbein & Ajzen, 1975), Rossiter and Percys (1980)
hypothesis theory Visual Loop Possibility assumes visual content as being especially capable of
stimulating visual imagery. Verbal content and music can also produce visual imagery but only
when the visual content is absent, such as when people are reading a book, or listening to a song
Page | 19

on the radio. When both visual and verbal content are present in an advertisement, this enhances
the attitude approach as both visual imagery and verbal belief are taken into account (Rossiter &
Percy, 1980). Moreover, through the experiment of studying the mental imagery process, Roger
Shepard positively confirmed the visual imaginary behaviors as being able to influence actual
behaviors. An individual has the same opinions or present the same feedback regarding an object
whether it is present or absent (Shepard, 1978). Rossiter and Percy designed an experiment by
using beer to test a total of 88 beer drinkers, with four printed advertisements in different
formats and grouping them into two, high visual versus low visual emphasis and explicit
versus implicit belief claims (Rossiter & Percy, 1980). The result has shown to support and
prove their visual and verbal loop theory. It must be mentioned that the visual loop possibility is
not as effective as it should be if only visual content was present, so do the purely verbal loop
possibility. Therefore, an appropriate combination of both visual and verbal contents is highly
necessary in order to reach a more efficient result when establishing a product attitude. The links
between the different aspects are shown below in Figure 2.
Visual Content
Visual Content

Product

Product Attitude

Verbal Belief

Visual Imagery
Visual Content

Product

Product Attitude

Verbal Belief

Figure 2 Visual and Verbal Loop Possibilities (Rossiter & Percy, 1980, p. 12)
Rossiter and Percy (1980) developed the theory of attitude change through visual imagery in
advertising in which using visual content can be equally effective as verbal content when
building a favorable attitude and convincing a purchasing behavior in advertisements. Through
reviewing their research works, it helps to understand to which extent the visual content as a
marketing instrument can create an attitude and encourage consumers to buy. It also explains
why imagery usage is important in cosmetics advertisements. Besides that, it is necessary to
understand why imagery is prevalently used in cosmetic advertising instead of simply negating
its importance. Not only has it helped in finding a solution to the research question, but also for
providing appropriate recommendations to marketers in between the overuse and proper use of
celebrity endorsers. Although the strong visuals can complement good and produce
synergistically effective advertisements (Rossiter & Percy, 1980), other factors might also be
more potent than the celebrity endorsers in cosmetics advertising.

Page | 20

3.2.5 The Social Adaptation Theory

People can tolerate a short period of inconsistencies in order to find out suitable social adaptive
strategies, as well as to improve their existing social behaviors. This general tactic will dominate
peoples behaviors, such as self-perception, self-esteem, values, or other social cognitions from a
broader point of view. According to the social adaptation theory, subjective self-esteem only
closes to interpersonal relations and acts in a specific way, though it determines the pattern of
multi-interpersonal relationships and behaviors (Kahle, Kulka, & Klingel, 1980, p. 496).
Individuals are more likely to receive selective information which they are interested, in rather
than all of it, because memory retrieving is limited at each time (John G & Thomas K, 1982).
This does not imply that the information will influence the individual, as it highly depends on the
usefulness of that information to their social adaption strategy. The individual will re-process
new information if the one received was not suitable. The impact of information, however, is
dependent on the meaning of the adaptation to an individual (Kahle & Homer, 1985, p. 954).
The match-up hypothesis of the celebrity endorser selection corresponds to the social adaption
theory. The image of a celebrity endorser and product information should be matched at an
intersection in an effective advertisement to attract target audiences because of their social
adaptation need. The hypothesis brings up the suggestion that an advertisement is more efficient
than verbal communication through the image match-up communication (Rossiter & Percy,
1980). Aristotle said that Personal beauty is a greater recommendation than any letter of
introduction (Aristotle).
Kahle and Homer (1985) conducted an experiment by using three elements: celebrity-source
physical attractiveness, celebrity-source likability and participant product involvement to
measure attitude and purchase intention. The research findings positively confirmed that the
attitude and buying intention changed, which are mainly influenced by celebrity-source physical
attractiveness. The variable of attractiveness has a significant difference at the 0.05 level,
resulting from a quantitative measurement, as well as the empirical observation. The results
explained the importance of this factor being used in advertising, as supporting the match-up
hypothesis, social adaptation prediction and the informational value of communications (Kahle
& Homer, 1985, p. 960). The elaboration likelihood model theory depicted attitudes change will
be received in two ways, which highly depends on to which extent consumers are involved with
a brand in an advertisement. This is also referred to in Figure 3.
Explaining the social adaptation theory, Kahle et al. revealed the logic behind using celebrities
in advertisement since it might influence consumers attitudes. The theory can be used in
predicting whether a personality characteristic will show causal predominance over some
elements of an action sphere. However, those elements are yet to be unidentified (Kahle, Kulka,
& Klingel, 1980). The social adaptation theory is popularly used in latter research studies in
finding the connection between product types and endorsers. Kahle and Homer (1985) developed
the match-up hypotheses which further proved that brand attitudes would change when the
product was related to attractive endorsers (physical or status). The purpose for providing the
review of the theory is to better understand how attitudes might change under certain occasions
to certain people. It further helps in understanding of the overused celebrity endorsers in
cosmetics advertisements. As to effectiveness in building brand loyalty, it must be tested in a
later study of this research project.
Page | 21

3.2.6 The Elaboration Likelihood Model Theory

Figure 3 The Elaboration Likelihood Model (ELM) of Persuasion (Solomon, Bamossy,

Askegaard, & Hogg, 2006, p. 223)

According to the elaboration likelihood model, individuals will process received information by
different routes, central route and peripheral route. Each individual follows one of these routes,
depending on what level of involvement of the individual, high-involvement processing or lowinvolvement processing. Under a high-involvement processing, the individual will process the
information through the central route. Under a low-involvement process, the individual will most
likely process the information through the peripheral route.
The central route of persuasion: if persuasive communication has the information of which an
individual is interested in or feels related to their situation, the individual is more willing to make
a diagnosis about this information and come up with a conclusion. The individual will carefully
consider a message sent by an advertisement, whether or not this message is practical or useful
regarding their situation.
The peripheral route of persuasion: an individual is not motivated to make a diagnosis of a
message, but simply taking either positive or negative cues associated to a product or topic in the
message in order to decide upon an attitude; a delicious meal, bad weather or a beautiful dress,
etc. The message may also be sent out by an endorser desire or trusted by the individual, so they
would agree with what these persons recommend because of their reputation, attractiveness or
authorization. The peripheral process can also create a negative attitude, because a message is
sent out by a negative reference, such as an endorser being disliked by the target consumers.
However, none of the theories have provided a comprehensive explanation regarding an attitude
change. The elaboration likelihood model of persuasion tested both attitudes through highinvolvement process and attitudes through low-involvement process (Petty, Cacioppo, &
Schumann, 1983, p. 135). It is important to point out that both high and low involvement follows
Page | 22

the same way to process information received in social adaption theory, while the Elaboration
Likelihood Model theory indicates that high and low involvements process information
separately (Kahle & Homer, 1985, p. 955).
The elaboration likelihood model, from a different angle explicates the external stimulus not
always following the same route in consumers minds. Through the experimental study, Petty et
al (1983) suggested that the celebrity status or credibility of the product endorsers play an
important role when consumers have a low product involvement. Under low involvement
conditions, attitudes or behaviors might respond to external stimulus. Unlikely to the low
involvement, the celebrity status had no effect on attitude or behavior change (Petty, Cacioppo,
& Schumann, 1983). The findings of the study are therefore positively supported by the purpose
of this research project, except it did not involve any loyalty perspectives.

3.3 Brand Loyalty

Brand loyalty is a phenomenon that has been investigated numerous times but with no clear sign
of capability to generalize the results attained. Though, it is commonly argued that brand loyalty
can be defined in two different ways, i.e. there are two different kinds of loyalty. Some
researchers define loyalty as something purely caused by behavior, behavioral loyalty, while
others argue for the existence of loyalty caused by attitude, attitude-behavior loyalty. The former
one suggests that a customers purchasing pattern (i.e. repeat purchase) of a specific product or
brand can lead towards the establishment of loyalty. However, the distinction between repeat
purchase and brand loyalty is very hard to define. Some researchers argue for the combination of
a favorable attitude and repeat purchase behavior in order to explain the concept of brand loyalty
(Dick & Basu, 1994). One type of loyalty is not sufficient enough without the other.
Bandyopadhyay and Martell stress the importance of including non-users when making the
distinction concerning brand loyalty. These users are the ones positioned on the bottom of the
loyalty hierarchy (where single users are at the top and multiple users in the middle), since
they are not users at the moment, but have the possibility to become users of the particular brand
in the future. This group might not be current consumers of the brand due to different reasons,
such as lack or purchasing power etc. Though, if consumers have a favorable attitude towards the
brand they can become loyal customers in the future. These are the ones focus should be directed
towards, since they already have a favorable attitude, and customer retention can be created.
Hence, this is not a loyal customer but the possibility is there, even without the repeat purchase
behavior (Bandyopadhyay & Martell, 2007).
The authors, Bandyopadhyay and Martell (2007), further define brand loyalty as a customer
who only buys one brand over the last 1 year whom are then called single users. In addition to
this category and the previously mentioned non-users, one more group of users is defined, the
multiple users. This group is the ones buying more than one brand, compared to single users,
whom only purchase one brand.

Page | 23

Attitudinal
Behavioral Loyalty
Loyalty
Single Users

Multiple Users

Non-Users

Strong

Brand loyal

Variety Seeker

Potential Buyer

Weak

Constrained Buyer

Deal Prone

Indifferent

Figure 4 Consumer Brand Loyalty Classification According to Their Behavioral and

Attitudinal Characteristics
The non-users are recognized due to their low amount of repeat purchasing and unfavorable
attitude. Though by stating that, these consumers should not be forgotten in the brand-loyalty
hierarchy, since as stated above, they are a great source of potential future customers. Between
the three different categories, multiple users are the biggest group, where the consumers are
exercising brand switching to a high extent. A favorable attitude is directed towards a number of
different brands and does not apply to only one specific brand. A possible cause for multiple
usages is the perception of different brands serving the same or a similar purpose without a
specific favorable attitude towards any of them, thereby increasing the brand switching
frequency.
The last category, single users, is the one with the highest rate and probability of brand loyalty.
Bandyopadhyay and Martell (2007) argue for a difference even within this category, where they
make a distinction between truly brand loyal customers and constrained buyers, where the
former category has a strong attitude and the latter a weak attitude. This weak attitude possibly
stems from a lack of options or favorable pricing strategies for the significant brand. Such
customers can be unpredictable and may become non-users if competitors offerings and
marketing of it are preferable in comparison to the brand they are committed to at the moment.
However, single users are considered to be difficult to influence with marketing tools, in the
sense they have a favorable attitude that is hard to affect with promotion and other means. Oliver
(1999, P34) describes this specific type of loyalty as;
A deeply held commitment to re-buy or re-patronize a preferred product/service consistently in the future, thereby
causing repetitive same brand or same brand set purchasing despite situational influences and marketing efforts
having the potential to cause switching behavior.
The theory of brand loyalty is of great importance to the research question in this study. Brand
loyalty is something that every marketer of a brand is striving for. The result of brand loyalty
stems from different things from time to time, but it can be argued that the most desirable is to
achieve loyalty based on a favorable attitude towards the brand. This is an interesting aspect for
this study, what is the reason for the creation of brand loyalty and why is that so. One has to
understand that brand loyalty is not a simple concept that is easily achieved by brands and
marketers. Different aspects need to be considered in order to see the link between an action and
its result. How can a marketing decision lead to the forming of brand loyalty, and why does it
happen in some cases and sometimes not. The theories examine varied aspects related to
different types of loyalty and the upcoming of it through both attitude and behavior.
Page | 24

Therefore, these concepts will exclusively be presented below, in order to provide the reader
with a deeper insight concerning the differences and similarities between the two of them.
3.3.1 Behavioral Loyalty

Behavioral loyalty is a function of a favorable repeat purchase pattern (Bandyopadhyay &

Martell, 2007). This statement argues for the creation of brand loyalty through exercising repeat
purchase of a specific brand over a considerable amount of time. A number of different aspects
have to be taken into consideration when explaining behavioral brand loyalty; it may have been
caused by brand favorable or non-favorable factors. A consumer might exercise brand loyalty
through a behavioral perspective by a habitual manner or lack of knowledge concerning other
alternatives (Chaudhuri & Holbrook, 2001).
3.3.2 Attitudinal Loyalty

Attitudinal loyalty is comprised of a favorable disposition towards the brand (Bandyopadhyay

& Martell, 2007). While repeat purchase is the only mean to define loyalty, in a behavioral
manner, many authors argue that an attitude in relation to the behavior is necessary in order to
reach the proper definition of brand loyalty. Bandyopadhyay and Martell (2007) review six
different conditions that are argued to lead towards loyalty, described in the following way;
Loyalty is (1) the biased (i.e. nonrandom) (2) behavioral response (i.e. purchase) (3) expressed over time (4) by
some decision making unit, (5) with respect to one or more alternative brands out of a set of such brand, and (6) is a
function of psychological (decision making, evaluative) processes.

The statement that follows is that these conditions are the important cornerstones when the
consumers are developing a commitment towards the brand. This is what distinguishes brand
loyalty from regular repeat purchase behavior. Figure 5 displays the Brand Loyalty Model
proposed by Dick and Basu (1994).

Page | 25

Figure 5 Brand Loyalty Model

The model explains the different factors influencing and leading to brand loyalty. Both
behavioral and attitudinal aspects are influential causes throughout the entire model, ending up at
the stage representing customer brand loyalty. Perceived brand-performance
brand
rmance fit considers the
attitude a customer has towards a brand and the ability to perform according to customers
expectations. Social and emotional identification with the brand points out the image a brand
mediates and the customers ability to relate to it. Habit and long history with the brand is related
to the behavioral aspect, where customers exercise a recurring behavior towards the same brand.
The three previously mentioned aspects lead to a further attitudinal behavior, where likeability is
the contributing factor. If it is positive it will develop even further, into a repeat purchase
behavior, strengthened by the actual likability and thereby transform into customer brand loyalty.
As can be seen both from the concept explanations of attitudinal
attitudinal and behavioral brand loyalty as
well as the model presented above, there is a linkage between customers perceptions, actions,
preferences and behavior. This link shows how all of the aspects involved result in brand loyalty
due to a certain behavior and attitude towards a brand. For a brand to be able to generate
favorable perceptions and attitudes from its existing and potential customers, it is important to
acknowledge and identify reasons resulting in that. The authors identified two important factors
for brands when exposing themselves and directing advertisements and statements towards
customers; brand affect and brand trust.
As can be seen both from the concept explanations of attitudinal and behavioral brand loyalty as
well as the model presentedd above, there is a linkage between customers perceptions, actions,
preferences and behavior. This link shows how all of the aspects involved result in brand loyalty
due to a certain behavior and attitude towards a brand. Furthermore, the link provides an insight
in how brand loyalty often is a result of marketers actions resulting in the creation of
Page | 26

preferences and perceptions in a consumers mind. For a brand to be able to generate favorable
perceptions and attitudes from its existing and potential customers, it is important to
acknowledge and identify reasons resulting in that. This is of great importance to our study since,
as stated in the section concerning brand loyalty, the reasons leading to loyalty are essential to
acknowledge in order to detect a pattern and make decision for further actions. Two important
factors for brands are identified for directing advertisements and statements towards customers;
brand affect and brand trust.
3.3.3 Brand Affect and Brand Trust

Chaudhuri and Holbrook (2001) explain brand affect and brand trust with the following
definitions and descriptions:
Brand affect: a brand's potential to elicit a positive emotional response in the average
consumer as a result of its use. (Chaudhuri & Holbrook, 2001) It consists of different factors
affecting customers leading to a factorable attitude towards the brand, where brand image and
likeability are some of the crucial factors. Brand affect is developed through a nature that is
impulsive and instantaneous. A positive (or negative) affect can take place in an instant.
Brand trust: the willingness of the average consumer to rely on the ability of the brand to
perform its stated function. (Chaudhuri & Holbrook, 2001) Trust is built up of three key
foundations, reliability, safety and honesty, and is of major importance for both the company and
its customers. It provides the customers with a sense of confidence through reducing
uncertainties in the environment they are present in, which in turn generates more value to the
brand as well. The nature of brand trust is developed in a manner more thought through and
considered before establishing the certain level of trust in comparison to brand affect
development.
Brand loyal customers tend to be more willing to pay a higher price for the product, stemming
from the perception of the brand providing additional value that no competing brands have the
ability to provide. For the brand itself, the most significant benefit is the increased market share,
since customers recurring purchase behavior generates more value to the company. A number of
affective factors are present when describing brand loyalty; where one suggests that customers
are continuously using the brand. Hence, the usage of the brand is likeable and customers might
relate to its image (Chaudhuri & Holbrook, 2001).
A combination between the two concepts of brand affect and brand trust makes up the basis for
and have an effect on attitudinal and behavioural brand loyalty. The considered brand provides
the customer with an image they can or want to relate to, in addition to trust regarding the image
the brand convey to the public. Chaudhuri and Holbrook (2001) argue for a number of benefits
obtained through brand affect and established brand loyalty; lowered marketing costs, an
increased number of new customers, favorable word of mouth and decreased willingness to try
competitors offerings.
As the concepts of brand affect and brand trust argues, a positive emotional response towards a
brand and the trust of a brand performing its function will generate an attitude that is dependent
on what the brand transmits. When consumers trust a brand and believe in the message, a long
Page | 27

term type of loyalty can be the possible outcome, whereas brand affect might be of a short term
nature. Thus, a strong attitude and loyalty are likely to be the result of brand trust, and a certain
behavior might be the result of brand affect. Here it can be seen that different factors are possibly
influencing the two types of loyalty considered.
Since the two different types of brand loyalty can be achieved through and are affected by
various different approaches, it is appropriate to emphasize a relationship between them. This
decision was made due to the possibility of loyalty creation, either through a basic separation or
through a combination between the two of them.
3.3.4 The Relationship Between Two Loyalties

According to Dick and Basu (1994), the concept of behavioral loyalty does not stress underlying
factors contributing to the creation of brand loyalty. It is difficult to explain the expansion of
loyalty by only arguing for a behavioral source of reason. It is therefore argued that brand loyalty
is most suitably explained by a combination between behavior and specific attitude towards a
brand. A continuous purchase behavior of a brand is therefore not the only factor resulting in
brand loyalty, but a positive attitude towards the brand. This attitude is the important aspect
resulting in loyalty, since it creates a want to perform repeat purchases, which would be a
different matter with the involvement of a negative attitude.
Dick and Basu (1994) argue for differences concerning the type of loyalty between and within
different markets, based on the nature of involvement and risk. The distinction made is related to
three types; consumable goods, durable goods, and services. The consumable goods markets
have a high rate of switching brands due to the low amount of involvement and risk related to the
products included in this category, i.e. food and clothing. Therefore, behavioral measures
without including attitude are suitable for this category of goods. On the other hand, durable
goods (i.e. furniture, laptops etc) are related to a significantly higher amount of involvement and
risk, resulting in the urge to switch brands not becoming very high in this specific category.
Services, however, is of another matter. They often involve a considerable amount of risk and
thereby loyalty towards one brand becomes more feasible (Bandyopadhyay & Martell, 2007)
The relationship between the two types of loyalty is interesting due to their difference in the
creation of brand loyalty. One aspect, the attitudinal, can be considered to be conscious while the
other, behavioral, is more related towards an unconscious upcoming. As Dick and Basu argues, it
is very likely that a combination of the two different types is needed in order to evaluate
something as complicated as brand loyalty. The consumers mind is complex, and therefore we
want to encompass the possibility of brand loyalty being created by a combination between
different characteristics, and not just a simple one way or another.
3.3.5 The Attribute Frequency Concept

Loyalty towards products with low rate of involvement in the purchasing process tends to be
connected with the attribute frequency concept. Hence, the greater the number of and the more
favorable attributes a brand products is considered to have, the greater the liking of the brand.
The number of positive attributes can itself influence the perception and the choice the consumer
makes when lacking other sufficient information. One brand can be preferred in front of another
Page | 28

if the consumer considers brand one to possess more favorable attributes in comparison to brand
two, thereby generating a stronger attitude towards brand one in this case (Bandyopadhyay &
Martell, 2007, p. 40).
Since the two different types of brand loyalty can be achieved through and are affected by
various approaches, it is also here appropriate to emphasize a relationship between them. This
decision was made due to the possibility of loyalty creation, either through a basic separation or
through a combination between the two of them.
The attribute frequency concept can be related to brand affect, since the more favorable attributes
a product possess, the more likely preference is created towards that product. As brand affect is
dependent on a positive emotional response, the attributes related to the product are of great
importance. Favorable attributes will generate consumers likeability when comparing the
product to other available offerings. If a consumer does not have a specific attitude towards the
brands in a specific product category, the most likely scenario is to purchase the one with most
favorable attributes, generating a behavioral loyalty as long as the attributes stay the same.
3.3.6 The ABC Model of Attitudes

It is widely argued that a consumers attitude has three different components: affect, behavior and
cognition (ABC). The first aspect, affect, is related to how the consumer feel about a specific
object. The second aspect, behavior, refers to the customers intention to actually do something
regarding the object in question. The third aspect, cognition, is a description of a consumers
beliefs about the object. The model in question puts an emphasis on the interrelationships
between the three previously mentioned aspects, feeling, doing and knowing. A customers
attitude towards a product cannot only be determined by one of these three factors, all of them
have to be present and known in order to make a proper interpretation. What might differ is the
level of commitment towards the product, through the level of use, heavy or light. The
development of hierarchy of effects intended to explain the different components impact, and a
series of steps within it leads to a certain attitude (Solomon 2007, p237). Figure 6 below displays
the ABC Attitude Model.

Page | 29

Attitude
Standard Learning Hierarchy
Cognition

Affect

Behavior

Attitude

Low-Involvement Hierarchy
Cognition

Based on
cognitive
information
processing

Behavior

Affect

Based on
behavioral
learning
processes

Experiential Hierarchy
Attitude
Affect

Behavior

Cognition

Based on
hedonic
consumption

Figure 6 ABC Attitude Model (Solomon, Bamossy, Askegaard, & Hogg, 2006, p. 238)
The standard learning hierarchy illustrates the formation process of an attitude most commonly
assumed. A belief about a product is formed, followed by the formulation of a feeling which over
time will result in a behavior. The low involvement hierarchy is different, here the consumer is
not as interested in the product or service in question. This type of involvement is based on
limited knowledge and not particularly strong preferences towards the brand. The last type, the
experiential hierarchy, suggest that consumers actions are based on emotional response
(Solomon 2007, pp238-239).
Each and every theory has its own contribution, and they are all in a sense linked together. The
different types of brand loyalty covered in this study, attitudinal and behavioral, are reached due
to different reasons. The creations of specific attitudes, explained by the ABC model, are of great
importance since it covers three important aspects, which by being achieved differently causes
different types of attitudes. It also takes behavior into account, pointing at the previously
mentioned relationship between attitudinal and behavioral loyalty. As Dick and Basu argued, it is
difficult for behavioral loyalty to exist without an attitude related to it. An individual will most
often have some sort of attitude towards a product or brand being purchased, on a different level
depending on the involvement regarding the product purchase. This could create difficulties for
brands management when evaluating product performance and its causes. Knowing what type of
loyalty that is achieved, and if it can be considered to be loyalty at all, can help in estimating the
effectiveness of its advertisement techniques.

3.4 Conclusion
In formulating the theoretical framework for this research, several important theories in relation
to celebrity endorsement, brand loyalty and its different functions provide essential support to the
current research project. The different theories covered in section 3 are all relevant for the study
and they contribute with a deeper insight concerning the relationships between them and how
and why they can be applied to the research problem.
Page | 30

A model covering all the different theories included in this study was created in order to provide
a better insight concerning the relationship between them. It includes the relevant theories related
to endorsement in the first two sections of the figure, and the ones related to brand loyalty in the
bottom two. The model shows a relationship where all theories meet in the middle, represented
by behavioral loyalty and attitudinal loyalty. The model is presented in Figure 7.

Meaning Transfer

Reference Group
Theory

Endorsement by
Types

Social Adaptation
Theory

Visual Imagery

The ELM Model

Behavioral Loyalty

Attitudinal Loyalty

ABC Model

Brand
Trust&Affect

Attribute
Frequency

Figure 7 Theoretical Framework

In order to sum up the literature review chapter, this research study is to explore whether the use
of endorsers or celebrity endorsers can attain brand loyalty or not. Based on the theoretical
framework and the presented literature reviews, we use the traditional approach (Creswell, 2003,
p. 132) for testing the relationships between variables. Four null hypotheses come to light from
these results and will therefore be tested later on:
H1: There is no significant relationship between the usage of endorsers in cosmetics
advertisements and behavioral loyalty.
H2: There is no significant relationship between the usage of famous endorsers in cosmetics
advertisements and behavioral loyalty.
H3: There is no significant relationship between the usage of endorsers in cosmetics
advertisements and attitudinal loyalty.

Page | 31

H4: There is no significant relationship between the usage of famous endorsers in cosmetics
advertisements and attitudinal loyalty.

Page | 32

CHAPTER 4 PRACTICAL METHODOLOGY

This chapter will focus on practical methods and data collection procedures applied in this
research project. This part will emphasize the data collection process and everything else
relevant to it. As data collection and its analysis have an important influence on the quality of
the research, it is imperative to carefully define each step of the process.

4.1 Research Survey

Quantitative data collection is used for the purpose of this study, where a research survey is
necessary and practical when collecting data required for later analysis. Many details have to be
taken into consideration, such as what questions to be asked in the survey in order to best reflect
data patterns, correlations, frequencies, etc. Based on the research problem, the survey will cover
three main sectors with sub questions under each sector. The main sector include: the influence
of advertisements on the purchase of cosmetics products, a measure of behavioral loyalty and a
measure of attitudinal loyalty.
A self-administered web-based survey was conducted, where the respondents visit a website
provided by the researchers and fill out the answers from there. This type of survey requires the
respondents to have access to the internet but also some knowledge concerning computers in
order to complete the questionnaire. In order for it to be as effective as possible, the questions
need to be clear and understandable, since the providers of the survey are not present to answer
questions that might arise. There are both pros and cons with performing a web-based survey.
There are a number of advantages to using a web-based survey over a traditional paper-based
survey. Foremost, it is very cost efficient, the expenses are few if any. You can also reach a
significant number of respondents through different databases and alike listing e-mail addresses
to a particular population. The absence of the researcher enables respondents to have more
privacy while participating in a survey. It also reduces the pressure on respondents, as they can
take the survey in an environment familiar to them.
However, using web-based survey also has some drawbacks to be bear in mind. Coverage errors
might occur in the sampling, which means everyone not having a non-zero chance of being
included. Also technical problems might arise, such as hardware and software errors in a webbased survey, and not all respondents are familiar or experienced with the internet. Reliability is
also a problem when using a web-based survey. Respondents may be forced to answer questions,
despite being unqualified to answer, or they may not want to disclose the truth. Sometimes, drop
out is a problem when using the Internet for a survey, because respondents might be
disconnected while filling out the survey. In addition, respondents might not even have the
interest of doing it. All these drawbacks could lead to a low respondent rate, by using a webbased survey (Cohen, Manion, & Morrison, 2007, pp. 230-235).

4.2 The Survey Design

The design of the survey will follow five different steps, set up according to Kent (2007, p. 152).
1.
2.
3.
4.

Specify information needed

Decide upon method of questionnaire administration
Determine the content, format and wording of each question
Create a draft of the complete questionnaire
Page | 33

5. Test the questionnaire

Based on the fundamental procedures of survey design, it is necessary to clarify specific steps of
the design for this particular study. The main purpose is to measure whether the use of celebrities
in cosmetics advertisements is effective in establishing behavioral loyalty and attitudinal loyalty
among consumers. The survey serves as a numeric description of consumers behavior and
attitudes regarding celebrity endorsement in cosmetics advertisements. Since there is not a direct
way to attain the purpose, the survey can only measure each variable separately. The survey is
therefore developed to measure four different types of variables:

Demographic variables, such as age group, purchasing frequency, important factors

considered when buying cosmetics, etc.
Opinions regarding celebrities and cosmetics advertisements
Behavioral loyalty attributes
Attitudinal loyalty attributes

In order to measure the four variables, proper measuring methods should be applied to each type.
The data for this study was collected through a survey that included 25 questions. For the
demographic variables, it is most appropriate to use multiple choices. Metric scales are
appropriate when asking questions regarding age group and amount spent on cosmetics. Since it
is impossible to list all ages and spending intervals, a use of a metric scale will provide a suitable
alternative. Given four or five scale options, respondents should be able to answer all questions
regarding ranges for age, spending, frequency, etc. However, it should be avoided to use scales
that are overlapping (Kent, 2007, p. 127).
The other variables, a decision of using rating scales was made, i.e. a likert scale for measuring
respondents opinions and attitudes. The likert scale is efficient in indicating these types of
questions as they afford the researcher the freedom to fuse measurement with opinions, quantity
and quality (Cohen, Manion, & Morrison, 2007, p. 327). Respondents are asked to specify the
level of agreement or disagreement with a series of statements concerning objects or attitudes in
the likert scale measurement. Here we decided to use a 5-point rating, since it would be less
complicated for respondents to indicate, as well as it reduces the complexity when analyzing the
data. Two types of ratings are used, strongly disagree to strongly agree, in order to test attitude
statements as well as respondents opinions regarding cosmetics advertisements. Very
unimportant to very important is used to indicate a salient factor which influences the
respondents cosmetics purchase behavior. The survey was tested before sending it out to the
target respondents, in order to assess credibility. According to the test results, the questions were
clear and easy to understand. A more thorough explanation concerning the test will be covered in
section 4.5.
Although rating scales can be useful and practical in research, one should always be aware of
certain limitations related to it. The study could be influenced through the use of sensitive data
in a sense that if it is out of the respondents acceptance, they might not answer the question in a
truthful manner. It is impossible for researchers to find out whether respondents are giving the
truth or not. In addition, respondents might have comments other than what are listed as options
in the survey. Very few of us would like to be considered as extremists, which is why
respondents often tend to be neutral. For this reason the mid-point of a scale is often opted.
Page | 34

Concerning those statements, rating scales are considered to be more sensitive than dichotomous
scales. However, dichotomous scales limit the number of choices the respondents could make.
For the reason of usefulness, rating scales are widely used in many social researches (Cohen,
Manion, & Morrison, 2007, pp. 327-328).

4.3 Sampling
Two main sampling strategies are often used in social science research: the probability sample
and the non-probability sample. Researchers need to decide upon one of them to be used in their
study. Using probability samples (e.g. random samples), participants of the case can be selected
into the sample with a known chance. Usually, the accurate probability would not be able to be
defined before the actual sample size has been selected. The word known here implies that an
exact probability can be calculated by a particular technique of selection, if the accurate
population size were identified by the time the sampling took place. A probability sample is
selected from a large population. Because of its representativeness, researchers wish to use such
a sample to make generalizations (Cohen, Manion, & Morrison, 2007, p. 110). The probability
sample involves two techniques to process the sample selection, randomization and systematic
selection. The randomization selection is based on the researchers independent judgment to
select sample sizes from a complete list of population units, whereas the systematic selection
imposes certain rules in the sampling selection process so that the human judgment is eliminated
from the selection (Kent, 2007, p. 231). With the randomization technique, a representative
sample is able to characterize the pattern of the entire population (Creswell, 2003, p. 156). Using
the probability sample will help in reducing the risk of bias, in comparison to the use of the nonprobability sample. Although, this does not imply that probability will ensure the sample to be
risk free (Cohen, Manion, & Morrison, 2007, p. 110).
Non-probability samples (e.g. nonrandom samples) are quite subjective. The use of nonprobability samples occur when a researcher has a specific target group or purposes to be studied.
A non-probability sample usually represents itself, but not the entire population (Cohen, Manion,
& Morrison, 2007, p. 113). Most of the time, the respondents in a non-probability sample are
chosen on the basis of their availability and convenience (Creswell, 2003, p. 156). Therefore, it is
not likely to calculate the chance of selecting a case from a population of cases. With a nonprobability sample, the population might or might not be represented by the selected samples.
The selection is very much based on the subjective judgment of researchers. Besides, this
subjective judgment is usually restricted in several different ways (Kent, 2007, pp. 231-232):

By types of people to be chosen in quantities, such as how many students or teachers

An accurate time of the interview taking place must be indicated
Locations where the interview will be taking place

In many small-scale researches, the non-probability sample is frequently used. The reason for
using the non-probability sample in small-scale research is because they are far less complicated
to set up and are considerably less expensive. They are also adequate where researchers do not
intend to generalize their findings beyond the sample in question (Cohen, Manion, & Morrison,
2007, p. 113).

Page | 35

The intention is to study a small group of females who uses cosmetics and are currently residing
in Sweden. Also, the sample included should possess certain brand awareness, which means they
have seen or heard of the most common cosmetics brands available on the market. This is only
something that can be assumed in beforehand. Our survey intends to recruit 100 respondents,
which are believed to be seen as a small-scale research. The characteristics of the sample are
defined based on the nature as well as the purpose of the study, something that is considered to
be subjective from our point of view.
Hence, the non-probability sample method is the most appropriate approach to be used in this
study. It was deemed that females are the most suitable respondents for this research, and was
therefore chosen as the sample. Moreover, we believe our sample might be a bit limited to be
considered as a probability sample, since the focus mainly is to recruit students.
We identify the sampling technique as a combination between snowball sampling and volunteer
sampling. Snowball sampling is a small amount of individuals whom are chosen by the
researchers. These individuals might have the desirable characteristics which researchers are
looking for, although not each individual has the exact traits. These attributes apply to the
intention of only including females, residing in Sweden, whom are recognizing cosmetics brands.
Regarding the volunteer sampling, it relates to researchers relying on personal friends, friends of
friends, or participants who reply to a newspaper advertisement. In the case of this study, we ask
potential respondents to forward the survey to friends who also reside in Sweden (which is
mentioned in later in this section). (Cohen, Manion, & Morrison, 2007)
Researchers should be aware of the limitations of using the volunteer sample method, regarding
the generalizability and representativeness, as those volunteers might have different motives for
performing the survey which cannot be identified. (Cohen, Manion, & Morrison, 2007)
The sample included in the study was attained through gaining access to a mailing list of USBE
(Ume School of Business) students, in addition to a list of international students email from the
International Office at Ume University. An email containing the web survey, an introductory
letter and an explanation of the study and its purpose was then distributed among a number of the
students represented on either of the two lists. The ones receiving the e-mail were also asked to
forward the email to other possible respondents fitting within the criteria; those being a female
residing in Sweden, with some knowledge or experience concerning cosmetics products. Hence,
possible respondents have a choice whether to participate in the study or not.
A total of 124 responses were collected, but we chose to conduct the analysis by the time we had
100 respondents. We decided to not take the rest of 24 respondents into our sampling pool
because that the 24 respondents took more than two weeks to complete the survey. Compared to
the 100 respondents whom completed the survey in one week, the 24 respondents took more than
one week to fill in the survey which their answers may be bias due to the time intervals are not
equal. Although 100 respondents cannot be considered large enough to represent the entire
population, a sample size of thirty is the minimum number of cases for researchers to use in
order to perform statistical analyses on the data. The type of the research is also an aspect to
take into consideration when recruiting the sample. Generally speaking, larger samples have a
greater reliability in comparison to smaller samples, as well as they enable more sufficient
Page | 36

statistics to be used in the data analysis. However, a large sample is not a guarantee for
representing the entire population. Researchers may consider four factors in making the sampling
size decision: 1) the sample size; 2) representativeness and parameters of the sample; 3) access
to the sample; 4) the sampling strategy to be used (Cohen, Manion, & Morrison, 2007, p. 101).
Although large sampling can better represent the population, there are still some general rules
that should be followed when large sampling is needed. Cohen, Manion and Morrison (2007)
suggested that the sampling size should be large when:

There are many variables

Only small differences or small relationships are expected or predicted
The sample will be broken down into subgroups
The sample is heterogeneous in terms of the variables under study
Reliable measures of the dependent variable are unavailable

Because the sample does not meet by the general rules, but the minimum sampling size of 30
respondents is reached, we decided upon a sampling size of 100 females to be used in our study.

4.4 Reliability
The sample size of 100 is lack of sufficiency to represent the entire population of the market.
However, using 100 females as the sample size shall be able to indicate the trend and behavior of
female consumers at a reliable extent. Reliability refers that a scale generates the same result if
measurements are repeated. It only matters the internal consistency instead of the sizes of
sampling. In this case, by using the same scales, all items should reflect some single underlying
dimension; statistically that means that they should correlate one with another (Kent, 2007)
Cronbachs coefficient alpha test is popular used in measuring the internal reliability for multiitem summated rating scales in surveys. It takes the average correlation among items in a scale
and adjusts for the number of items. If scales are reliable, they should have high average
correlation with a rather great number of items. The coefficient varies between zero for no
reliability to unity for maximum reliability. In most of the social science researches generally
apply the recommendation that an alpha value of 0.7 of any scale is sufficient enough to be
considered to be reliable. However, for preliminary research the minimal acceptable reliability
can be in the range of 0.5 to 0.6 as well. (Kent, 2007, pp. 142-143)
Hence, this research study will apply the Cronbachs coefficient alpha test with a value of 0.7 to
check the reliability of the scales used in the survey.

4.5 Validity
Validity is an important concept in order to make the results trustworthy and enhance the
possibilities for further studies and research. Tests and research designs possess validity if they
actually measure and test what they are supposed to (Graziano & Raulin, 2010, p. 162).
In order to ensure the study can be considered valid, it had to be made sure that the questions
asked in the survey actually measure the problem in the research question. This should be related
to the theories used, which should function as a stable basis that is clear and well supported
(Graziano & Raulin, 2010, p. 163). Each and every question was developed in relation to the
Page | 37

different theories and the sample included in the study can be generalized due to the times,
places and conditions covered (Graziano & Raulin, 2010, p. 163). Hence, the questionnaire and
its results are considered to be provided with validity.
With the purpose of making sure the questionnaire was not misunderstood or interpreted in an
incorrect manner, the authors decided to perform a pilot study. This type of study is of great
importance when administering a self-completion questionnaire, to make sure both the survey
questions and the research instrument itself are functioning properly (Bryman & Bell, 2007, p.
273). After performing the pilot study on ten different respondents, chosen due to their
characteristics being the same as the sample group of the study, a few minor revisions of the
questionnaire were completed based on the respondents comments.

4.6 Statistical Tools

Statistical tools are used as a mean for analyzing and interpreting the data generated through the
questionnaire answers obtained in the study. A number of different ways of handling the data are
available, and one has to evaluate which ones are the most appropriate for this specific set. When
these tests are completed, conclusions can be drawn from the results and interpreted thereafter.
The vast amount of raw data collected through the questionnaires needs to be handled and
analyzed in a suitable manner. A number of different statistical packages are available for the
purpose of processing the data. For this study, the optimum way is perceived to be using the
Statistical Package for the Social Sciences (SPSS), which can provide relationships, correlations
and other statistical measures. As the survey mainly consists of numerical data, SPSS could be
utilized to simplify the data analysis process. SPSS is also utilized in many social science
researches because of its easy and powerful functions (Cohen, Manion, & Morrison, 2007,
p.501). Before handling the data in SPSS, the results are processed and converted into numbers
in Microsoft Excel to facilitate the transformation into usable data that can be generalized.
Demographical findings will be processed and presented in bar charts to facilitate the
interpretation of this specific type of data and to easily contrast the differences. A reliability test
with Cronbachs Alpha (mentioned in section 4.4), an ANOVA analysis and a correlation
analysis will all be used in order to find correlations, relationships and strength of relationships
between the different factors found in the study. Furthermore, a descriptive analysis of the
factors will be presented to provide a picture of which ones are found to commonly be the most
important. The test and their functions and purpose are further explained in the research findings
and analysis section. These specific tests were chosen due to the nature of the study and what the
authors intended to investigate.

4.7 Limitations
The limitations regarding the methodology chapter considers different factors not included or not
performed in the study due to different reasons, which will be further explained in this section.
One of the most important parts of the practical methodology chapter is the actual sample and the
methods for reaching the respondents. The sample size of 100 respondents can be considered to
be rather small when performing a study of this type. Though, as mentioned before in Section 4.4,
it is showing a trend of behavior and attitudes, and thus the results can be generalized throughout
the specific segment it represents.
Page | 38

Additional statistical tests can be performed to provide a broader picture of possible outcomes of
the results and its contributions to the study. However, it was decided upon to only utilize the
tests already mentioned to be included in the analysis, with the intention of only covering the
most important aspects related to the problem. Hence, it was decided to exclude tests such as a
regression analysis.
Beforehand, we considered performing an experimental study in addition to the questionnaire,
where respondents were to be asked about celebrities and brands exclusively. Separate pictures
of the celebrity and the products would be displayed individually in order to obtain non-biased
results by removing the influence from one to another. This would provide a different point of
view where the respondents would have the possibility to express themselves in their own words
without the affect of pre-made questions. The addition to the study by performing this would be
very valuable in the sense that it could provide a deeper insight and depth. This thought had to be
disregarded due to lack of knowledge and means needed for performing this type of study.

Page | 39

CHAPTER 5 RESEARCH FINDING AND DATA ANALYSIS

The research finding and data analysis will be represented and explained in this chapter. It is
not only support the research problems of this research project, but also used in further
discussion for managerial recommendations and the final conclusion. The data analysis is
mainly focused on the research problem and questions. Here, the authors will also setup
hypotheses developed from the main research questions in order to answer the questions.

5.1 Demographical Data Analysis

This section covers a presentation and brief analysis of the different demographical factors
considered to be important in the study. Among these we find age, type of cosmetics purchased,
purchase frequency, spending and sources (where the customers got to know about the products).
5.1.1 Age Groups
100%

80%

60%
44%
40%

30%

20%

10%

9%

7%

30-34

35-39

40 and above

0%
20-24

25-29

Figure 8 Age Groups of the Respondents

The age groups included in the study and the questionnaire ranged from 20-24 to 40 and
above. As can be seen from Figure 8, the majority, 74 percent, of the respondents were between
20 and 30 years old. In addition, a total of 26 percent can be found in the categories representing
30 and above.
Women between 20 and 30 are an important segment when investigating cosmetics and its
related purchasing pattern and other issues. The reason stems from this specific category
containing a large number of existing and potential customers, for every different type of product
categorized within cosmetics, with purchasing power. Additionally, cosmetics advertisements are
very often directed towards this specific segment.

Page | 40

5.1.2 Types of Cosmetics Purchase

60%
52%
50%
41%
40%
30%
20%
7%

10%
0%
Makeup

Skincare

Fragrance

Figure 9 Types of Cosmetics Most Frequently Purchased by Respondents

Figure 9 displays that the types of cosmetics that are most frequently purchased are skincare,
with 52 percent of the respondents purchasing some sort. This category includes cleansers, toners,
day/night cream and other skincare products. The second most common type is makeup, where
41 percent of the respondents are purchasing this product most frequently. The reason for
fragrance being the least frequently purchased product might be due to the fact that it often is the
most expensive product, but also the product lasting the longest.
5.1.3 Purchase Frequency

Less often

13%

Every three months

30%

Every two months

29%

Once a month

28%
0%

5%

10%

15%

20%

25%

30%

35%

Figure 10 Respondents Purchase Frequency

Figure 10 displays the respondents purchase frequency of cosmetics products, where the choices
vary from once a month, through every two months, through every three months, to less
often. There are only small differences between the options, where it differs one percent
Page | 41

between once a month and every two months, and also only one percent between every two
months and every three months. The least common purchase frequency is less often, where
only 13 percent of the respondents are present.
If a consumer is using cosmetics on a regular basis it takes a certain amount of time before using
up the product, depending on how much they use at every occasion. If the consumer is a heavy
user they might have to purchase a new product every month or perhaps every two months,
depending on what type it is. However, it should be mentioned for clarification that the time of
purchase is a relevant factor as well, since the consumer might need a new makeup product at
one point, a new skincare product at another and vice versa.
5.1.4 Spending
37%

27%
20%
16%

Under 100

100-199

200-299

300 and above

Figure 11 Spending on Each Purchase

Figure 11 represents the spending pattern, and displays that the most common spending amount
during each purchase opportunity is under 100SEK, represented by 37 percent. The next
category is declared to be the second most common alternative with 27 percent of the
respondents. It can also be seen that the category of 300SEK and above is more frequent, with its
20 percent, in comparison to the 200-299SEK option.
The price range of products differ depending on the brand, but when looking at the most
frequently appearing brands in cosmetics commercials using celebrity endorsement, the prices
are somewhat similar. A mascara or foundation (two common products in advertising) usually
has a price range between 79 and 199SEK, covering what many of the respondents usually spend
on one occasion.

Page | 42

5.1.5 Sources
41%
30%

14%
6%

7%

2%

Figure 12 Sources of Getting to Know Cosmetic Brands

Figure 12 displays that the most common source of information where the respondents find out
about cosmetics products is TV advertisements, followed by magazines. The least common
source, with a total of only 2 percent of the respondents, is newspapers, followed by internet and
in shops with 6 and 7 percent each.
Cosmetics advertisements are commonly seen on TV and in magazines, reaching the segments
the ad is directed towards more easily, in comparison to, for example a newspaper ad. Hence, the
result of most people getting to know about a product through these sources, proving this type of
advertisement being effective when wanting to get more people to become familiar with the
product and the brand.

5.2 Reliability Test

In order to measure to what extent the items used to measure the concepts correlate with each
other, the reliability test is necessary. As stated earlier, a value of 0.7 should be achieved in order
to be considered reliable. To test the variables used in the survey, four main variable groups were
tested by using the Cronbachs Alpha reliability test. The results show that the variables are
reliable in terms of measuring each corresponding underlying concept, where a value of 0.7 is
achieved in each variable group.

Page | 43

Variables

Cronbach's
Alpha

Cronbach's Alpha Based on

Standardized Items

N of
Items

Behavioral Loyalty

.762

.778

Attitudinal Loyalty

.723

.716

Attitudes toward
Ads

.727

.709

Factors influence
purchasing

.748

.733

14

Table 1 Cronbachs Alpha Test

As Table 1 shows, the different variable groups have met the value of 0.7. Each one of the nine
items used to measure behavioral loyalty and attitudinal loyalty are added together by using
computing tools. In this way, it is more convenient for later analysis. After computing, each of
the nine items will become a new variable with a new name (Behavioral Loyalty and
Attitudinal Loyalty) in SPSS.

5.3 ANOVA Analysis

The one way analysis of variance (ANOVA) is employed to test the four different hypotheses
stated earlier. ANOVA is used to compare the variance of scores between tested groups with
their variance within the groups. When the variance within the compared groups is large, usually
a 0.05 significant level is applied, the groups are different. Otherwise the results are not
statistically significant. The one-way ANOVA usually consists of one independent and one
dependent variable (Kent, 2007, p. 385).
Four hypotheses will be tested:
H1: There is no significant relationship between the usage of endorsers in cosmetics
advertisements and behavioral loyalty.
Here, factor 3, the importance of the usage of endorsers in cosmetics advertisements for the
influence on purchasing decisions and the new behavioral loyalty variable is used to test
whether there is a statistical relationship existing between these two variables. The dependent
variable is behavioral loyalty and the independent variable is brand endorsers.
Source
Corrected Model
Intercept
factor3
Error
Total
Corrected Total

Type III Sum of Squares

df

Mean Square

484.952a

21760.782

484.952

161.651

2752.758

96

28.675

161.651

Sig.

5.637 .001

21760.782 758.888 .000

5.637 .001

46211.000 100
3237.710

99

a. R Squared = .150 (Adjusted R Squared = .123)

Table 2 Dependent Variable: Behavioral Loyalty

Page | 44

Table 2 shows a result of the significance level being 0.001< 0.05. This means that there is
statistical significance (.001) of differences between the brand endorsers and behavioral loyalty.
Therefore, hypothesis H1 has to be rejected. Brand endorsers in cosmetics advertisement have an
influence on behavioral loyalty in purchasing situations.
H2: There is no significant relationship between the usage of famous endorsers in cosmetics
advertisements and behavioral loyalty.
Factor 4, the importance of the usage of famous endorsers in cosmetics advertisements for the
influence on purchasing decisions and the new behavioral loyalty variable is used to test the
hypothesis. The independent variable is famous endorsers and the dependent variable is
behavioral loyalty.

Source

Type III Sum

of Squares

df

Mean Square

Sig.

358.540a

119.513

3.985

.010

18554.123

18554.123

618.649

.000

358.540

119.513

3.985

.010

Error

2879.170

96

29.991

Total

46211.000

100

3237.710

99

Corrected Model
Intercept
factor4

Corrected Total

a. R Squared = .111 (Adjusted R Squared = .083)

Table 3 Dependent Variable: Behavioral Loyalty

Table 3 provides results showing that the significance level is 0.010< 0.05. This means that there
is a statically significant (.010) difference between the famous endorsers and behavioral loyalty.
Therefore, hypothesis H2 also has to be rejected. Famous endorsers in cosmetics advertisements
do have an influence on behavioral loyalty.
H3: There is no significant relationship between the usage of brand endorsers in cosmetics
advertisements and attitudinal loyalty.
This hypothesis is designed to test the relationship between brand endorsers and attitudinal
loyalty. Firstly, factor 3, the importance of the usage of endorsers in cosmetics advertisements
for the influence on purchasing decisions and the new attitudinal loyalty variable is used to test
whether there is a statistical relationship existing between these two variables. The dependent
variable is attitudinal loyalty and the independent variable is brand endorsers.

Page | 45

Source

Type III Sum

of Squares

df

Mean Square

Sig.

91.883a

30.628

1.055

.372

50715.269

50715.269

1746.872

.000

91.883

30.628

1.055

.372

Error

2787.077

96

29.032

Total

102230.000

100

2878.960

99

Corrected Model
Intercept
factor3

Corrected Total

a. R Squared = .032 (Adjusted R Squared = .002)

Table 4 Dependent Variable: Attitudinal Loyalty

Table 4 shows the result from the ANOVA analysis that there is no statistically significance
since .372> .05. It means that the Null hypothesis cannot be rejected. So, there is no significant
relationship between using brand endorsers and attitudinal loyalty.
H4: There is no significant relationship between using famous endorsers in cosmetics
advertisements and attitudinal loyalty.
The last hypothesis is designed to test the relationship between using famous endorsers and
attitudinal loyalty. First, the factor 4 which how important using famous endorsers in cosmetics
advertisements influence your purchase decision and the new attitudinal loyalty variable is used
to test whether there is a statistical relationship existing between these two variables. The
dependent variable is: attitudinal loyalty. The independent variable is famous endorsers.
Type III Sum of
Source

Squares

df

Mean Square

Sig.

128.888a

42.963

1.500

.220

44376.007

44376.007

1549.085

.000

128.888

42.963

1.500

.220

Error

2750.072

96

28.647

Total

102230.000

100

2878.960

99

Corrected Model
Intercept
factor4

Corrected Total

a. R Squared = .045 (Adjusted R Squared = .015)

Table 5 Dependent Variable: Attitudinal Loyalty

Table 5 demonstrates the result from the ANOVA analysis, showing that there is no statistical
significance, since 0.220> 0.05. It means that the hypothesis cannot be rejected. Hence, there is
no significant relationship between the usage of famous endorsers and attitudinal loyalty.

Page | 46

Two hypotheses concerning behavioral loyalty are rejected because of the statistically significant
differences. But the other two hypotheses, related to attitudinal loyalty cannot be rejected
because there are no statistically significant differences. The results elucidate that using
endorsers or famous endorsers in cosmetics advertisements do influence the respondents buying
decisions. To further test the strength of the relationship between using brand endorsers and
famous endorsers to behavioral loyalty and attitudinal loyalty, a correlation analysis will be
provided in section 5.4.
In addition, the data findings and analysis illustrate that the usage of brand endorsers or celebrity
endorsers in advertisements are not the reason leading to an attitudinal loyalty towards cosmetics
brands. Other factors are more important to take into consideration when making purchase
decisions. This study provides evidence that respondents are less likely to take famous endorsers
into consideration for their decisions. 36 percent of the respondents would buy brands often
promoted in advertisements, but only when they are not familiar with the brands offered in shops.
42 percent of the respondents tend to stick to brands they have previously been using instead of
switching to other brands. 74 percent of the respondents show that they would buy a brand even
though it does not use famous endorsers. The results also show that 58 percent will buy other
products of the same brand as they are already familiar with. And lastly, most of the respondents
(70 percent) agree that they will buy a brand recommended by an unknown spokesperson in
advertisements as well.

5.4 Correlation Analysis

Although both correlation and regression techniques concern about the relations between
variables, they are used for different proposes. Correlation analysis is employed when
researchers need to find out how strong of the relationship between variables, whereas regression
analysis is used to determine the pattern of the relationship between variables. (Kent, 2007, p.
364)
To further determine the strength between variables tested previously, the Pearson correlation
analysis is given here.

Brand endorsers and behavioral loyalty
The fist one is to test the correlation of the factor 3 which how important using brand endorsers
in cosmetics advertisements influence your purchase decision and the behavioral loyalty
variables. The result shows below in table 6.
Behavioral
loyalty

factor3
factor3

Pearson Correlation

Sig. (2-tailed)

.000

N
Behavioral Pearson Correlation
loyalty
Sig. (2-tailed)

.363**

100

100

.363**

1.000

.000

N
100
**. Correlation is significant at the 0.01 level (2-tailed).

100.000

Table 6 Brand Endorsers and Behavioral Loyalty

Page | 47

Based on the results, the correlation between the usage of brand endorsers and behavioral loyalty
appear to be positive. As to the strength of the relationship, Cohen (1983) suggested that an rvalue from 0.30 to 0.49 can be considered to show a medium strength relationship between the
variables. The maximum strength is 1 or -1, depending on the direction of the relationship. The
result shows that the r-value is 0.363 between usage of brand endorsers and behavioral loyalty
variables. Thus, the relationship is medium strong, where there also is statistical significance
between the variables (.000).

Famous endorsers and behavioral loyalty
Behavioral
loyalty

factor4
factor4

Pearson Correlation

Sig. (2-tailed)

.326**
.001

N
Behavioral Pearson Correlation
loyalty
Sig. (2-tailed)

100

100

**

1.000

.326

.001

N
100
**. Correlation is significant at the 0.01 level (2-tailed).
Table 7 Famous Endorsers and Behavioral Loyalty

100.000

The results shown in Table 7 concern the correlation of factor 4, the importance of the usage of
famous endorsers in cosmetics advertisements for the influence on purchasing decisions and the
behavioral loyalty variables. Based on the results, the correlation between using famous
endorsers and behavioral loyalty appear to be positive as well. Although the correlation seems to
be smaller (.326) compared to using brand endorses, the result is considered to have a medium
strong relationship between the variables. Also, there is significant difference between the
variables at the .01 level.
Table 8 and Table 9 display the result of the relationship of the usage of brand endorsers and
attitudinal loyalty, as well as famous endorsers and attitudinal loyalty.

Brand endorsers and the attitudinal loyalty
The result in table 8 shows that the correlation between using brand endorsers and attitudinal
loyalty is negative (-.071), there is also no statistical significance between the variables. The
correlation is in the range from 0.10 to 0.29, it is therefore considered to be a small negative
relationship between the variables.
factor3
factor3

Pearson Correlation

Sig. (2-tailed)
N
Attitudinal Loyalty

Pearson Correlation

Attitudinal Loyalty

-.071
.480

100

100

-.071

1.000

Sig. (2-tailed)

.480

100

100.000

Table 8 Brand Endorsers and Attitudinal Loyalty

Page | 48

Attitudinal
Loyalty

factor4
factor4

Pearson Correlation

-.064

Sig. (2-tailed)
tailed)

.529

N
Attitudinal Loyalty Pearson Correlation
Sig. (2-tailed)
tailed)

100

100

-.064

1.000

.529

N
100
Table 9 Famous Endorsers and Attitudinal Loyalty
L

100.000

Table 9 shows the correlation to be negative even here, and there is no statistical significance
between the variables. However, there is a small negative ((-.064)
.064) relationship between using
famous endorsers and attitudinal loyalty.

5.5 Descriptive Analysis

In this section, descriptive statistics regarding the attitudes and factors included and investigated
in the study will be presented. The results from the questionnaire have been processed in SPSS
and provided different interesting numbers, where the mean is the most important
impor
number for the
research question.

4.13
3.84
3.65
3.12
2.84

Attads1

Attads2

Attads3

Attads4

Attads5

Figure 13 The Means of Attitudes about Cosmetics Advertisements

A

Looking at the means for attitudes about cosmetics advertisements, it can be seen that the
respondents tend to agree with number two, three and five. These are: exaggerated product effect,
too much use of celebrities and all ads convey the same message. Attitude one and five, too
much irrelevant information and message in ads is unclear, are mostly perceived to be neutral or
Page | 49

more directed towards disagreeing. It can be seen from these results that the cosmetics industry is
perceived to very often exercise the use of celebrities in advertisements. The appearance of
celebrities in many different commercials for different brands might cause a sense of feeling of
overuse of this type of endorsement. This can also be linked to the attitude all ads convey the
same things, since brands are using the same type of advertising technique, where the message
often is similar or the same. The different commercials often have an exaggerated view of its
products effect, such as super long eyelashes, lipstick that last all day, hairspray that keep your
hair the same no matter what etc., which most of the time is not the case. But, it can still
stil have a
better effect when comparing it to competitors products. Most of the time the customer knows
about this, but they still buy the products, even though they do not fully believe the message.
All of these three factors are linked together, in the sense
sense that one perception might give another.
If you are of the opinion cosmetics brands are overusing celebrity endorsement in their
advertisement, you might also presume that the product effect is exaggerated. This is because
celebrities look stunning most
most of the time at official appearances, due to makeup artists,
hairdressers and stylists, something the average person does not have access to. In addition, due
to the similar messages different brands convey, it is hard to differentiate between them, for
example
xample knowing exactly which brand that promised both thick and long lashes.

4.26

3.84

3.93

3.86

3.73
3.59

factor1

factor6

factor7

factor8

factor9

factor14

Figure 14 The Means of Important Factors Influence Cosmetics Purchasing Decisions

D

Out of the fourteen different factors, five of them have a mean around 4, which is important.
This suggests that these factors are considered to be important when purchasing cosmetics
products. The factors mentioned are; reputation of the brand which is factor 1 (3.73), reasonable
price factor 6 (4.26), friends recommendations factor 7 (3.84), previous purchase experience
factor 8(3.93) and special benefits factor 14 (3.86) (organic, allergy tested etc.).
None of the factors with the highest means (the ones considered to be the most important when
purchasing cosmetics) are related to advertising or endorsement, but only to experience, the
product itself and what opinion others have concerning the product. It can be seen from these
results that customers value their own and others experience with a brand and their perception of
Page | 50

what a brand stands for the highest. However, it should be mentioned that a brands reputation is
created both due to the quality of their products, but also by projecting themselves
themselve in a specific
way in media and through advertisements. Thus, the use of a famous endorser in advertisement
might not have a direct effect on customers, but are important for the brands entirety.
3.04
2.52

2.73

2.63

2.17

2.07

BhaL1

BhaL2

BhaL6

BhaL7

BhaL8

BhaL9

Figure 15 The Means of Behavioral Loyalty Variables

V

Furthermore, we present several behavioral loyalty variables which have the highest means
among the nine different variables. As can be seen in Figure 15, behavioral loyalty variable
(BhaL)) 1, If I am not familiar with cosmetics brands, I choose to buy brands often promoted in
advertisements has the highest mean (3.04), indicating that respondents rely on advertisements
information in an unfamiliar purchase situation. BhaL8, association between the endorsing
celebrity and the brand has a mean of 2.73 and BhaL6, celebrities help me to get to know a
brand better has a mean of 2.63. Furthermore, BhaL2, no brand preference has a mean of 2.52,
whiles BhaL9, only celebrities can represent the substance has a mean of 2.17. Lastly, BhaL7,
celebrities as a guarantee for the quality of a brand has a mean of 2.07. All these variables have
an average mean above 2, which is represented by Disagree, showing that the respondents
express their negative opinions regarding these statements.

Page | 51

4.05
3.97

3.58
3.5
3.38

3.39

3.38

AttL2

AttL3

AttL5

AttL6

AttL7

AttL8

AttL9

Figure 16 The Means of Attitudinal Loyalty Varaiables

In Figure 16, the mean of attitudinal loyalty variables are presented. Attitudinal loyalty (AttL)
variable 6, I would buy a brand regardless whether a celebrity endorses that brand or not has
the highest mean of 4.05, which indicates Agree. Followed by AttL9, unnecessary to use
celebrities in advertisements with a mean of 3.97. Respondents indicate that they would buy a
brand even if it was recommended by an unknown spokesperson in advertisements. All the other
variables; AttL2, stick to a brand I usually buy rather than trying an unfamiliar brand, AttL3
seldom switch brands, AttL5 would not switch brand because of lack of knowledge, AttL7
If I find a brand I like, I will always buy products of that brand and AttL8 buying a cosmetics
brand is not all about celebrities I like have an average mean of 3.45. This shows that the
respondents have positive opinions regarding information search before buying cosmetics
products, sticking to cosmetics brands they are familiar with, and it also claims that the use of
celebrity endorsers in advertisements might be somewhat unnecessary.

5.6 Conclusion
The results presented in this chapter are all generated through data from the web-based survey.
Various statistical analysis skills are employed to predict patterns and relationships of the data.
In this chapter, both descriptive and an inferential statistics approach are used to analyze the raw
data collected from the survey. Descriptive statistics do exactly what they say (Cohen, Manion,
& Morrison, 2007, p. 503). They are often used in summarizing frequencies, including the mode,
the mean, the median, minimum and maximum scores etc. (Cohen, Manion, & Morrison, 2007, p.
503) Several data findings presented in this chapter are analyzed by using descriptive statistical
skills. It is necessary to detect a general pattern of the sampling, such as belonging to which age
group, what type of cosmetics often bought and amount spent on each purchase. Without
knowing these factors, we are less likely to understand the demographical pattern of the data. It
would also make it difficult to give constructive recommendations in relation to the rest of the
findings. Nevertheless, using frequencies and descriptive statistics speaking for themselves is
important in academic researches.
Page | 52

Although the descriptive data can be used to portray the truth in a simple way, such statistical
skills would not be able to assist researchers in making inferences or predictions. Therefore,
using inferential statistical analyses will help to see inferences and predictions based on the data
collected. In striving to present the data in a rigorous way, several inferential statistics were also
included in this chapter, including hypotheses testing by one way analysis of variance and
correlations testing (Cohen, Manion, & Morrison, 2007, pp. 503-504). To test the four
hypotheses, we chose to use the null hypothesis by applying inferential statistics to indicate
whether a statistical significance existed. Kirk (1999) suggested that a statistically significant
result is one for which chance is an unlikely explanation. Using hypothesis testing is essential in
quantitative research. In such a research as carried out in this thesis, a significance level of 0.05
is often the boundary of whether to accept or to reject a null hypothesis (Cohen, Manion, &
Morrison, 2007, pp. 515-516). The ANOVA was used to test four given hypotheses based on the
research question. After presenting and explaining the results of the ANOVA analysis, a
correlation analysis was carried out, aiming to identify the strength between each of two sets of
variables (Cohen, Manion, & Morrison, 2007, pp. 531-538). Doing so enhanced a further
indication of to what extent the usage of brand endorsers/celebrity endorsers are having an effect
on or no effect on respondents. A detailed discussion concerning the results of this undertaken
research in combination to prior studies can be found in chapter 6.
All discussions and data analysis only represent the response generated from the samplings
collected for the purpose of the study. We do not attempt to overlook the limitations of the study.
In light of the chosen theories, the data is correspondently collected for the purpose of supporting
this research.

Page | 53

CHAPTER 6 DISCUSSION AND FURTHER RESEARCH

This chapter presents a conclusion to the research problem in relation to previous studies made
within similar fields. Managerial recommendations will be presented in order to provide an
insight regarding other possible courses of action for managers within the cosmetics and
advertising field. Moreover, further research directions are at hand to propose additional and
different ways of investigating the subject.

6.1 Discussion
We have given a detailed discussion in the empirical finding analysis related to our research
questions and our topic. In short to answer our research questions clearly, using celebrity
endorsers in cosmetics advertisements has very little influenced on our respondents brand loyalty
behaviors, mostly it influenced our respondents behavioral loyalty under certain occasions.
Especially if respondents have no brand preferences or brand awareness in mind, they will
mostly choose the brands often promoted in advertisements. As price is also an important factor
influencing purchase decisions, respondents tend to react in a behavioral loyalty way to the
question which promotional offering is important as well. But celebrity endorsers have no
influence on attitudinal loyalty of the respondents according to the analysis. Other factors have
more important influence upon forming attitudinal loyalty, such as friends recommendations,
reasonable prices, special benefits etc.
In this study, brand loyalty is particularly measured by both behavioral loyalty and attitudinal
loyalty. It is the aim of the study to identify whether using famous endorsers in cosmetics
advertisements has an influence on behavioral loyalty or attitudinal loyalty. The survey was
conducted based on brands often promoted in advertisements in general, rather than specific
brands. Prior researches mostly focused on particular brands, such as Crest Original and Crest
Tartar, two market-leading toothpaste brands from Procter & Gamble (Bandyopadhyay &
Martell, 2007). Another study let respondents list cosmetics in preferred order and the researches
picked up top two or three of those brands to make a survey (Guthrie & Kim, 2009). As
discussed earlier, there is no pre-determined loyalty in the survey, and all respondents are
assumed to have either behavioral loyalty or attitudinal loyalty. In prior researches, brand
attitudes (Guthrie & Kim, 2009) or attitudinal loyalty (Bandyopadhyay & Martell, 2007) were
the perquisite of the studies. 100 females were selected to be the sampling units in this study,
which can be considered a sufficient sampling size. Prior research had rather large sampling
sizes, usually more than 200 and up to 1000 respondents. The findings of this research confirm
that a consumers behavioral response towards cosmetics brands does not necessarily lead to
attitudinal loyalty. Neither brand loyalty nor attitudinal loyalty is proved to be created by brand
endorsers or famous endorsers. However, brand endorsers help respondents in getting to know a
brand better when they are confronted with unfamiliar brands. Moreover, this research further
supports the suggestion that the concept of loyalty to be considered should be dependent on
both market type and situations (Thiele & Bennett, 2001, p. 25). However, many prior studies
prefer to measure brand loyalty in a uniform way. This equates to brand loyalty measuring based
on market types is necessary because all market types do not have the same way to capture or
measure brand loyalty (Thiele & Bennett, 2001). A similar study, pinpointing both behavioral
loyalty and attitudinal loyalty in relation to cosmetics ads has not yet been conducted.

Page | 54

The survey used in this research included nine behaviorally constituted questions and nine
attitudinal questions to measure respondents behaviors and attitudes concerning endorsers in
cosmetics advertisements. Prior studies set up eight behavioral questions and 30 attribute-related
questions to test the respondents usage patterns, their satisfaction and beliefs about selected
brands in the test (Bandyopadhyay & Martell, 2007). Although this research does not categorize
users, the survey was conducted only among females whom well reflect the main consumption
groups of cosmetics. Although, Thiele and Bennett (2001) suggested in their prior research that
using only behavioral loyalty was enough to measure the consumable goods market since it was
rather stable and unlikely to change. Consumers have low involvement in the consumable goods
market and purchasing was usually the outcome of habitual behavior. Therefore, attitudinal
loyalty measuring was not essential (Thiele & Bennett, 2001, pp. 29-30). Different from Thiele
and Bennetts suggestion, respondents in this survey present their high interest of involvement in
gathering information about cosmetics brands before buying it. Promotion offering factors have a
lower mean than other important factors for influencing the respondents cosmetics purchase
decisions and a majority of the respondents consider a reasonable price factor to be important as
well. The data findings in this research positively confirm that both behavioral and attitudinal
loyalty should be included in measuring real brand loyalty in the cosmetics products category. At
the end of the survey, respondents were asked to give opinions ranging from very unimportant to
very important on factors influencing their cosmetics buying decisions. 14 factors were provided,
and few respondents indicated that brand endorsers or famous endorsers have an influence in
making purchase decisions. However, other factors related to brand reputation; previous buying
experiences and friends recommendations were indicated to be important. Using brand
endorsers might help brands to build up brand awareness and get attention in a proper way. Too
much is dependent on the celebrity endorsers and they do not necessarily lead consumers to a
behavioral or attitudinal loyalty. Instead, cosmetic marketers should focus on other more
important factors instead of the perfection effect of their products. Moreover, the negative
information of celebrity endorsers would have an impact on consumers perceptions on the
endorsed brands (White, Goddard, & Wilbur, 2009). Therefore, marketers should use celebrity
endorsement technique wisely.
Furthermore, Baldinger and Rubinson (1996) proposed several suggestions in their experiment
Brand builder, based on their experimental results; a large shared brand is likely to have more
loyal buyers but not necessarily to retain them at a higher rate over time. This is due to
behavioral loyalty dominating consumers when making purchase decisions. When both behavior
and attitudes were used to measure loyalty, two results were generated, high-loyal buyers who
have consistent attitudes tend to stay loyal to the same brand. Here both behavioral and
attitudinal loyalty was both the same. The second result: high-loyal buyers who have
inconsistent attitudes tend to switch away. (Baldinger & Rubinson, 1996, p. 32) In this case
behavioral and attitudinal loyalty was not the same. To marketers, behavioral buyers would help
in order to generate profit. But only consumers who have both behavioral and attitudinal loyalty
are truly valuable, because they will not switch to another brand because of external variable
changes, unless internal conditions would change (Baldinger & Rubinson, 1996). Their study
positively supports this research to some extent. The respondents of this study expressed their
behavioral responses to brand endorsers/celebrity endorsers. However, there were not
correlations between attitudinal loyalty and brand endorsers/celebrity endorsers.

Page | 55

This research study has outlined issues for marketers who are seeking and trying to understand
what brand loyalty is about and how to attain it and keep it. The study brings up the question of
whether heavy usage of celebrity endorsers in cosmetics advertisements could help marketers to
build up brand loyalty or not. Also, brand loyalty is the main thread of the entire research.
Theoretically, the real brand loyalty should be measured from both behavioral and attitudinal
aspects, as they both are complementary elements to measure brand loyalty. However, due to
many constrains, marketers sometimes neglect both aspects but simply treating behavioral
loyalty as the real loyalty. As presented in the research findings, using brand endorsers or famous
endorsers do have an effect on consumers behavioral loyalty but not on their attitudinal loyalty.
Brand loyalty is not only important to marketers, but also important to consumers. It provides
marketers competitive positions in markets, and it saves consumers time and resources to search
for alternatives when making purchase decisions. Behavioral loyalty might easily change if there
are more competitive brands or products offered on the market. Consumers will be more
committed to brands if they have both behavioral and attitudinal loyalty, since markets are
constantly changing because of new technologies and other reasons. Thereby, consumers are
facing more and more choices when making purchase decisions. For marketers, it is imperative
to win the marketing battles not only concerning behavioral loyalty, but also attitudinal loyalty as
it is a commitment of the consumers towards brands. Advertising provides a platform between
marketers and consumers to send and receive information about brands. What matters is
imperative is the quality and not the quantity of advertisements. As the managerial
recommendations provided in Section 6.2 state, there are many other ways marketers can use for
creating brand loyalty and interact with consumers.

6.2 Managerial Recommendations

There are a number of different ways to approach a problem from a managerial point of view
when considering the research problem discussed in this study. Therefore, this part will cover
recommendations for different managerial courses of action related to brands and its marketing
techniques.
As found in the study, customers generally do not value a brand endorser more than the brands
product itself. Usually, the reputation of the brand, previous experience and friends
recommendations are of greater importance when one is deciding what product to buy or not.
Due to these findings, one recommendation is the use of real life examples with regular people,
where they get tips from professionals regarding the best products for their specific needs and
problems. A similar idea is present in a Chinese TV show, Women Queen, where several women
participate and get individual help by professional makeup artists. They reveal their problems
and get matched with the best product/s for their needs, thereby also helping other women
watching the show who find themselves with similar problems. Engaging in a different type of
marketing technique for cosmetics helps a brand to differentiate themselves as most of todays
cosmetics brands use a similar type of advertisement.
Additionally, differentiating advertising and its techniques in comparison to competitors is a
great possibility for managers within this industry. Most cosmetic companies are using the same
or similar type of advertising and the amount of money spent on the use of celebrities in
advertisements is enormous. One suggestion is to reallocate these resources towards the
consumers and their specific needs instead, and possibly use more regular people as
Page | 56

spokespersons. By doing this, the average consumer might find it easier to relate to this person
and believe the message transmitted. The monetary resources saved just by not having a celebrity
as an endorser or spokesperson are a great source for further research in order to find the best
products with the best match for customers needs.

6.3 Contribution
This research has shed new light on the usage of celebrity endorsement in cosmetics
advertisements. In the study, a detailed definition of brand loyalty is extensively discussed,
including both behavioral loyalty and attitudinal loyalty. Based on the data findings, research
inferences by hypotheses through statistical software were tested. The results are not only useful
in providing constructive managerial recommendations, but also to pinpoint a new path for latter
research in this area. Our data findings negate the high costs of celebrity endorsements as it is not
effective in creating attitudinal loyalty, although behavioral loyalty is widely accepted as the
definition of brand loyalty. We firmly believe that consumers are truly committed to a brand only
when both behavioral and attitudinal loyalty is present. Although the data findings cannot speak
for all voices of consumers, they still represent a voice of the respondents at the least. We have
never doubted the deep-rooted influence of celebrities in any cases. However, marketers should
not put too much focus on the halo effect of celebrities. More efforts could be on other aspects in
the process instead. Moreover, our research does not attempt to overlook the efficiency of
advertising, neither the efficiency of using celebrity endorsers. The research is merely trying to
raise a suggestion from a consumers point of view for marketers. It is not only trying to help
marketers to improve branding and marketing strategies, but as reciprocated benefits to
consumers themselves as well.

6.4 Limitations and Further Research

With the present research, the participants used in the survey were mainly students ranging
within the ages between 20 and 30 years old. Using students as primary samples has certain
limitations on external validity. It might also limit the generalizability of the results. Validity
issues should also not be neglected when using students as samples when comparing mean
differences in variables, because students usually have certain similarities in common compared
to a general sample (Silvera & Austad, 2003). The representativeness of the data is limited since
the sample is based on the snowball and volunteer sampling methods. Other limitations have
been discussed in the data analysis and conclusion discussion. To summarize them, they are
concerned with cultural differences which might lead to different results, as well as the product
categories being too broad to assign attributes or characteristics.
Although this research has been able to provide contributions in identifying real brand loyalty in
association with celebrity endorsement in cosmetics advertisements, the process of the research
and data collection are rather simplified because of resource constrains and other limitations. The
generalization and representativeness of the data presented is rather limited due to the
characteristics of the sampling. Therefore, a further empirical research in this topic is
recommended which evaluates the behavioral loyalty and the attitudinal loyalty separately,
regarding the use of famous endorsers in cosmetics advertisements. A probability sample would
help researchers to overcome the disadvantages of nonprobability samples. To achieve better
results, researchers could ask respondents to rank cosmetics brands and then select a sufficient
Page | 57

number of brands to be included in a survey. In such a way, researchers could better assign
behavioral and attitudinal attributes to target brands. To respondents, this would result in less
ambiguous decisions and opinions because they might already have clear impressions in mind.
User groups could be detailed categorized as well, such as high-end brand oriented, priceoriented and function-oriented, as different users are dominated by either behavioral loyalty or
attitudinal loyalty. By using focus groups, researchers would get a clearer picture concerning
consumers opinions, attitudes and attentions. Furthermore, cultural background might play a key
role in a cosmetics purchasing decision as well, since many cultural differences exist.
Aforementioned aspects are all suggestions for new research opportunities related to the
effectiveness of celebrity endorsement and brand loyalty towards cosmetics products.

Page | 58

REFERENCES
Aristotle. (n.d.). Aristotle quotes. Retrieved April 2, 2010, from Thinkexist.com:
http://thinkexist.com/quotation/personal_beauty_is_a_greater_recommendation_than/180845.html
Asch, S. (1952). Social psychology. Englewood Cliffs, NJ: Prentice-Hall.
Baldinger, A., & Rubinson, J. (1996). Brand Loyalty: The Link Between Attitude And Behavior. Journal
of Advertising Research , 36(6), 22-34.
Bandyopadhyay, S., & Martell, M. (2007). Does attitudinal loyalty influence behavioral loyalty? A
theoretical and empirical study . Journal of Retailing and Consumer Services , 12(5),35-44.
Brock, T. C. (1965). Communicator-Recipient Similarity and Decision Changes. Journal of Personality
and Social Pyschology , 1(6),650-654.
Bryman, A., & Bell, E. (2007). Business Research Methods Second Edition. Oxford University Press.
Byrne, A., & Whitehead, M. (2003). The naked truth of celebrity endorsement. British Food Journal ,
105(4/5),288-296.
C.Whan, P., & V.Parker, L. (1977, September). students and Housewives: Differences in Susceptibility to
Reference Group Infiuence. Journal of Consumer Research 4 , pp. 4(2),102-10.
Chaudhuri, A., & Holbrook, M. B. (2001). The Chain of Effects From Brand Trust and Brand Affect to
Brand Performance: The Role of Brand Loyalty. Journal of Marketing , 63(4),81-93.
Coca-Cola, T. (n.d.). Coke Lore. Retrieved April 8, 2010, from The Coca-Cola Company:
http://www.thecoca-colacompany.com/heritage/cokelore_newcoke.html
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral
sciences 2nd Edition. New York: Erlbaum .
Cohen, L., Manion, L., & Morrison, K. (2007). Research Methods in Education, Sixth Edition. Bodmin:
Routledge.
Creswell, J. W. (2003). Research Design, Qualitative, Quantitative, and Mixed Methods Approaches.
California: Sage Publications Inc.
Dick, A. S., & Basu, K. (1994). Consumer loyalty: towards an integrated conceptual approach. Journal of
the Academy of Marketing Science , 22(2),99-113.
Fishbein, M., & Ajzen, I. (1975). Belief, Attitude, Intention, and Behavior:An Introduction to Theory and
Research . Massachusetts: Addison-Wesley.
Fisher, A. B. (1985, August 5). Coke's Brand Loyalty Lesson. Retrieved April 8, 2010, from Fortune:
http://money.cnn.com/magazines/fortune/fortune_archive/1985/08/05/66245/index.htm
Friedman, H. H., & Friedman, L. (1979). Endorser Effectiveness by Product Type- Celebrities worked
best with expectations. Journal of Advertising Research , 19,63-71.
Page | 59

Graziano, A. M., & Raulin, M. L. (2010). Research Methods, A Process of Inquiry Seveth Edition. Boston:
Pearson Education Inc.
Guthrie, M. F., & Kim, H.-S. (2009). The relationship between consumer involvement and brand
perceptions of female cosmetic consumers. Journal of Brand Management , 17(2),114-133.
Heidelberg, S. B. (2006, May 27). Definition of Cosmetics. Retrieved April 26, 2010, from SpringerLink:
http://resources.metapress.com/pdf-preview.axd?code=w50p665l28jw27rt&size=largest
Jacoby, J., & Kyner, D. B. (1973). Brand Loyalty Vs. Repeating Purchase. Journal of Marketing
Research , 10(1),1-8.
John G, L. J., & Thomas K, S. (1982). Memory and Attentional Factors in Consumer Choice: Concepts
and Research Methods. Journal of Consumer Research , 9(6),18-37.
Jones, E. E., David E., K., Harold H., K., Nisbett, R. E., Valins, S., & Weiner, B. (1972). Attribution:
Perceiving the Causes of Behavior. General Learning Press.
Kahle, L. R., & Homer, P. M. (1985). Physical Attractiveness of the Celebrity Endorser: A Social
Adaptation Perspective. Journal of Consumer Research , 11,954.
Kahle, L. R., Kulka, R. A., & Klingel, D. M. (1980). Low Adolescent Self-Esteem Leads to Multiple
Interpersonal Problems: A Test of Social-Adaptation Theory. Journal of Personality and Social
Psychology , 39(3),496.
Kamins, M. A. (1990). An Investigation into the "Match-Up" Hypothesis in Celebrity Adertising: When
Beauty May be Only Skin Deep. Journal of Advertising , 19(1),4-13.
Kelman, H. C. (1961). Process of Opinion Change. Public Opinion Quarterly , 25,57-78.
Kent, R. (2007). Marketing Research, Approaches, Methods and Applications in Europe. London:
Thomson Learning.
Kirk, R. E. (1999). Statistics: An Introduction. London: Harcourt Brace.
McCracken, G. (1989). Who is the Celebrity Endorser? Cultural Fundations of the Endorsement Process.
Journal of Consumer Research , 16(12),310-319.
Oliver, R. L. (1999). Whence Consumer Loyalty? Journal of Marking , 63,33-44.
Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and Peripheral Routes to Advertising
Effectivess: The Moderating Role of Involvement. Journal of Consumer Research , 10(2),135-146.
Prance, L. (2007, March 1). Study predicts backlash against celebrity endorsed products. Retrieved April
14, 2010, from cosmeticsdesign.com: http://www.cosmeticsdesign.com/Products-Markets/Study-predictsbacklash-against-celebrity-endorsed-products
Rossiter, J. R., & Percy, L. (1980). Attitude Change Through Visual Imagery in Advertising. Journal of
Advertising .
Page | 60

Rossiter, J. R., & Percy, L. (1980). Attitude Change Through Visual Imagery in Advertising. Journal of
Advertising , 9(2),10-16.
Saunders, M., Lewis, P., & Thornhill, A. (2007). Research methods for business students, fifth edition.
Harlow: Pearson Education Limited.
Shepard, R. N. (1978). The Mental Image. American Psychologist , 33(2),125-137.
Silvera, D. H., & Austad, B. (2003). Factors predicting the effectiveness of celebrity endorsement
advertisements. European Journal of Marketing , 38(11/12),1509-1526.
Solomon, M., Bamossy, G., Askegaard, S., & Hogg, M. K. (2006). Consumer Behavior, A European
Perspective. Harlow: Prentice Hall Europe.
Study skills. (n.d.). Retrieved April 13, 2010, from http://efn.hud.ac.uk/studyskills/referen.html
The World Factbook-Sweden. (n.d.). Retrieved April 28, 2010, from Central Inteligence Agency:
https://www.cia.gov/library/publications/the-world-factbook/geos/sw.html

Thiele, S. R., & Bennett, R. (2001). A brand for all reasons? A discussion of brand loyalty approaches
and their applicability for different markets. Journal of Product and Brand Management , 10(1),25-37.
Till, B. D. (1998). Using celebrity endorsers effectively: lessons from associative learning. Journal of
product and brand management , 15(6),400-409.
Till, B. D., & Busler, M. (1998). Matching products with endorsers: attractiveness versus expertise.
Journal of Consumer Marketing , 7(5),576-586.
Trout, J. (n.d.). ibranz. Retrieved April 8, 2010, from ibranz: http://www.ibranz.com/article8.html
White, D. W., Goddard, L., & Wilbur, N. (2009). The effects of negative information transference in the
celebrity endorsement relationship. International Journal of Retail& Distribution Management ,
37(4),322-335.

Page | 61

APPENDICES
Appendix 1-Survey in English
Marketing Survey
We are students at Ume University, and we are conducting this research survey in order to complete our
Masters thesis. This survey aims to investigate celebrity endorsement in cosmetics advertisements and its
effect on consumers brand loyalty. Your opinions will be valuable to our research. This survey will take
approximately 4 to 5 minutes and your help is highly appreciated. Thank you!

1. In which of the following age groups are you?

2.

3.
4.

5.

20-24
25-29
30-34
35-39
40 and above
What type of cosmetics products do you most often buy?
Makeup (face, eyes& lips)
Skincare (cleaners/toners, masks, day/night cream etc)
Fragrance (Perfume, Toilette, Cologne)
How often do you buy cosmetics?
Once a month
Every two months
Every three months
Less often
On average, how much do you spend on cosmetics every purchase? (In Swedish kronor)
Less than 50
50 to 99
100 to 199
200 to 399
400 and above
Where do you usually get information about cosmetics brands? More than one option is
possible
TV ads
Magazines
Newspapers
Internet
Friends/ relatives/ acquaintances
In shops
Other: _________

Please indicate to which extend you agree with the statements below
6. If I am not familiar with cosmetics brands, I choose those often promoted in advertisements.
Strongly Disagree

Strongly Agree

7. I change cosmetics brands regularly because I dont have any brand preference.
Strongly Disagree

Strongly Agree

8. I buy a brand because I like the celebrity endorsing the brand.

Strongly Disagree

Strongly Agree

9. I would only use brands which are promoted by famous celebrities.

Strongly Disagree

Strongly Agree

10. If my favorite celebrity endorsed a brand, I will buy it no matter what.

Strongly Disagree

Strongly Agree

Page | 62

11. I think using celebrities in advertisements is good, because it helps me to get to know a brand
better.
Strongly Disagree

Strongly Agree

12. I think the use of celebrities in advertisement is a guarantee for the quality of a brand
Strongly Disagree

Strongly Agree

13. I associate the endorsing celebrity with the brand (think of the brand when thinking about the
celebrity).
Strongly Disagree

Strongly Agree

14. I think only celebrities can represent the substance (e.g. value, content, and image) of cosmetics
brands in advertisement.
Strongly Disagree

Strongly Agree

15. I often gather information and do some research about cosmetics brands before I buy it.
Strongly Disagree

Strongly Agree

16. I would stick to a brand I usually buy rather than try other brands I am not familiar with.
Strongly Disagree

Strongly Agree

17. If I like a brand, I seldom switch to another just for trying something different.
Strongly Disagree

Strongly Agree

18. I rarely take chances by buying unfamiliar brands even if it is endorsed by celebrities.
Strongly Disagree

Strongly Agree

19. I would not switch to an unfamiliar brand due to the lack of knowledge about that brand.
Strongly Disagree

Strongly Agree

20. I would buy a brand regardless whether a celebrity endorses that brand or not.
Strongly Disagree

Strongly Agree

21. If I find a brand I like, I will always buy products of that brand.
Page | 63

Strongly Disagree

Strongly Agree

22. I want to get to know more about a new brand before I buy it, even if it is endorsed by
celebrities I like.
Strongly Disagree

Strongly Agree

23. It is not necessary to use celebrities in advertisement, because I will buy a brand recommended
by an unknown spokesperson in the advertisement as well.
Strongly Disagree

Strongly Agree

24. What are your opinions about cosmetics advertisements?

Strongly
disagree

Disagree

Neutral

Agree

Strongly
agree

Too much irrelevant information

Exaggerated product effect
Too much use of celebrities
Message in ads is unclear
All ads convey the same message
Please indicate the importance of the following factors to you in cosmetics purchasing
25. Indicate the important factors that influence your purchase of cosmetics products
Very
unimportant

Unimportant

Neutral

Important

Very important

Reputation of the brand

Wide selections of products of the brand

Brand endorser
Famous endorsers
Unique brand advertisement
Reasonable price
Friends recommendations
Previous purchase experience
Brands availability in stores
Salespersons knowledge about the brand
Product Packaging
Promotion offering of the brand
Clear instructions of how-to
Special benefits, (e.g. organic, natural
ingredients, allergy tested, etc)

Thank you for your cooperation!!

Page | 64

Appendix 2-Survey in Swedish

Marknadsunderskning
Vi r studenter inom marknadsfring vid Ume universitet, och vi genomfr en underskning fr
kunna slutfra vr Master-uppsats. Syftet med denna underskning r att underska anvndningen
knda personer i reklam och marknadsfring, och dess effekt p konsumenters lojalitet gentemot
mrke. Dina sikter r mycket vrdefulla fr vr underskning. Den tar ungefr 3 till 4 minuter
slutfra och din hjlp r vldigt uppskattad. Tack s mycket!

att
av
ett
att

1. Vilken ldersgrupp tillhr du?

20-24

25-29

30-34

35-39

ver 40

2. Vilken typ av kosmetikaprodukter kper du oftast?

Smink (ansikte, gon & lppar) Hudprodukter (rengring, masker, dag/nattkrm etc.)
Dofter (Parfym, Eau de Toilette, Cologne)

3. Hur ofta kper du kosmetika?

En gng i mnaden

Varannan mnad

Var tredje mnad

Mer sllan

4. verlag, hur mycket spenderar du p kosmetika vid ett tillflle? (Svenska kronor)
Mindre n 50

50 till 99

100 till 199

200 till 399

400 eller mer

5. Var fr du oftast information om kosmetikamrken? (Mer n ett alternativ r mjligt)

TV reklam
Vecko-/Mode Tidningar
Vnner/Slktingar/Bekanta I butiken

Dagstidningar Internet
Annat: _____________

Ange till vilken grad du hller med om nedanstende pstenden.

6. Om jag inte knner till ett kosmetikamrke vljer jag det som oftast frekommer i reklam.
Hller absolut inte med

Hller med fullstndigt

7. Jag byter kosmetikamrke regelbundet, jag fredrar inget srskilt mrke.

Hller absolut inte med

Hller med fullstndigt

8. Jag kper ett mrke fr att en kndis jag tycker om rekommenderar just det mrket i
reklamannonser.
Hller absolut inte med

Hller med fullstndigt

9. Jag anvnder bara mrken som r marknadsfrda av knda personer

Hller absolut inte med

Hller med fullstndigt

10. Om min favoritkndis rekommenderade ett mrke i reklam skulle jag kpa det ven om jag
aldrig hrt talas om det frut.
Hller absolut inte med

Hller med fullstndigt

Page | 65

11. Jag tycker att anvndningen av kndisar i reklam r bra, fr det hjlper mig att lttare bli
bekant med ett mrke.
Hller absolut inte med

Hller med fullstndigt

12. Jag tycker att anvndningen av en kndis i annonser r en garanti fr ett mrkes kvalitet.
Hller absolut inte med

Hller med fullstndigt

13. Jag frknippar kndisen i en annons med mrket (jag tnker p mrket nr jag tnker p
kndisen)
Hller absolut inte med

Hller med fullstndigt

14. Jag tror att bara en kndis kan representera substansen (t.ex. vrdet, innehllet, bilden) av ett
kosmetikamrke genom marknadsfring.
Hller absolut inte med

Hller med fullstndigt

15. Jag samlar ofta information om och undersker kosmetikamrken innan jag kper dem
Hller absolut inte med

Hller med fullstndigt

16. Jag skulle hellre fortstta kpa det mrke jag redan anvnder, n att testa andra mrken som
jag inte knner till.
Hller absolut inte med

Hller med fullstndigt

17. Om jag tycker om ett mrke, s byter jag sllan till ngot annat bara fr att testa ngot nytt.
Hller absolut inte med

Hller med fullstndigt

18. Jag tar sllan chanser genom att testa oknda mrken, ven om det r rekommenderat i
reklam av en attraktiv kndis.
Hller absolut inte med

Hller med fullstndigt

19. Jag skulle inte byta till ett oknt mrke som jag inte vet ngot om.
Hller absolut inte med

Hller med fullstndigt

20. Jag kper ett mrke oavsett om en kndis rekommenderar mrket i reklam eller inte.
Hller absolut inte med

Hller med fullstndigt

21. Om jag hittar ett mrke jag tycker om s kommer jag alltid att kpa dessa produkter.
Page | 66

Hller absolut inte med

Hller med fullstndigt

22. Jag vill veta mer om ett mrke innan jag kper det, ven om det r rekommenderat i reklam
av en kndis jag tycker om.
Hller absolut inte med

Hller med fullstndigt

23. Det r inte ndvndigt att anvnda kndisar i reklam, jag skulle ven kpa ett mrke
rekommenderat av en oknd talesman.
Hller absolut inte med

Hller med fullstndigt

24. Vad r dina sikter om marknadsfring av kosmetika?

Hller
absolut
inte med

Hller inte
med

Neutral

Hller
med

Hller med
fullstndigt

Fr mycket irrelevant information

verdriven effekt av produkten
Anvnder kndisar fr mycket
Oklart vad de vill frmedla
All reklam sger samma sak
Ange hur viktiga fljande faktorer r fr dig nr du kper kosmetikaprodukter.

25. Ange hur viktiga de faktorer som influerar dig nr du kper kosmetikaprodukter r.
Vldigt
oviktigt

Oviktigt

Neutral

Viktigt

Vldigt viktigt

Mrkets rykte
Brett utbud av produkter

Talesman fr mrket
Kndisar som talesmn
Unik marknadsfring av mrket
Rimligt pris
Vnners rekommendationer
Tidigare erfarenhet av mrket
Tillgnglighet i affren
Sljarens kunskap om mrket
Produktens frpackning
Specialerbjudanden
Bra anvndarinstruktioner
Speciella egenskaper, (t.ex. organisk,
naturliga ingredienser, allergi testad, etc)

Tack fr din medverkan!

Page | 67

Appendix 3-Survey Coding

Q1: 0=20-24,

1=25-29;

2=30-34;

3=35-39

4=40and above

Q2: 0= Makeup (face, eyes& lips); 1= Skincare (Cleaners/toners, masks, day/night cream etc)
2=Fragrance
Q3: 0= Once a month, 1=Every two months, 2=Every three months,

3= Less often

Q4: 0=Less than 50;

4= 400 and above

Q5: 0=TV ads;

1= 50 to 99;

2= 100 to 199; 3= 200 to 399;

1=Magazines; 2=Newspapers;

3=Internet;

4=Friends/relatives/acquaintances

5= In shops; 6=Other
Behavioral loyalty: BhaL
Q6: BhaL1: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q7: BhaL2: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q8: BhaL3: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q9: BhaL4: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q10:BhaL5: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q11: BhaL6: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q12: BhaL7: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q13: BhaL8: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q14: BhaL9: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Attitudinal Loyalty: AttL
Q15 AttL1: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q16 AttL2: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q17 AttL3: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q18 AttL4: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q19 AttL5: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q20 AttL6: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q21 AttL7: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q22 AttL8: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Page | 68

Q23AttL9: 1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree

Q24 Attads: Attitudes towards advertisements :
1=Strongly Disagree, 2= Disagree, 3=Neutral, 4= Agree, 5=Strongly Agree
Q25 Imfacotrs: Important factors influence purchasing decisions
1=Very unimportant, 2= Unimportant, 3=Neutral, 4=Important, 5=Very important

Page | 69

Appendix 4- The email of respondent recruitment

Dear Students,
We are students in the Marketing Masters program, and currently working on our masters
thesis. In order to be able to complete it, we would like to invite you to take part of a small
online survey. The survey is part of our thesis and completely anonymous. The data collected
will be used for further analysis in this study, and this study only.
This survey will take approximately 5 minutes and is about brand loyalty and celebrity
endorsement in cosmetics advertisements. Please be aware of the fact that we only need female
respondents.
If possible, please also help us to forward this survey to your female friends as well. Your
opinions are valuable to our study and your help is highly appreciated!
Thank you very much for your time!
Best Regards,
Emma Lfgren and Juan Li

Page | 70

Appendix 5-Raw Data Set

Q1 Age Group
Frequency
Valid

Percent

Valid Percent

Cumulative
Percent

20-24

30

30.0

30.0

30.0

25-29

44

44.0

44.0

74.0

30-34

10

10.0

10.0

84.0

35-39

9.0

9.0

93.0

40 and above

7.0

7.0

100.0

100

100.0

100.0

Total

Q2 Types of Cosmetics Purchase

Frequency
Valid

Percent

Valid Percent

Cumulative
Percent

Makeup

41

41.0

41.0

41.0

Skincare

52

52.0

52.0

93.0

Fragrance

7.0

7.0

100.0

100

100.0

100.0

Total

Q3 Purchase Frequency
Frequency
Valid Once a month

Percent

Valid Percent

Cumulative
Percent

28

28.0

28.0

28.0

Every two months

29

29.0

29.0

57.0

Every three months

30

30.0

30.0

87.0

Less often

13

13.0

13.0

100.0

100

100.0

100.0

Total

Q4 Spending
Frequency
Valid

Percent

Valid Percent

Cumulative
Percent

Less than 50

13

13.0

13.0

13.0

50-99

24

24.0

24.0

37.0

100-199

27

27.0

27.0

64.0

200-399

23

23.0

23.0

87.0

400 and above

13

13.0

13.0

100.0

100

100.0

100.0

Total

Page | 71

Q5-Sources
Frequency
Valid

Percent

Valid Percent

Cumulative
Percent

TV ads

41

41.0

41.0

41.0

Magazines

30

30.0

30.0

71.0

Newspapers

2.0

2.0

73.0

Internet

6.0

6.0

79.0

14

14.0

14.0

93.0

7.0

7.0

100.0

100

100.0

100.0

Friends etc
In shops
Total

Q6-Q14: Descriptive Analysis of Behavioral Loyalty

N

Minimum Maximum

Mean

Std. Deviation

BhaL1

100

3.04

1.136

BhaL2

100

2.52

1.243

BhaL3

100

1.96

1.053

BhaL4

100

1.69

.873

BhaL5

100

1.92

.950

BhaL6

100

2.63

1.186

BhaL7

100

2.07

1.075

BhaL8

100

2.73

1.145

BhaL9

100

2.17

1.035

Valid N (listwise)

100

Page | 72

Q15-Q23: Descriptive Analysis of Behavioral Loyalty

N

Minimum

Maximum

Mean

Std. Deviation

AttL1

100

2.95

1.167

AttL2

100

3.38

1.071

AttL3

100

3.39

1.100

AttL4

100

3.32

1.118

AttL5

100

3.38

1.170

AttL6

100

4.05

.968

AttL7

100

3.50

1.040

AttL8

100

3.58

1.084

AttL9

100

3.97

.926

Valid N (listwise)

100
Q24. Attitudes about cosmetics advertisements

Minimum Maximum

Statistic

Statistic

Statistic

Mean

Std.
Deviation

Skewness

Statistic

Statistic

Statistic Std. Error

Attads1

100

3.12

.935

-.017

.241

Attads2

100

4.13

.800

-1.087

.241

Attads3

100

3.84

.564

-.374

.241

Attads4

100

2.84

.896

-.107

.241

Attads5

100

3.65

.903

-.671

.241

Valid N
(listwise)

100

Page | 73

Q25-Important Factors Influencing Cosmetics Purchase Decisions

N

Minimum Maximum

Statistic

Statistic

Statistic

Mean

Std.
Deviation

Skewness

Statistic

Statistic

Statistic Std. Error

factor1

100

3.73

.839

-.603

.241

factor2

100

3.50

.948

-.290

.241

factor3

100

2.66

.768

-.010

.241

factor4

100

2.34

.755

-.084

.241

factor5

100

2.99

.882

-.251

.241

factor6
factor7

100
100

1
1

5
5

4.26
3.84

.733
.918

-1.236
-.553

.241
.241

factor8
factor9

100

3.93

1.018

-1.090

.241

100

3.59

1.074

-.813

.241

factor10
factor11

100

3.33

1.064

-.390

.241

100

2.91

1.129

-.207

.241

factor12

100

3.58

.901

-1.005

.241

factor13

100

3.28

1.190

-.342

.241

factor14
Valid N
(listwise)

100

3.86

.995

-1.093

.241

100

Page | 74

Ume School of Business

Ume University
SE-901 87 Ume, Sweden
www.usbe.umu.se