A study of Brand Personality construct and development of a scale to measure product Brand
Transcript of A study of Brand Personality construct and development of a scale to measure product Brand
A study of Brand Personality construct and development of a scale to measure product
Brand Personality in the Indian context
Thesis submitted to PONDICHERRY UNIVERSITY
for the award of the Degree of
DOCTOR OF PHILOSOPHY in
Management
by
JIJO GEORGE (Reg. No. 141170008)
under the supervision of Dr. S. VICTOR ANANDKUMAR
Associate Professor
DEPARTMENT OF MANAGEMENT STUDIES SCHOOL OF MANAGEMENT PONDICHERRY UNIVERSITY PUDUCHERRY – 605014. INDIA
MARCH 2014
JIJO GEORGE [Reg. No. 141170008] Ph. D. Research Scholar (Full–Time) Department of Management Studies School of Management Pondicherry University Puducherry – 605 014 -----------------------------------------------------------------------------------------------
DECLARATION
I, JIJO GEORGE, hereby declare that the thesis titled “A study of Brand
Personality construct and development of a scale to measure product
Brand Personality in the Indian context”, submitted to Pondicherry
University in fulfilment of the degree of DOCTOR OF PHILOSOPHY in
MANAGEMENT, is a record of original research work done by me under the
guidance and supervision of Dr. S. VICTOR ANANDKUMAR, Associate
Professor, Department of Management Studies, School of Management,
Pondicherry University.
I further declare that the thesis has not previously formed the basis for the
award of any Degree/ Diploma/ Associateship/ Fellowship or any other similar
title.
Place: Puducherry JIJO GEORGE
Date:
ii
DEPARTMENT OF MANAGEMENT STUDIES
SCHOOL OF MANAGEMENT PONDICHERRY UNIVERSITY PUDUCHERRY – 605014. INDIA
----------------------------------------------------------------------------------------------- Dr. S. VICTOR ANANDKUMAR Associate Professor
CERTIFICATE
This is to certify that the thesis titled “A study of Brand Personality
construct and development of a scale to measure product Brand
Personality in the Indian context” submitted to Pondicherry University in
fulfilment of the requirements for the award of the degree of DOCTOR OF
PHILOSOPHY in MANAGEMENT, is a record of original and bonafide
research work done by Mr. JIJO GEORGE [Reg. No.141170008] during the
period 2010-2014, at the Department of Management Studies, School of
Management of Pondicherry University, Puducherry, under my supervision and
guidance. This thesis has not previously formed the basis for the award of any
Degree/Diploma/Associateship/ Fellowship or similar titles and it represents
independent work on the part of the candidate.
I further certify that the work is worth submitting for the award of the said
degree.
Place: Puducherry Dr. S. VICTOR ANANDKUMAR
Date: Research Supervisor
iii
ACKNOWLEDGEMENTS
I would like to thank Dr. S. Victor Anandkumar, my research supervisor, for
his guidance, encouragement, patience and friendship throughout this research. His
influence on me has had a great effect on my learning experience. Throughout my
doctoral studies, he has provided encouragement, sound advice, good teaching and
lots of good ideas. I owe him lots of gratitude. Your guidance and mentorship over the
last four years has been instrumental in my development as a scholar and person as a
whole. I am eternally grateful for the time and effort you have put forth on my behalf.
I would also like to thank my doctoral committee members Dr. G.
Anjaneyaswamy (Professor, Department of Tourism Studies, Pondicherry University)
and Dr. T. Nambirajan (Professor & Head, Department of Management Studies,
Pondicherry University). Your contribution to my research work has been
immeasurable. The time, effort, and expertise that you have provided have immensely
improved the quality of my research.
A special thanks to Dr. R. Prabhakara Raya (Dean, School of Management),
Dr. K. Chandrasekara Rao (Professor & Head, Department of Banking Technology,
Pondicherry University), Dr. P. Natarajan (Professor & Head, Department of
Commerce), Dr. P. Sridharan (Head, Department of International Business), Dr.
Yedla Venkata Rao (Head, Department of Tourism Studies), Dr. Binu Zachariah
(Assistant Professor, Department of English), Dr. Lakhimai Mili (Assistant Professor,
Department of English), Dr. S. Visaka Devi (Assistant Professor, Department of
English), Dr. Bushan D. Sudhakar (Associate Professor, Department of International
Business), Dr. B. Rangaiah (Associate Professor, Department of Applied
Psychology), Dr. Surendra Kumar Sia (Associate Professor, Department of Applied
Psychology) and Dr. Pradeep Krishnatray (former Dean, School of Media &
Communication, Pondicherry University) for the assistance and support they
rendered at different levels of my research work.
I am also indebted to my many research scholar colleagues for providing a
stimulating and fun environment which helped me a lot to learn and grow. I thank
each of you for the valuable suggestions and contributions in my research.
iv
I am grateful to my parents (Mr. George P.V. and Mrs. Annamma George) and
my sister (Ms. Jisha George) for their love and support. My parents consistently
reinforced the value of education and learning. Without their tireless effort, sacrifice,
and encouragement, I never would have reached this point. I look forward to returning
the love and support they have given me. I would also like to thank my relatives who
motivated and prayed for the successful completion of my PhD work.
I would also like to thank the experts and respondents who helped me to
produce this research. I owe a great thanks to more than 1000 respondents who spent
their valuable time to give their response. A big ‘thanks’ to all of you.
When it comes to friends, I am indebted to many, both inside Pondicherry
University and outside. However, I cannot bring my acknowledgements to an end
without mentioning to Mr. Biju M.A, Mr. Prince Alex and Ms. Lincy P.T.
A word of thanks to the faculty members of the Department of Management
Studies for their kind support and suggestions rendered at various stages of this work;
especially during different presentations of this work in the department. I would like
to extend my thanks to the administrative staff at the Department of Management
Studies, for their kind help and assistance. I would like to thank all whose direct and
indirect support helped me to complete my thesis in time.
Last but most important, I want to thank my Lord who surely planned it all-
from its humble beginning to its fruitful end. To my fellow brothers and sisters who
intend to embark on this academic journey, I say: “Put all your trust in the Lord and
do not rely on your own understanding. At every step you take, keep the Lord in mind
and He will direct your path.” (Proverbs 3:5-6, The Revised English Bible).
Place: Puducherry
Date: JIJO GEORGE
v
TABLE OF CONTENTS No. TOPIC PAGE
No. Declaration of authorship ii Certificate of authorship iii Acknowledgements iv List of tables x List of figures xi CHAPTER I
INTRODUCTION
1.1. Introduction 1 1.2. The concept of brand personality 1 1.2.1. Sources of Brand Personality 2 1.2.2. A Brief Review on Brand Personality Measurement 2 1.2.3. Importance and significance of Brand Personality 3 1.2.4. Objectives and scope of the study 4 1.2.5. Need for the Study 4 1.2.6. The need for a brand personality scale in the Indian
context 5
1.2.7. Significance of the study 5 1.2.8. Original contributions 6 1.2.9. Scope of the Research 6 1.2.10. Period of the study 7 1.3. Overview of the scale development process 8 1.3.1. Phase1: Construct Definition 10 1.3.2. Phase 2: Content Validity 10 1.3.3. Phase 3: Measure Purification 11 1.3.4. Phase 4: Checking the reliability of the scale 11 1.3.5. Phase 5: Confirming brand personality dimensions 12 1.3.6. Phase 6: Validity and reliability 12 1.3.7. Phase 7: Developing a short version of the scale 12 1.3.8. Naming of factors 13 1.4. Discussion and conclusion 13 1.4.1. Limitations and delimitations of the study 13 1.5. Structure of the thesis 14 Chapter 2
THEORETICAL AND CONCEPTUAL BACKGROUND
2.1. Introduction 16 2.1.1. Creating brand personality 17 2.1.2. Brand personality as a relationship builder 17
vi
2.1.3. The position of brand personality in marketing 18 2.1.4. Brand personality in brand building 20 2.2. Conceptualisation of brand personality 20 2.2.1. Demarcating Brand personality 21 2.3. Theories related to brand personality 23 2.3.1. Brand Equity 23 2.3.2. Anthropomorphism Theory 25 2.3.3. Self Concept Theory 25 2.3.4. Self Congruity Theory 26 2.3.5. Personality Theory 27 2.4. Review of select Brand Personality literature 29 2.4.1. Review of select Brand Personality literature 29 2.4.2. Measuring Brand Personality 31 2.4.3. Brand Personality Scales 32 2.4.4. A review of the applications of brand personality
concept 58
2.4.5. Brand Personality Researches in the Indian Context 77 2.5. Chapter discussion and conclusion 78 2.5.1. Gaps considered for the present study 79 2.6. Summary of the literature review 79 Chapter 3
RESEARCH METHODOLOGY
3.1. The concept of measurement 80 3.2. Approaches to brand personality measurement 80 3.2.1. Qualitative Vs. Quantitative approaches to Brand
Personality Scale development 81
3.3. A review of methodology of building brand personality scales
87
3.3.1. Research approaches 87 3.3.2. Research methods adopted in Brand Personality scale
development 90
3.3.3. Generation of Items 94 3.3.4. Initial item reduction 97 3.3.5. Assessment o f dimensionality 98 3.3.6. Reliability and Validity of the Scale 100 3.4. Proposed scale development procedure 102 3.4.1. Sources of data 102 3.4.2. Brand selection 104 3.4.3. Methodology design for Item Generation 105 3.4.4. Methodology design for Language expert opinion
survey 106
3.4.5. Methodology design for Expert opinion survey 106 3.4.6. Methodology design for identifying the brand 107
vii
personality dimensions 3.4.7. Methodology design for confirming the brand
personality dimensions 108
3.5. Developing a short version of the scale 109 3.6. Research framework 110 3.7. Chapter conclusion 111 Chapter 4
SCALE DEVELOPMENT AND EVALUATION
4.1. Introduction 112 4.2. Defining the Construct 112 4.3. Content Validity 112 4.4. Stimuli Selection 113 4.5. Item Generation 114 4.5.1. Qualitative Study: Free Association Test 115 4.5.2. Items from Literature Review 117 4.6. Language Expert Opinion Survey 121 4.7. Expert Opinion Survey 121 4.8. Consumer Survey 122 4.8.1. Measure Purification: Exploratory Factor Analysis 123 4.8.2. Scale Reliability 124 4.9. Confirming Product Brand Personality Scale (PBPS)
dimensions 127
4.9.1. Construct Validity 130 4.10. A short version of the PBPS 133 4.10.1. Checking the reliability of the short version of PBPS 134 4.10.2. Confirmatory Factor Analysis (CFA) for short version
of PBPS 135
4.10.3. Construct validity 137 4.11. Naming of seven factors 140 4.12. Chapter conclusion 141 Chapter 5
FINDINGS, DISCUSSION OF RESULTS AND CONCLUSION
5.1. Introduction 142 5.2. Major findings 143 5.3. Originality and novelty 146 5.4. Implications 146 5.4.1. How a brand manager may use the PBPS 147 5.4.2. Implications for Academics 147 5.5. A critical discussion on PBPS 148 5.6. Comparison of PBPS dimensions with other scales 151 5.7. Limitations 153
viii
5.8. Future research 154 5.9. Chapter conclusion 158 REFERENCES
159
APPENDICES
177
I Free elicitation questionnaire II Language Expert Opinion Survey III Expert Opinion Survey IV Questionnaire for consumer survey V Questionnaire for holdout sample survey VI Statistical Outputs A Exploratory factor Analysis Results B Cronbach’s Alpha value Calculations of full version of
the scale
C Test Re-test Results of Full version PBPS D Confirmatory factor analysis outputs of Full version of
PBPS
E Results of Short version of the scale
ix
LIST OF TABLES
TABLE-2. 1 BRAND PERSONALITY SCALE (BPS) DEVELOPED BY AAKER (1997) ............. 35
TABLE-2. 2 BOSNJAK, BOCHMANN, AND HUFSCHMIDT’S BPS – FACTORS AND
ITEMS (2007) ............................................................................................................................... 43
TABLE-2. 3 SUMMARY OF BRAND PERSONALITY SCALES ............................... 49
TABLE-2. 4 COMPARISON OF BRAND PERSONALITY SCALES .............................................. 54
TABLE- 4. 1 BRANDS SELECTED FOR THE STUDY ................................................................. 114
TABLE- 4. 2 LIST OF ADJECTIVES FROM FREE ASSOCIATION TEST .................................. 116
TABLE- 4. 3 LIST OF ADJECTIVES FROM LITERATURE SURVEY ........................................ 117
TABLE- 4. 4 EXPLORATORY FACTOR ANALYSIS RESULTS ................................................. 124
TABLE- 4. 5 FACTOR WISE TEST RE-TEST CORRELATION VALUES .................................. 125
TABLE- 4. 6 VARIABLE-WISE TEST RE-TEST CORRELATION VALUES .............................. 126
TABLE- 4. 7 EVALUATION OF MODEL FIT ............................................................................... 129
TABLE- 4. 8 AVERAGE VARIANCE EXTRACTED .................................................................... 130
TABLE- 4. 9 CONSTRUCT RELIABILITY VALUES .................................................................... 131
TABLE- 4. 10 EVALUATION OF DISCRIMINANT VALIDITY .................................................. 132
TABLE- 4. 11 SHORT VERSION OF THE PRODUCT BPS (PBPS) ............................................. 133
TABLE- 4. 12 CRONBACH’S ALPHA VALUES OF THE SHORT SCALE ................................. 134
TABLE- 4. 13 TEST RE-TEST CORRELATION VALUES OF THE FACTORS IN THE
SHORT SCALE .......................................................................................................................... 135
TABLE- 4. 14 - FIT INDICES FOR THE EVALUATION OF CFA MODEL OF THE
SHORT PBPS ............................................................................................................................ 137
TABLE- 4. 15 AVERAGE VARIANCE EXTRACTED VALUE IN THE CASE OF SHORT
PBPS ........................................................................................................................................... 138
TABLE- 4. 16 - CONSTRUCT RELIABILITY VALUES OF SHORT PBPS DIMENSIONS ........ 139
TABLE- 4. 17 EVALUATION OF THE DISCRIMINANT VALIDITY OF THE SHORT
PBPS ........................................................................................................................................... 140
TABLE- 4. 18 NAMES OF SEVEN FACTORS IDENTIFIED ........................................................ 140
x
LIST OF FIGURES
FIGURE-1. 1 SCHEMATIC PRESENTATION OF THE STUDY POPULATION OF
BRANDS ........................................................................................................................................ 7
FIGURE-1. 2 FLOW CHART OF THE SCALE DEVELOPMENT PROCESS ................................... 9
FIGURE-2. 1. BRAND PYRAMID ..................................................................................................... 19
FIGURE-2. 2 BRAND BUILDING MODELS .................................................................................... 20
FIGURE-2. 3 BRAND IDENTITY PRISM ......................................................................................... 22
FIGURE-2. 4 FIVE FACTOR MODEL OF HUMAN PERSONALITY ............................................. 28
FIGURE-2. 5 AMBROISE, FERRANDI AND MERUNKA’S BRAND PERSONALITY
SCALE (BPS) (2005) ................................................................................................................... 39
FIGURE-2. 6 SWEENEY AND BRANDON’S BRAND PERSONALITY SCALE (2006) .............. 40
FIGURE-2. 7 BOSNJAK, BOCHMANN, AND HUFSCHMIDT’S BPS (2007) .............................. 422
FIGURE-2. 8 GEUENS, WEIITERS AND WULF’S BRAND PERSONALITY SCALE
(2009) .......................................................................................................................................... 444
FIGURE-2. 9 KUENZEL AND PHAIROR’S BRAND PERSONALITY SCALE (2009) ............... 455
FIGURE-2. 10 HEINE’S LUXURY BPS (2009) ............................................................................... 466
FIGURE-2. 11 HEERE’S BRAND PERSONALITY DIMENSIONS (2010) ................................... 477
FIGURE-2. 12 BRAND PERSONALITY DIMENSIONS FOR CITIES .......................................... 644
FIGURE- 3. 1 FRAMEWORK FOR SCALE DEVELOPMENT ...................................................... 110
FIGURE- 4. 1 CONFIRMATORY FACTOR ANALYSIS MODEL OF PBPS ................................ 128
FIGURE- 4. 2 CONFIRMATORY FACTOR ANALYSIS MODEL FOR SHORT VERSION
OF PRODUCT BPS .................................................................................................................... 136
FIGURE- 5. 1 PRODUCT BRAND PERSONALITY SCALE ......................................................... 145
xi
Chapter 1
INTRODUCTION
This chapter intends to introduce the area in which this research is
conducted by presenting the concept of brand personality in general and then
narrowing down to the objectives of this research. An attempt is also made to
demonstrate the importance, significance and originality of the work.
1.1. Introduction
In a competitive business environment, consumers have numerous
choices of brands with several of them providing similar functional benefits.
Because of competition and innovative technologies, it is now easy for
competitors to copy the offerings of its rivals. So, differentiating one brand
from its competitors merely based on its functional attributes alone has become
almost impossible. Offering symbolic benefits to consumers and thereby
gaining an edge over the competitors has become important in this regard.
Brand personality plays a crucial role in offering symbolic benefits to
consumers. It is of greater importance than other concepts like brand identity,
since consumers pay more attention to it when brands become more
homogeneous (Hussey and Duncombe, 1999).
1.2. The concept of brand personality
The tendency to attribute human characteristics to brands existed in the
marketing domain as early as 1950s. According to Azoulay and Kapferer
(2003), it was Martineau (1958) who first introduced this term in marketing
research to refer to the non-material features of a store which makes it special.
King (1970) quoted the research works of J. Walter Thompson Advertising
Agency which indicate that consumers attribute personality characteristics to
brands and talk about it.
1
Before 1980s brand differentiation was mainly based on product
performance. But with increasing number of products which provided almost
the same set of functional benefits, it became difficult for marketers to
differentiate brands merely on the basis of its performance. In order to
overcome this trouble, the advertising agency Ted Bates introduced the concept
of Unique Selling Personality following the concept of Unique Selling
Proposition created by Rosser Reeves. As a consequence, from 1970 onwards
brand personality becomes an inevitable component of advertising strategy and
it captured everything related to the non-material features of the brand
(Azoulay and Kapferer, 2003).
1.2.1. Sources of Brand Personality
Brand personality is a concept that has its roots in human personality
concept in Psychology. However, the two concepts differ mainly in terms of
how they are formed in the minds of people. While the former is formed based
on some indirect sources like celebrity endorsers, typical users, product
features and so on (Aaker, 1997), the latter derives its image out of direct
sources such as individual behaviour, physical characteristics, attitudes and
beliefs, and demographic characteristics (Park, Jaworski, and Maclnnis, 1986).
1.2.2. A Brief Review on Brand Personality Measurement
Brand personality measurement research can be traced back to 1960’s
(Birdwell, 1968; Dolich, 1969). Early research works focused on the self
congruity concept or in other words, understanding the relationship between
the consumers’ self and brand personality. For example, Birdwell (1968)
conducted a study using a compiled list of bi-polar items containing
appropriate adjectives which described both automobile and human
personalities, to understand the relationship between consumers’ perception
about their car and their self perception. Another early study on product
personality was one done by Dolich (1969). It was done with an intention to
understand the relationship between consumers’ actual and ideal self image and
2
personalities of four product categories namely beer, cigarettes, bar soap and
tooth paste. The researcher used a scale adapted from human personality scales
to study the same. The research works of Malhotra (1981), on construction of a
scale which measured self-concepts, person concepts and product concepts and,
Karande, Zinkhan, and Lum (1997) on brand personality and self-concept are
some of the other noteworthy works.
A breakthrough in the field of brand personality research came with the
work of Aaker (1997) who developed the first reliable, valid and generalizable
brand personality scale (BPS). This scale consisted of five dimensions, 15
facets and 42 traits. However, it was criticized on various grounds such as
loose definition of the construct (Azoulay and Kapferer, 2003), considering
only positive brand attributes for measuring brand personality (Geuens,
Weijters, and De Wulf, 2009) and its culture-specific nature since the five
factors identified by Aaker (1997) in the United States (U.S.) were not found to
be replicable in the case of other cultures. For example Aaker, Benet-Martinez,
and Garolera (2001) found that Peacefulness replaced Ruggedness in the case
of Japan and in the case of Spain, only three factors (Sincerity, Excitement, and
Sophistication) out of five were found to be applicable. The other two factors in
the case of Spain were Passion and Peacefulness. This limitation of Aaker’s
scale has lead to a large number of culture-specific studies. A French BPS was
created by Ferrandi, Valette-Florence, and Fine-Falcy (2000), an Italian one
was developed by Caprara, Barbaranelli, and Guido (2001), a German one by
Bosnjak, Bochmann, and Hufschmidt (2007), and a Belgian BPS by Geuens et
al. (2009). Some researchers who followed Aaker (1997), created scales
(Bosnjak et al., 2007; Geuens et al., 2009) which successfully addressed some
criticisms regarding Aaker’s scale.
1.2.3. Importance and significance of Brand Personality
Brand personality assumes significance since it plays a vital role in
understanding brand choice (Plummer, 2000) and is invaluable in building
brand equity (Van Rekom, Jacobs, and Verlegh, 2006). A research by Sirgy
3
(1982) reveals that brand personality plays an essential role in increasing the
consumer preference for the brand as well as its usage. Further, the research
works of Biel (1992) and Keller, Parameswaran, and Jacob (2011) established
that the aforementioned concept helps consumers to differentiate a brand from
its myriad of competitors. Also, it serves as a mode of self expression (Belk,
1988). Functional differences are often minimal among competing brands and
symbolic benefits like brand personality remains as the key to create a strong
point of differentiation, which puts the concept of brand personality in the
forefront of marketing (Biel, 1992; Aaker and Joachimsthaler, 2000).
By understanding the personality of the brand, managers can manage
their brands in a better manner. This can also help them to create insights
regarding the problems related with a brand’s personality and generate possible
solutions. Hence measuring brand personality is vital for brand management.
1.2.4. Objectives and scope of the study
The aim of this study is to create a scale that incorporates various
advancements in the field of brand personality measurement and which
measures brand personality in the Indian context. The scope of this study was
limited to product brands. The major objectives of this study are:
• To develop a product Brand Personality scale in the Indian context
• To test the reliability of the scale
• To assess the validity of the scale
• To create a short version of the product Brand Personality Scale
1.2.5. Need for the Study
The need for a Product Brand Personality scale arises out of two
reasons. Firstly, the sources of product and service brand personality
dimensions vary significantly. Further, confining the study to product brands
can help the tool to be more precise. It is better to have separate brand
4
personality scales for product and service brands. Since Aaker (1997) created a
scale to measure brand personality in general, many successive studies were
done to develop category specific brand personality scales, especially in the
case of service brands. For example, Ekinci and Hosany (2006) created a
destination personality scale, d’Astous and Lévesque (2003) developed a scale
to measure store personality, Venable (2005) investigated the personality of
non-profit organizations and so on. This is the first scale to measure Product
Brand Personality.
1.2.6. The need for a brand personality scale in the Indian context
As Aaker’s (1997) scale was not found to be valid in contexts other than
the U.S., it prompted many successors of Aaker to develop country-specific
scales (Aaker, Benet-Martínez, and Garolera, 2001; Caprara et al., 2001; Sung
and Tinkham, 2005; Geuens et al., 2009). Further, Aaker’s scale was not found
to be applicable in the Indian context also (Thomas and Sekar, 2008). Brand
personality research in India is mainly on the basis of Aaker’s (1997) brand
personality scale (For example, Thomas and Sekar, 2008; Purkayastha, 2009).
Again, as discussed earlier, Aaker’s scale was criticised on many grounds. All
these necessitated the study for developing a Product brand personality scale in
the Indian context.
1.2.7. Significance of the study
This scale identifies the product brand personality dimensions and the
items to measure those dimensions in the Indian context. This work will help
Indian researchers to consider a more reliable and valid scale in the Indian
context to conduct studies on brand personality. This scale is expected to foster
marketing research in general and brand personality research in particular, in
India.
From an academic point of view, this scale will help to simplify
theorizing and hypothesis generation. For marketing practitioners, this scale
will help to assess the match or mismatch between the perceived and projected
5
personality of a brand and to assess the congruity of the brand’s personality
with that of its target customers in the Indian context. Also it can help in
making decisions to monitor the intended personality of their brand.
1.2.8. Original contributions
This is the only study hitherto conducted to uncover the brand
personality dimensions focusing on product brands alone (product brand
personality). Further, context-wise, this is the first India-specific study to
develop a scale to measure product brand personality.
1.2.9. Scope of the Research
1.2.9.1. Location and Area of the Study
The area of study was selected to be Indian sub-continent, which is
significantly different from other contexts where brand personality research has
taken place. Most of the brand personality researches took place in developed
countries, whereas India is a developing country and an emerging market. This
makes India a significantly different context.
1.2.9.2. Selection of Sample and Respondents
Stimuli Selection: Stimuli refer to the specific objects based on which
respondents will make their responses during a survey. In the case of this
research, stimuli represented a bouquet of carefully chosen brands. Instead of
relying on a general brand domain, this study was restricted to the product
brand sphere. Again, regional level or state level brands were excluded. As the
study is intended to represent a pan-Indian population, a set of brands which
were relevant in the national stage was opted. In the item generation phase of
this research, respondents were given the option to select brands of their choice
as stimuli, whereas a random sample of brands was opted for in the successive
stages of the study. Figure 1.1 gives a schematic representation of the
population of brands considered.
6
Figure-1.1. Schematic representation of the study population of brands
[Source: Original]
Respondent selection: The respondent selection method varied from phase to
phase, to meet the specific requirements of each stage of this research. In the
item generation phase of the study, a judgement sampling method was adopted.
A set of post-graduate management students was chosen at this phase. In the
Consumer survey phase wherein the Indian product brand personality
dimensions were identified, a quota sampling was adopted to get a
representation of Indian population in terms of area and gender.
1.2.10. Period of the study
The entire scale development process spread over February 2011 to
March 2013. Item generation from literature was done during the period from
February 2011 to December 2011. Item generation based on qualitative study
was done during the month of January and February 2012. This was followed
Brands
Study population
7
by the language expert opinion survey in March 2012. The expert opinion
survey phase was between April 2012 and July 2012, followed by the
Consumer survey phase from July to December 2012. After this, data for Test
re-test study was collected in January 2013. And the final hold out sample
study was done during February to March 2013.
1.3. Overview of the scale development process
The methodology adopted for this study is presented in the form of a
schematic diagram (Figure-1.2) illustrated in the next page.
8
Figure-1.2. Flow chart of the Scale development process
[Source: Adapted from Churchill Jr (1979)
Construct Definition
Content Validity • Item Generation • Stimuli Selection
Measure Purification • Language Expert
Opinion Survey • Expert Opinion Survey • Consumer Survey
Item Selection
Reliability Assessment • Cronbach’s Alpha • Test-retest
Analysis using Hold out sample
• Confirmatory Factor Analysis
Validity Assessment • Discriminant Validity • Convergent Validity
Developing a short version of the scale Assessment of the reliability and validity of the short scale
9
1.3.1. Phase 1: Construct Definition
Among the multitude of definitions of brand personality, the one by
Azoulay and Kapferer (2003) was opted for this study. They define brand
personality as ‘the set of human personality traits that are both applicable to
and relevant for brands’. This definition was adopted since it has been accepted
by different brand personality researchers to be a better and more precise
definition.
1.3.2. Phase 2: Content Validity
Content validity or item sample adequacy refers to the degree to which a
particular set of items reflects a content domain (DeVellis, 2011). Aaker (1997)
entails two steps to ensure content validity, which concerns with item selection
and stimuli selection. Firstly, a broad and representative set of items which can
be used to measure brand personality are to be pooled and secondly, a set of
stimuli (brands) against which the appropriateness of the items shall be
examined using an empirical study are to be selected. Nunnally (2010) argues
that if construct definition, item selection and stimuli selection are done in a
well defined manner, then the content validity of the scale can be established
before its construction.
Item Generation: The item generation phase of this study constituted two parts:
first, generation of item pool from an extensive literature survey and second, a
qualitative study using the method of free elicitation to generate the indigenous
items which will make the scale more localized.
Stimuli Selection: To ensure content validity, it is necessary to choose a broad
and representative set of stimuli (that is, brands). For this, 21 brands were
selected from a list of top brands according to 2012 surveys by Trust Research
Agency and Brand Equity (The Economic Times).
10
1.3.3. Phase 3: Measure Purification
This phase consisted of a Language expert opinion survey, an Expert
opinion survey and a Consumer survey.
Language expert opinion survey: This survey was intended to identify and
eliminate/replace items that are not commonly used in the Indian context. This
study was carried out using a structured questionnaire administered among
three English department faculty members in a University with expertise in
English language teaching and usage assessment and considerable teaching and
research experience.
Expert opinion survey: Following the language expert opinion study, to further
reduce the items to a manageable level, an expert opinion survey was
conducted among 12 experts of which three were marketing area faculty
members in Business Schools, three Brand Managers, three Advertising
Professionals and three Consumer Psychologists.
Consumer survey: The purpose of this survey was to explore the factor
structure of brand personality as perceived by Indian consumers. For this,
responses were collected from 606 respondents who were University students
representing a pan- Indian sample. A total of 21 brands were considered at this
stage and each respondent was required to rate 3 brands of his/her choice based
on his/her familiarity with the brand. The responses were collected using a
structured questionnaire with 5-point Likert scale. The data collected was
analysed using exploratory factor analysis with principal component analysis
and Varimax rotation. This helped in identifying a brand personality structure
consisting of seven factors and 45 items.
1.3.4. Phase 4: Checking the reliability of the scale
The external reliability was checked using test re-test reliability test for
which a survey was conducted among 91 respondents chosen from the first
consumer survey. Internal reliability of the scale was checked using
Cronbach’s alpha. The results of these tests helped to prove that the scale had
both internal and external reliability.
11
1.3.5. Phase 5: Confirming brand personality dimensions
The result obtained by the factor analysis of consumer survey data was
addressed with two questions: First, whether the same dimensions will be
obtained if another set of brands were used for the study? Second, whether
there will be any change in the result if a different subject sample was used? A
hold out sample survey was conducted to address these questions. Brands and
subjects considered for this study were different from that of first consumer
survey. The sample size for this study was restricted to 217. A total of 12
brands were selected using simple random sampling from a list of top brands
according to 2012 surveys by Trust Research Agency and Brand Equity (The
Economic Times). A confirmatory factor analysis of the data from the hold out
sample survey showed adequate level of fit with the model developed using
exploratory factor analysis of consumer survey in Phase 3.
1.3.6. Phase 6: Validity and reliability
Construct validity of the new measure was established by evaluating
convergent validity and discriminant validity of the measure. Three main
indicators of convergent validity, namely factor loadings, average variance
extracted and construct reliability were examined and were found to have
recommended values to indicate convergent validity. The scale was also found
to have adequate discriminant validity.
1.3.7. Phase 7: Developing a short version of the scale
The objective of this part was to create a short version of the Indian BPS.
Since, brand personality is often measured along with other concepts such as
self concept, celebrity personality and so on, a large number of items in the
scale will make it practically insignificant. Hence many researchers have
developed a short version of their scales (For example, Aaker, 1997; Burisch,
1997; Geuens et al., 2009; Rammstedt and John, 2007). To meet this purpose
as well as to satisfy structural model validity, the maximum number of items to
be included under a particular factor was decided to be 4 and minimum as 3.
12
Item selection was done with the help of factor analysis. This resulted in the
final version of the scale with 7 factors and 26 items. The psychometric
properties of this short version were also evaluated by means of testing test re-
test reliability, Cronbach’s alpha values, convergent validity and discriminant
validity.
1.3.8. Naming of Factors
A name was given for each dimension which reflected the general
characteristics of that dimension. The chosen names were as follows: Factor 1 -
Happy, Factor 2 - Youthful, Factor 3 - Reliable, Factor 4 - Adventurous, Factor
5 - Competent, Factor 6 - Accountable and Factor 7 - Appeal.
1.4. Discussion and conclusion
Some of the brand personality dimensions identified by this India-
specific study have similarities with those of Aaker’s (1997) scale. For
instance, the Appeal dimension from this study and Sophistication dimension
of Aaker’s scale try to capture similar aspects of brand personality. However,
Charming is the only item shared in common between the two. The Competent
dimension from this study and Competence dimension of Aaker’s scale to
measure how well a brand appears to be successful and efficient to its
consumers. The Excitement and Sincerity dimensions of Aaker’s scale seem to
be captured by the Happy dimension in the scale developed through this study.
The Ruggedness dimension of Aaker’s scale and Adventure dimension from
this study are distantly similar. However, the dimensions Reliable, Youthful and
Accountable emerge as India-specific dimensions.
1.4.1. Limitations and Delimitations of the study
The study was delimited to a student sample which represented a pan-
Indian population, and was intended to develop a scale which can be used
exclusively to measure the personality of product brands. But these
delimitations may be treated as some of the limitations of this work. The use of
13
a student sample may hinder the generalizability of this study. However, a
study by Aaker (1997) showed that there is no significant difference in the
brand personality dimensions based on demographic features. But, it is worth
examining whether a significant difference exists in brand personality
perception based on demographic characteristics in the Indian context. Next,
the number of brands considered for the study for identifying the number of
dimensions, was limited to 21. The use of a large number of brands would have
enhanced the generalizability.
Area of study is another limitation of this work. The study is limited to
the Indian sub-continent, the brands selected and participants of the survey - all
of which reflect an Indian scenario, which may not reflect the global scenario.
Like in the case of other brand personality researches, the outcome of this study
will have to be tested if it is valid in other contexts.
1.5. Structure of the thesis
This thesis titled ‘A study of Brand Personality construct and
development of a scale to measure product Brand Personality in the Indian
context’ has been presented in five chapters. The organization and brief content
of the chapters are as follows:
Chapter I titled ‘Introduction’ presents an overview of the topic and
discusses on the concept of brand personality, importance of the construct,
need for the study, objectives and research framework, scope of the study,
overview of the scale development process, overview of results, limitations and
delimitations of the study.
Chapter II titled ‘Theoretical background’ details the theoretical
background of the concept which discusses aspects such as the origin of the
concept of brand personality, how the concept is different from similar
concepts like brand identity and brand image, the antecedents and
consequences of the concept of brand personality and the popular scales
available for brand personality measurement.
14
Chapter III titled ‘The measurement concept and methodological
framework’ presents the concept of measurement, its origin and role in
marketing and concepts of reliability and validity. This chapter also discusses
the different scale development practices so as to bring out the best framework
to create a measurement tool.
Chapter IV titled ‘Scale development process’ details the systematic
process adopted for the study. The various steps include construct definition,
content validity, measure purification using exploratory factor analysis, hold
out sample analysis for confirming brand personality dimensions, naming of
factors and checking the reliability and validity of the scale.
Chapter V titled ‘Summary, discussion and conclusion’ brings out the
summary of results, scope for further research, implications for academics and
practitioners.
The ‘Appendix’ contains the data collection instruments and statistical analysis outputs.
15
Chapter 2 THEORETICAL AND CONCEPTUAL BACKGROUND
Objective: The structure of any research should rise from a thorough review of
literature. The objective of this chapter is to review the published literature
related to the research problem considered herein. The shared goals of this
chapter include: developing and demonstrating a familiarity with the body of
knowledge in the area of brand personality research, to establish the credibility
of this research, to summarize and link between the various research works
done in this area, and to link the current work with the existing literature.
2.1. Introduction
Consumers perceive brands to possess a personality, which can be used
as an avenue for self expression or to experience the emotional benefits by
which the brand differentiates itself from others (Phau and Lau, 2000).
Researches related to this aspect of brands are termed as brand personality
researches. To conduct any study in this domain, it is necessary to develop a
strong understanding of this concept. Moorthi (2003) argues that many
practitioners and academicians are of the opinion that, it is easier to understand
brand personality intuitively than defining it. For example one may describe the
natural health care product brand Dabur to be like his ‘grandfather’. It is a very
simple presentation of the personality of Dabur as a brand. This in turn can
imply that Dabur
1) Has a strong heritage
2) Can refer to the ayurvedic and herbal benefits of the brand
3) Healthy
4) Nurturing
5) Dependable
The lack of a unique personality may cause a brand go unnoticed for a
long time. Hence it is essential for any brand to have a unique brand
16
personality. A unique and enduring brand personality becomes a need of the
hour given the intense competition among and proliferation of brands. Further,
consumer finds it easier to remember brands that have a strong personality.
Amrutanjan takes on a personality of strength and competence which makes it
hard for consumers to miss it in the pain balm segment. Another benefit of
brand personality is that it helps to distinguish one brand from another or a
brand from another product.
There are mainly two issues related to brand personality; the first one is
connected to creation of brand personality and, the second is related to the
measurement of the personality. Creation of brand personality is mainly
achieved by using high level of advertisements and celebrity endorsements
along with packaging, slogans, logos, user imagery and other elements of
marketing mix (Plummer, 1985; Batra, Lehmann and Singh, 1993; Aaker,
2012).
2.1.1. Creating brand personality
Consumers derive a particular perception regarding the personality of a
brand by mixing their knowledge regarding the brand such as its benefits and
attributes to the identity created by marketers (Keller, 1993). Thus it can be
seen that brand personality is an outcome of the marketer-generated images and
consumer-perceived images and attributes.
Brand personality is one among the intangible associations along with
brand’s values, vision, and philosophy and so on, which brings in emotional
ties beyond product satisfaction with the consumer (Kapferer, 2012). Most
brands did not create brand personality as such. It is the conscious effort of the
marketers to acquire more consumer equity and brand value which results in
the establishment of a strong brand personality.
2.1.2. Brand personality as a relationship builder
Brand personality arises when one talks about a brand as if talking about
a friend or person. One of the factors that contribute to the singularity of the
17
brand is its personality, which usually arises out of an anthropomorphic
conception of the brand. This is an outcome of the need to pursue relationship
marketing; or in other words, it is about giving more importance to building an
enduring relationship between the brand and consumer, rather than focussing
on immediate profit. This is more similar to that of building relationship
between two persons: brand as a person and the consumer as another, hence the
notion of brand personality (Aaker and Biel, 2013)
To communicate a particular kind of personality, a brand may use a
multitude of strategies including advertisements using celebrity endorsers. For
example, ThumsUp is a cola brand which seeks to position itself as a masculine
and adventure-seeking brand. To communicate this, they used celebrity
endorsers like Akshay Kumar, who is a synonym of dare and adventurism in
the Bollywood movies. The impact is that consumers will get a feeling that
Thums Up is a brand which is closely associated with Akshay Kumar; so
naturally they start attributing his characteristics to the brand.
Another example is that of L’Oréal Paris which tries to position itself as
a ‘glamour’ brand. It attains this by employing a ‘dream team’ of Hollywood
stars and globally popular models in their advertisements.
Yet another classic example for crafting of brand personality cited by
Kapferer (2012) was the one made by French automobile manufacturer
Citroen. They named some of their brands with the name ‘Picasso’ (Citroën C3
Picasso, Citroen C4 Picasso and so on). Associating the brand with the world
famous Pablo Picasso, increased the perception and novelty of its brands,
which in turn helped them to compete successfully with its competitor brands
from Renault.
2.1.3. The position of brand personality in marketing
Kapferer (2012) says it is the common spirit, vision and ideals embodied
in the products of a brand that helps to unify them rather than other external
18
signs and features. He argues that major brands can be compared to a pyramid
as shown below
Figure-2.1. Brand Pyramid [Source: Kapferer (2012)]
According to him, the top of the pyramid stands for brand vision and
purpose, followed by brand personality codes, which should be designed in
such a manner as to reflect a singular character of the brand. The next level
constitutes the brands strategic features arising out of its overall vision and
objectives. The bottom of the pyramid consists of each model’s positioning
strategy.
It can be seen that, in this brand pyramid, brand personality and vision
form the top of the pyramid. Therefore product or model positioning is
something which arises out of these (that is, vision and brand personality) when
it comes to brand building or brand management process whereas, consumers
19
derive their perceptions in the reverse manner - generating an understanding of
the personality of the brand from the positioning of the products.
2.1.4. Brand personality in brand building
According to Kapferer (2012), there are two models to build strong
brands. The first one starts with the tangible added values such as benefits,
attributes and ingredients and move on to intangible ones like brand
personality, vision and values. The second model starts with intangible added
values like brand personality and moves on to products that deliver these
values.
Figure-2.2. Brand building models [Source:Kapferer (2012)]
Brand associations are typically considered to move upward a ladder in
time. For example,a brand will initially focus on ingredient (for example, Dove
is having hydrating cream),then attributes (for example, softening), then benefit
of using the brand (say, protection) and then at the top comes the symbolic or
intangible benefits such as brand personality.
2.2. Conceptualisation of brand personality
A brand refers to ‘a name, term, sign, symbol, or design, or a
combination of them, intended to identify the goods or services of one seller or
group of sellers and to differentiate them from those of competitors’ (American
Marketing Association, 1960). According to Keller et al. (2011) a brand can
provide three types of benefits to its customer, namely functional or utilitarian
20
benefits, symbolic benefits and experiential benefits. Functional benefit usually
refers to the ability of a brand to provide solution to a problem faced by a
customer. For example, Unilever’s Surf detergent helps to remove stain from
clothes. Experiential benefits are related with the sensory pleasure which the
customer derives out of the brand (For example, the taste of Dairy Milk
chocolate from Cadbury). Finally, symbolic benefits of a brand indicate the
signs or signals which a brand delivers to its consumer (for instance, the
prestige of owning a Mercedes Benz car). The signal effects of the brand may
originate out of the typical user of the brand, or even from the personality of
the brand. Proliferation of brands and intense competition among them has
forced marketers to move from utilitarian and experiential benefits to symbolic
benefits to differentiate their brands. According to Zentes, Morschett and
Schramm-Klein (2008), brand personality is an important one among such
symbolic benefits. Aaker and Fournier (1995) consider brand personality as
“set of meanings constructed by an observer to describe the ‘inner’
characteristics of another person… that can be used to summarise complex
behaviours and form expectations of future behaviours” or in other words,
brand personality refers to the inner features of a brand as perceived by
consumers based on the characteristics and behaviour of a brand. Batra et al.
(1993) describe this concept as dealing with how consumers use traits that are
used to describe the personality of a person to describe that of a brand. For
Aaker (1997), brand personality refers to ‘the set of human characteristic
associated with a brand’.
2.2.1. Demarcating Brand personality
There is close relationship between the concepts of brand personality,
brand identity and brand image. Hence, it is worth reviewing the inter-
relationships among these concepts in detail. Brand personality is considered to
be an important component of brand identity and brand image (Geuens et al.,
2009). According to Kapferer (2012), brand identity is what the firm wants to
project to its target customers. Whereas, brand image is what the target
21
consumers understand or interpret about the brand identity (De Pelsmacker,
Geuens and Van den Bergh, 2007). Brand image and brand identity are
generally considered to be multidimensional constructs and brand personality is
just a component of them. For instance, as per Keller et al. (2011), personality
of the brand is just a component among many other aspects such as user
profiles, heritage, history, values, purchase and usage situation, experiences
and so on. An important typology of brand identity worth mentioning is the one
proposed by Aaker and Joachimsthaler (2000). Their brand identity typology
consists of perspectives which treats brand as a Product, Organization, Person
and a Symbol.
Kapferer (2012) considers brand identity to consist of six facets namely
physique, personality, relationship, culture, reflection and self-image. It is
obvious that brand personality is just a single facet of brand identity, or in other
words one can say that brand personality is contained in brand identity.
Figure- 2.3. Brand Identity Prism [Source: Kapferer (2012)]
22
Based on the literature, it is observed that brand personality has two
faces; the first one is portrayed personality which constitutes the brand identity
and the second is perceived personality dimension which forms a part of brand
image.
Another concept which is closely related to brand personality and often
misunderstood as brand personality is user imagery (Aaker, 2012). According
to Aaker (2012), ‘human characteristics associated with the typical user of a
brand’ is termed as user imagery. Researches have indicated that user imagery
and brand personality are distinct constructs and they will not give similar
results (Keller et al., 2011). The case of P&G’s Oil of Olay given by Plummer
(1984) may be a good example for this. As per Plummer (1984), a typical user
of Oil of Olay is a down-to-earth, solid, pretty, female citizen where as the
brand personality of this brand is more aspirational and upscale one.
2.3. Theories related to brand personality
An understanding of the theories related to or underlying the concept of
brand personality is important to operationally define the construct. It can also
help in understanding the relevance of this construct to the consumer, how a
product’s value is enhanced by the virtue of this attribute and so on. This
section tries to explore the various theories that support and form a background
to the concept of brand personality.
2.3.1. Brand Equity
Brand personality helps in building brand equity. According to Stein
(2004), the role of brand personality in the creation of brand equity can be
explained by relying on different models such as Relationship basis model,
Self-Expression model, and Functional benefit representation model.
Relationship basis model: The ability of a brand to connect with its consumer
is explained based on relationship basis model. A consumer forms relationship
with a brand in two ways; the first one is his direct relationship with the brand,
similar to relationship between two persons and the second is through brand
23
personality, which evokes strong feelings and favourable attitudes towards the
brand in the minds of consumers (Aaker, 2012). Along with brand personality,
a customer sometimes forms a ‘strictly business’ relationship with a brand
which is based on the utilitarian benefits offered by the brand. Among the two
types of relationships discussed, the first is a functional or utilitarian benefit-
based and the second is more a symbolic one.
When customers tend to identify human personality traits relevant to
brands, the partnership between the two becomes stronger. When a brand and
its customers become relationship partners in this manner, the actions and
behaviours of the brand can affect the brand-customer relationships
significantly (Aaker, 2012). For example, a customer may feel difficulty in
maintaining a relationship with a brand, if it goes out of stock or if it is not in
line with his expectations (Aaker, 2012).
Self-Expression model: Brand personality helps a user to express himself/
herself. This is achieved by a brand by attaching with itself, feelings and
emotions which are similar to those the customer wants to express (Aaker,
2012). A customer becomes most fulfilled, when the brand helps him to
express his/her exact feeling. According to Aaker (2012), “a warm person will
be most fulfilled when a warm feeling occurs; similarly, an aggressive person
will seek out context where aggression is accepted”. Brand personality
becomes effective in the process of ‘self expression’ only when it fits with the
needs and desires of the customer. When the personality of the brand is in the
right context and fits the consumer’s self-expression needs, any brand
personality can assist in self-expression. However, symbolic and high
involvement product categories might produce larger effects in the self
expression of a customer (Aaker, 2012).
Functional benefit representation model: Both self expression model and
relationship basis model, basically explain the role played by brand personality
in value creation in terms of linking the consumer and the brand. One can
observe a direct role of brand personality in both these models. But, in the case
24
of functional benefit model its role is more indirect. In this case brand
personality adds value indirectly by acting as a medium for representing and
prompting functional benefits along with brand attributes (Aaker, 2012). If a
brand succeeds in using distinctive symbols that have appropriate associations
with it, then the brand may be considered to be successful in executing the
‘functional benefit representation model’ (Batra, Myers and Aaker, 1996). For
instance, in the case of MRF tyres, the brand uses ‘a muscle man character’
which helps the brand hint the functional benefits such as ‘strength and
durability’ offered by the product. Aaker (2012) argues that a brand may be
able to create a feeling of high quality by associating itself with a country or
region.
2.3.2. Anthropomorphism Theory
‘Anthropomorphism’ is said to have its origin from a Greek word
‘Anthropos’ (meaning ‘human’) and ‘morphic’ (meaning ‘form’ or ‘shape’)
(Ranjbar, 2010). It refers to the attribution of human like mental or physical
characteristics to a non-human thing (for instance, my dog loves me) (Epley,
Waytz and Cacioppo, 2007). It can be considered as a process of forming
inferences regarding the unobservable features of a non-human agent (can be
anything like animals, religious deity, mechanical or electronic devices, natural
forces and so on) and is less related to descriptions of its observable behaviour.
The essence of Anthropomorphism lies in the creation of links between
imagined or real behaviour of non-human agents with human-like emotions,
characteristics or motivations (Ranjbar, 2010).
2.3.3. Self Concept Theory
The self concept refers to an internal model of a person that uses self
assesments to define his or her own ideas and beliefs or self-schemas. Features
such as occupation and hobbies, personality, skills and abilities, physical
characteristics, and so on are assessed and applied to self-schemas, which are
ideas of oneself in a particular dimension (For instance, someone who
25
considers himself a ‘daring person’ will associate ‘daring-like’ features to
himself). A group of self-schemas forms one's general self-concept. For
instance, the statement "I am intelligent" is a self assessment that adds to self
concept. But, a statement like "I am sleepy" cannot be considered as a part of
one’s self concept as it is a temporary state (Wylie, 1979).
2.3.4. Self Congruity Theory
Self congruity refers to the extent to which the personality of the
consumer goes in line with that of the typical user imagery of the brand. The
theory proposes that consumer behaviour, to some extent, depends on the
match between the consumers’ self image and the image of the brand they
choose. Self congruity is a very old widely researched topic in marketing (See
for example Levy, 1959; Dolich, 1969; Sirgy, 1982). The researches in this
area showed that consumers have a tendency to opt brands that match more
with their personality over those do not. These researches further demonstrated
the usefulness of this concept in predicting and explaining the concepts such as
customers intention to purchase, brand loyalty, brand attitude, product usage
and so on (Sirgy et al.,1997).
Sirgy (1982) classifies self congruity in to four types, namely, actual self
congruity, ideal self congruity, social self congruity and ideal social self
congruity. Among these four congruity variants, actual self congruity refers to
the match between the present personality of the person and a stereotypical user
of the brand. Ideal self congruity arises out of ideal self concept and the
personality of the brand. It indicates the match between the personalities which
he would like to present as having and that of the brand’s personality. For
example, if someone wants to present himself as having a macho personality,
he may use a Marlboro cigarette. In this case, a match between the ideal self of
the consumer with a stereotypical user of the brand can be seen. The match
between the social self concept and typical user imagery or personality of the
brand results in social self congruity. Social self concept refers to how others
perceive the self of a person. Ideal social self congruity arises out of how others
26
would like to see the self of a person and the personality of the brand. The
research by Sirgy (1982) showed that all the four variants of self congruity
concepts affect consumers’ attitude towards a brand. However, actual and ideal
self congruity is having highest empirical evidence on their influence on
consumer behaviour.
Self Congruity and Brand personality: User imagery of the brand which
forms the basis for self congruity assessment is one among the many ‘direct’
sources that builds the personality of a brand (Aaker, 1997). Brand personality
is a much broader and inclusive concept than self congruity. It can be noticed
that brand personality is derived out of many direct as well as indirect sources.
2.3.5. Personality Theory
According to Goldberg (1993), the personality of a person refers to a
series of dynamic and organised features of him/her that he/she owns and
specifically affects his/her behaviour and motives in various contexts. Different
theories in personality psychology insist on providing a clear structure and
framework of personality and its dimension to make any individual different
from others (Ranjbar, 2010). Among the different models developed by
psychologists, the big five model is considered to be the most accepted frame
work to measure human personality. This framework was based on the lexical
hypothesis (also known as Lexical Approach, or Sedimentation Hypothesis)
developed by researchers such as Gordon Allport and Raymond Cattell in the
1940’s (Opoku, 2005). The lexical hypothesis assumes that the various
personality characteristics that are important in people’s lives will become part
of their language and the important personality characteristics are likely to be
encoded in to language as a single word. Further, these words will be
transferred from one generation to next, through the process of socialization.
2.3.5.1. Big Five Human Personality Dimensions
The five dimensions of human personality and the corresponding traits
are given in the following figure.
27
Figure-2.4. Five-factor model of Human Personality [Source: Adapted from (Opoku, 2005)]
Big Five Factors
Extroversion
ActiveCompetitive DominantEnergetic
LivelyResoluteStrong
Agreeableness
AffectionateAltruisticAuthenticCordialFaithful
GenerousLoyal
Conscientiousness
Conscientious ConstantEfficientPrecise
ProductiveRegularReliable
Scrupulous
Emotional Stability
CalmLevel-headedLight-hearted
PatientRelaxedSereneStable
Tranquil
Openness
CreativeFancifulInformed
InnovativeModernOriginalRecent
Up-to-date
28
2.3.5.2. Human Personality and Brand Personality
Many researchers consider brand personality to be an application of
human personality theory to brands. During the initial years of brand
personality research, researchers depended on human personality scales to
measure brand personality. Further, Aaker (1997), whose work is considered to
be a seminal one in brand personality, relied on different human personality
scales and conceptualisations, to a great extent to conceptualise and develop the
brand personality scale. Also, most of the brand personality scales try to select
items from human personality scales that are relevant for brands.
2.4. Review of brand personality research
This section reviews a few carefully selected cases of brand personality
scale development and discusses them in detail by
1) Reviewing the existing approaches to brand personality measurements –
starting from the earliest (Aaker, 1997) to the most recent (Heere, 2010).
2) Identifying the commonalities and the basis for comparison, and
3) Developing a comparative framework of Brand Personality
Measurement approaches.
Such a comparison can reveal the strengths and weaknesses of each scale and
the specific contexts wherein they will be most suitable. This will reduce the
efforts of future researchers in identifying and choosing the best as well as the
most suitable scale to measure brand personality in a specific context.
2.4.1. Review of select Brand Personality literature
The tendency to attribute human characteristics to brands existed in the
marketing domain as early as 1950s. Azoulay and Kapferer (2003) noticed that
Martineau (1958) used the word to refer to the non-material dimensions that
make a store special or to have a character of its own. King (1970) quotes the
29
research works of J. Walter Thompson Advertising Agency which indicate that
consumers attribute personality characteristics to brands and talk about it.
Prior to 1980s, brand differentiation was mainly on the basis of product
performance. But with the increasing number of products which provided
almost the same set of functional benefits, it became difficult for marketers to
differentiate brands merely on the basis of its performance. In order to
overcome this trouble, the advertising agency Ted Bates introduced the concept
of Unique Selling Personality following the concept of Unique Selling
Proposition created by Rosser Reeves. As a consequence, the account
executives started to use a new item called brand personality in their copy
strategy. From 1970 onwards every advertising agency started giving a
provision for describing brand personality (along with the target, brand promise
and the reason why). The use of brand personality originated as a non-
product- based definition of the brand: it captured all that was not bound to the
product’s use, performance, attributes, and so on. In copy strategies, brand
personality was used as a common, practical, but rather loose, word for
assessing non- product- based, non-functional dimensions of the brand; it
captured the singularity of the source of the product as if it were a person
(Azoulay and Kapferer, 2003).
Sources of Brand Personality
The major difference in the formation of human personality and brand
personality lies in terms of how these perceptions are formed. Human
personality traits are inferred on the basis of individual behaviour, physical
characteristics, attitudes and beliefs, and demographic characteristics (Park et
al., 1986). But in the case of brands, consumers derive perceptions about a
brand’s personality through direct or indirect contact with the brand (Plummer,
1985; Aaker, 1997). According to Aaker (1997), the direct sources of brand
personality includes the brands user imagery, its employees, Chief Executive
Officer (CEO), brand’s product endorsers and indirect sources may include
product-related attributes, product category associations, brand name, symbols
30
or logo, advertising style, price and distribution channel (Batra et al., 1993;
Aaker, 1997).
Brand Personality definitions
Brand personality has been defined by different researchers in different
ways. Aaker (1997) defines it as “the set of human characteristics associated
with a brand”. Yet another popular definition of brand personality is the one
proposed by Azoulay and Kapferer (2003). They define brand personality as
the “the set of human personality traits that are both applicable to and relevant
for brands”. Sweeney and Brandon (2006) tried to define brand personality
based on interpersonal circumplex model and defined it as “the set of human
personality traits that correspond to the interpersonal domain of human
personality and are relevant to describing the brand as a relationship partner.”
These different definitions on brand personality arise from the attempts of
researchers to make the concept more precise and valid.
2.4.2. Measuring Brand Personality
The product or brand personality measurement research can be traced
back to 1960’s (Birdwell, 1968; Dolich, 1969). Early research studies mainly
focused on relationship between consumer self and product personality. For
example, Birdwell (1968) conducted a study using a compiled list of bipolar
items containing appropriate adjectives which described both automobile and
human personalities, to understand the relationship between consumers’
perception about their cars and their self perception. Another earlier study on
product personality was one done by Dolich (1969). It was done with an
intention to understand the relationship between consumers’ actual and ideal
self image and personalities of four product categories namely beer, cigarettes,
bar soap and tooth paste. The researcher used a scale adapted from human
personality scales to study the same. The research works of Malhotra (1981),
on construction of a scale which measured self concepts, person concepts and
31
product concepts and, Karande, Zinkhan, and Lum (1997) on brand personality
and self concept are noteworthy.
Aaker (1997) describes two types of brand personality scales used by
earlier researchers to study how the relationship between brand and human
personality drives consumer preferences. The first one is Adhoc type scales
(consisting of traits ranging from 20 to 300). The main limitation was that these
scales were developed for specific research studies and also, they lacked a
theoretical background. The second type consists of brand personality scales
based on human personality scales. The validity of these scales was questioned
since the antecedents of both constructs were different. Though some of the
dimensions of human personality may be mirrored on brands, the whole cannot
occur (Kassarjian, 1971).
Ambroise, Ferrandi and Merunka (2005) describe three approaches used
by researchers to measure brand personality. The first one is a hierarchical
approach used by Aaker (1997). Second, a Lexical approach used by Caprara et
al. (2001), and the third one is the direct application of human personality
scales to measure brand personality. For example, Ferrandi and Valette-
Florence (2002) applied Saucier (1994) human personality scales to brands and
the results of their study indicated some congruence between purchasers’
personality and the brand personality of the brand they purchased.
2.4.3. Brand Personality Scales (BPS)
In order to obtain a comprehensive understanding of brand personality
measurements, the following brand Personality scales are discussed in detail.
The scales included in this study were chosen on the basis of two criteria: it
should either introduce a new approach to measure brand personality or be
based on a new conceptual definition of the construct. The selected scales were
compared on the basis of their theoretical base, strengths, limitations and cross
cultural validity.
32
1) Aaker (1997) - This seminal work of Aaker is counted to be the first robust,
reliable, and valid scale to measure brand personality.
2) Ambroise, Ferrandi and Merunka (2005) - This scale was developed in
France, considering two product categories and using Factor analysis to find
out the underlying dimensions of brand personality.
3) Sweeney and Brandon (2006) – This is a brand personality scale developed
in Australia based on the Inter personal Circumplex model.
4) Bosnjak, Bochmann, and Hufschmidt (2007) - This scale was developed in
Germany considering multiple product categories and using a person-centric
perspective.
5) Geuens, Weiiters and Wulf (2009) - This scale was originally developed in a
Belgian cultural context considering multiple product categories. The
uniqueness of this scale lies in the fact that it was the only scale which proved
to be cross culturally valid.
6) Kuenzel and Phairor (2009) – This scale was developed in Germany to
measure the brand personality of an automobile brand.
7) Heine (2009) – This is a scale developed using Repertory Grid Method to
measure brand personality of luxury goods.
8) Lee, Soutar and Quintal (2010) – This is a scale developed to measure
destination personality using a Best-Worst Scaling (BWS) method.
9) Heere (2010) – This scale was developed based on a sports brand and it
yielded a two dimensional brand personality scale.
2.4.3.1. Aaker’s BPS (1997)
Aaker (1997) was the first one to generate a robust, reliable and valid scale
to measure brand personality. Her study was set in the United States of
America (USA). She developed a framework of brand personality dimensions
33
by drawing from research on the “Big Five” human personality structure
(Norman, 1963). The major sources used to generate candidate traits included
1) Personality scales used by psychologists
2) Personality scales used by Marketers and
3) Original qualitative research of personality traits associated with a
number of brands
To identify the brand personality dimensions, a total of 631 subjects
rated a subset of 37 brands on 114 personality traits. An exploratory principal
component factor analysis yielded a five factor structure. The facets of each
factor were found by using a principal component analysis and the specific
features of each facet were found out using a cluster analysis. The robustness of
the structure was proved using a series of factor analysis run on subsets of
subjects, and reliability was verified using test- retest correlation and
Cronbach’s alpha. An additional support for the five factor structure was
provided by conducting a confirmatory factor analysis on 180 subjects, based
on 20 brands in 10 product categories and 42 personality traits
Aaker’s BPS is given in the following table. It has five dimensions, 15
facets and 42 traits.
34
[Source: Adapted from Aaker, J.L. (1997)]
Sl.No. Dimensions Facets Traits 1 Sincerity • Down-to-earth
• Honest • Wholesome • Cheerful
o Down-to-earth o Family-oriented o Small town o Honest o Sincere o Real o Wholesome o Original o Cheerful o Sentimental o Friendly
2 Excitement • Daring • Spirited • Imaginative • Up-to-date
o Daring o Trendy o Exciting o Spirited o Cool o Young o Imaginative o Unique o Up-to-date o Independent o Contemporary
3 Competence • Reliable • Intelligent • Successful
o Reliable o Hard working o Secure o Intelligent o Technical o Corporate o Successful o Leader o Confident
4 Sophistication • Upper class • Charming
o Upper class o Glamorous o Good looking o Charming o Feminine o Smooth
5 Ruggedness • Outdoorsy • Tough
o Outdoorsy o Masculine o Western o Tough o Rugged
Table-2.1. Brand Personality Scale developed by Aaker (1997)
35
A major criticism pertaining to Aaker’s BPS is regarding the loose
definition of the construct. Aaker (1997) defined Brand personality as ‘the set
of human characteristics associated with a brand’. The adoption of such a wide
definition has lead to a concept validity problem. The above mentioned
definition encompasses everything related to a human being and applicable to
brand. Azoulay and Kapferer (2003) argue that Brand Personality is a concept
which has close relations with the personality concept in psychology and
therefore its definition must be in line with the same. Psychologists excluded
intellectual abilities, gender and social class from their definitions and scales on
personality whereas Aaker (1997) included many of these. The main
problematic items in Aaker’s scale, as pointed out by Azoulay and Kapferer
(2003), include Competence (which is a cognitive ability and psychologists
donot include it in their personality scales), Feminine (Gender is absent from
psychology scales of personality and also its meaning is tied to cultures), items
related to age and social class. Some other questionable items include Western,
Small town and so on, which belongs to the cultural facet in the brand identity
prism by Kapferer (2012), and brand personality is just a facet of the same.
(Azoulay and Kapferer, 2003)
Another limitation of the BPS generated by Aaker (1997) is that it is
culture-specific or, the five factors identified by Aaker in the case of USA were
not found to be replicable in the case of other cultures. For example, Aaker et
al. (2001) found that Peacefulness replaced Ruggedness in the case of Japan
and in the case of Spain only three factors (namely Sincerity, Excitement and
Sophistication) out of five were found to be applicable. The other two factors in
the case of Spain were Passion and Peacefulness. This limitation of Aaker’s
(1997) scale has lead to a large number of culture-specific studies. A French
one was developed by Ferrandi, Valette-Florence, and Fine-Falcy (2000),an
Italian one was developed by Caprara et al. (2001), a German BPS was
developed by Bosnjak et al.(2007) and so on.
36
Austin, Siguaw and Matila (2003) in their study titled ‘A re-examination
of the generalizability of Aaker brand personality measurement framework’,
found that Aaker’s framework does not generalize the factor structure at the
respondent level (for a specific brand or within a specific product category).
This issue occurred because Aaker (1997) conducted all analysis on data
aggregated across respondents instead of measuring the personality of
individual brands or aggregating data within a specific product category
(Austin et al., 2003).
Another limitation of Aaker’s (1997) scale is that it considers only
positive brand attributes (for example, Down-to-earth, Spirited, Charming and
Imaginative). But this need not be the case always. A brand may also be
described as arrogant, calculating, cocky and so on. For effective brand
management, the brand manager needs to have a full understanding of the
brand personality.
The researchers who conducted studies in brand personality
measurement tried to address these limitations and come out with various
approaches and results which answered many of these criticisms. Some major
research works which followed Aaker’s (1997) seminal work on brand
personality are discussed inthe following sections.
2.4.3.2. Ambroise, Ferrandi and Merunka’s scale (2005)
Ambroise, Ferrandi and Merunka (2005) conducted a study in France to
develop a new brand personality scale. The researchers considered four brands
for this study and used the items from previously published works on brand
personality. The appropriateness of the items was evaluated by a convenience
sample of 387 business students on a seven point Likert scale. Principal
component factor analysis was used with promax rotation. The final result of
the study consisted of 12 factors and 33 items. The validity of the scale was
proved with the help of a confirmatory factor analysis. The key strength of this
37
scale as compared to Aaker’s (1997) scale is that it uses a more strict definition
of brand personality. The scale derived from this study is shown below.
38
Figure-2.5. Ambroise, Ferrandi and Merunka’s brand personality scale (2004)
[Source: Ambroise et al. (2005)]
Brand Personality Dimensions
GLAMOROUSCharming
VoluptuousSophisticated
ELEGANTGood Looking
ElegantStylish
EXCITINGFunny
ImaginativeOriginal
RELIABLESecureRobust
Comforting
NATURALNatural
Environmentally-friendly
SWEETLikable
AffectionateFriendly
CHEERFULSpiritedTrendyJoyful
RIGOROUSDeterminedIntelligentEfficient
MATUREAdult
DiscreetThoughtful
MISCHIEVOUSYouthfulComical
SECUREReliable
Successful
OUTGOINGLively
PopularSporty
39
2.4.3.3. Sweeney and Brandon’s scale (2006)
Sweeney and Brandon (2006) explored the potential to move from factor
analytical model to Circumplex model. The study was conducted in Australia
and used three sources to generate items required for developing the scale,
which included human personality items from Wiggins’ (1979) interpersonal
adjective scale, items from the Trapnell and Wiggins’ (1990) five-factor model
and Aaker’s (1997) brand personality measures. Thirty two judges rated the
appropriateness of these traits based on the interpersonal definition of brand
personality. The results of the study showed that brand personality can be
viewed in less wholesome ways than those traits originally proposed by Aaker
(1997). They proposed that three dimensions namely, agreeableness,
extroversion and conscientiousness from human personality model and the five
dimensions from Aaker (1997) brand personality framework together, are
suitable and appropriate for measuring interpersonal brand personality.
Figure-2.6. Sweeney and Brandon’s Brand Personality Scale (2006)
[Source: Adapted from Sweeney and Brandon (2006)]
Brand Personality
Sincerity
Excitement
Competence
Sophistication
Ruggedness
Agreeableness
Extroversion
Conscientiousness
40
2.4.3.4. Bosnjak, Bochmann, and Hufschmidt’s scale (2007)
Bosnjak et al. (2007) conducted a study in Germany with respect to
commercial brands with a person-centric perspective which explored both
positive as well as negative human personality dimensions which are both
applicable and relevant to brands. The study found that a four factor structure
with 20 items could explain the indigenous German brand personality. The
dimensions found were drive, conscientiousness, emotion and superficiality. Of
these, the dimension drive has two facets namely excitement and boredom. The
use of a person-centric approach has resulted in the emergence of a negatively
connoted factor, superficiality and the facet, boredom (belonging to the factor -
drive). A pictorial representation of the relationships between different factors
and facets is given in the following figure.
41
(Note: + denotes positive relationships, -denotes negative relationships and ?indicates that no specific direction of influence is assumed)
Figure-2.7. Bosnjak, Bochmann, and Hufschmidt’s BPS (2007)
[Source: Adapted from Bosnjak et al. (2007)]
42
The various items associated with each factor are given in the following table.
Table-2.2. Bosnjak, Bochmann, and Hufschmidt’s BPS – Factors and Items (2007)
Sl. No. Factors Items
1 Superficiality 1. Hypocritical 2. Obtrusive 3. Arrogant 4. Selfish
2 Emotion 1. Good-natured 2. Cordial 3. Sentimental4. Loving
3 Conscientiousness
1. Competent 2. Responsible 3. Orderly 4.Reliable
4 Drive 1. Excitement 2. Boredom - Exciting - Small-minded - Adventurous - Bourgeois - Spirited - Boring - Saucy - Old-fashioned
[Source: Bosnjak et al. (2007)]
2.4.3.5. Geuens, Weiiters and Wulf’s scale (2009)
One of the most recent and notable development in the field of measuring
brand personality is the work done by Geuens et al.(2009).They tried to address
the criticisms on brand personality scales and come up with a scale which
consisted of personality items only. The study considered 193 brands and was
conducted among 12,789 Belgium respondents. This study yielded a scale with 12
items and five factors (Activity, Responsibility, Aggressiveness, Simplicity and
Emotionality). The scale was proved to be reliable in the case of between-brand
between-category comparisons, between-brand within-category comparisons and
for between-respondent comparisons. The scale had high test-retest reliability and
cross cultural validity in USA and nine other European countries. The uniqueness
of this work lies in the fact that, it is the only scale which proved to be valid to
measure brand personality irrespective of cultural variations.
43
Figure-2.8. Geuens, Weiiters and Wulf’s Brand Personality Scale (2009)
[Source: Geuens et al. (2009)]
2.4.3.6. Kuenzel and Phairor’s scale (2009)
Based on the warning of Austin et al. (2003) to apply general brand
personality scales to single brands, Kuenzel and Phairor (2009) conducted a study
based on an automobile brand in Germany. Data for the study was collected from
1170 consumers of the brand using a five-point Likert-type scale. The analysis,
carried out using an exploratory factor analysis with principal component analysis
and oblique rotation, yielded a two factor structure with four items describing each
factor. The major advantage of this scale lies in its shortness, which can help in
reducing respondent fatigue as compared to that of Aaker’s scale (Koebel and
Ladwein, 1999).
Brand Personality Dimensions
Activity- active - dynamic- innovative
Responsibility- down-to-earth- stable- responsible
Aggressiveness- aggressive- bold
Simplicity- ordinary- simple
Emotionality- romantic- sentimental
44
Figure-2.9. Kuenzel and Phairor’s Brand Personality Scale (2009)
[Source: Kuenzel and Phairor (2009)]
2.4.3.7. Heine’s scale (2009)
Heine (2009) used a qualitative methodology including Repertory Grid
Method (developed by Kelly, 1955) to develop a scale to measure luxury brand
personality. The need for this research lies in the fact that brand personality is a
key source of symbolic benefits and luxury brands derive their value mainly from
these symbolic benefits. The study was conducted among 31 German millionaires
and a content analysis of the data collected through interview revealed that luxury
brand personality consisted of five dimensions
Brand Personality Dimensions
Sincerity (Successful, Down-to-earth, Honest, Original)
Passion (Spirited, Imaginative, Daring, Passionate)
45
T M D E DT O DC ET ST SG
[Note: T-Traditional, M-Modern, D-Decent, E-Eccentric, DT-Discreet, O-Opulent,
D-Democratic, ET-Elitist, ST-Soft, SG-Strong]
Figure-2.10. Heine’s Luxury BPS (2009) [Source: Heine (2009)]
Further Heine (2009) states that the dimensions presented above are
not independent from each other and that the left hand side of the above figure can
be characterised as quiet and the opposite side as rather loud.
2.4.3.8. Lee, Soutar and Quintal’s scale (2010)
Lee et al. (2010) tried to introduce a new brand personality measurement
approach which requires less time to obtain data and removes some of the
response biases. For this purpose, they used a Best Worst Scaling (BWS)
technique to measure the personality of three destinations namely France, U.S. and
China based on the perceptions of people from Australia, Germany, Brazil and
South Korea. They used personality traits from previous studies on brand
personality and with the help of a partially balanced incomplete block design; they
collected the most and least descriptive terms for each of the three destinations
from the respondents. The results of the study showed that France was seen as
Luxury Brand Personality
Modernity
The temporal perspective of a
brand
Eccentricity
The level of discrepancy from social norms and
expectations
Opulence
The level of conspicuousness of the symbols of
wealth
Elitism
The level of status and exclusivity
Strength
the level of toughness and
masculinity
46
significantly more sophisticated, the U.S. as significantly more rugged and lively
and China as significantly more trendy, lively and genuine, than the other
destinations. Though the study does not provide a rigid scale to measure brand
personality, it introduces a new approach with which brand personality can be
measured.
2.4.3.9. Heere’s scale (2010)
A study was done by Heere (2010) based on the notion that managers can
sum up personality traits since they are responsible to a great extent for the
creation of brand personality. He used the free listing technique to elicit the
adjectives they used to describe their brand. The information gathered from
various managers were processed and summarized in to a final set of items which
was rated by 224 respondents on a seven point scale. An exploratory factor
analysis of the data resulted in the clustering of adjectives around two factors,
namely Game-related (associations that are directly related to elite sport and the
emotions resulting from the game) and Event-related (more generic associations
related to the event). The event-related factor consisted of adjectives like
accessible, warm, cool and attractive and game-related factor included adjectives
such as proud and exciting.
Figure-2.11. Heere’s brand personality dimensions (2010)
[Source: Heere (2010)]
Brand Personality Dimensions
Game-related Factor
Event-related Factor
47
2.4.3.10. Summary of Brand Personality scales
The different approaches to measure brand personality are summarized in
Table-2.3 consisting of the context (that is, the country wherein the study was
done and the number of brands considered for the study), sources of traits (that is,
sources used by the researcher to identify the adjectives needed for the study),
sample description (that is, the number and type of respondents participated in the
study), design and analysis (that is, how the researcher identified the various
dimensions of brand personality) and the major findings of the study.
48
Table-2.3. Summary of Brand Personality scales
Author Context Sources of Traits Sample description
Design and Analysis
Major Findings
Aaker (1997)
Study done in U.S based on 37 brands across multiple product categories
Sources include: -Personality scales used by psychologists(Big Five model – Norman, 1963; Tupes and Christal, 1958; NEO Model - McCrae and Costa, 1989; Big Five Prototypes - John, 1990), Adjective Check List (Piedmont, McCrae, and Costa, 1991), and Inter-Circumplex Model - McCrae and Costa, 1989) -Personality scales used by marketers and academicians (Alt and Griggs, 1988; Batra et al., 1993; Levy, 1959; Malhotra, 1981; Plummer, 1985; Wells
Size: 631 Panel: A national mail panel Subjects: A representation of U.S. people based on gender, age, ethnicity, household income and geographic location
Structure: -Used an exploratory principal component factor analysis to identify the dimensions. - Robustness was established using a series of factor analysis run on subsets of subjects. - Reliability was established using test-retest correlations and Cronbach’s alpha - Confirmatory factor analysis was used to find additional support for the stability of the 5 dimensions
-Identified a five dimensional structure with 15 facets and 42 traits -The dimensions include Sincerity, Excitement, Ruggedness, Sophistication and Competence.
49
et al., 1957) -Original qualitative research of personality traits associated with a number of brands - Traits from a free association test conducted among consumers.
Ambroise et al. (2005)
Study done in France considering four brands, namely Pepsi, Coke, Nike and Adidas
Two sources were mainly used - Existing scales which include Aaker et al., 2001; Caprara et al., 2001; Ferrandi et al., 2000 and Ferrandi and Valette-Florence, 2002. and - Generation of new items by consumers and experts (using nominal group technique)
Size: 387 Panel: University Subjects: Business students
Principal component factor analysis with a promax rotation and confirmatory factor analysis were used for the study
Found the brand personality to have an order 1 structure with 33 items loading on 12dimensions. The dimensions found were glamorous, secure, outgoing, sweet, exciting, elegant, mischievous, cheerful, mature, natural, rigorous and reliable.
Sweeney et al. (2006)
A study done in Australia considering the features of multiple product categories
Three sources were used: - Human personality items from Wiggins (1979), inter-personal adjective scale items from the Trapnell and
Size: 32 Panel: Online survey Subjects: Academics
32 judges rated the appropriateness of a set of traits based on the interpersonal definition of brand personality
-Brand personality can be viewed in less wholesome ways than those traits originally proposed by Aaker (1997) - Study indicated that
50
Wiggins (1990) and five-factor model- Aaker’s (1997)brand personality measures
broadly Aaker's (1997) brand personality dimensions and the agreeableness, extroversion and conscientiousness dimensions from human personality model are suitable and appropriate for measuring Interpersonal brand personality
Bosnjak et al. (2007)
Study in Germany based on 13 brands in multiple product categories
Two sources were used - Items created by eliciting attributes with the aid of original qualitative research -Items translated from Aaker’s (1997) scale
Size: 131 Panel: Online survey Subjects: People with diverse demographic features
Structure: Factor analysis was done with principal component analysis and varimax rotation
-Found that indigenous German brand personality scale consists of Four factors and 20 items. The dimensions includes drive, conscientiousness, emotion and superficiality
Geuens et al. (2009)
Study in Belgium considering 20 brands in stage-1 and 193 brands in stage-2 from
Consists of items from -Aaker’s (1997) scale -items from Costa and McCrae's revised NEO-PI scale (1992), Mervielde's Dutch Big Five version (1992),
Size: Stage-1 - 1235 Stage-2 - 2789 Panel: Online survey Subjects: People with
Structure: Factor analysis was done with principal component analysis and varimax rotation
-A new scale was formulated consisting of 5 factors and 12 items. - The five dimensions found were Activity, Responsibility,
51
multiple product categories
and Saucier's brief version of Goldberg's Big Five markers (1994) -Items generated using brainstorming sections in two focus groups
diverse demographic features.
Aggressiveness, Simplicity, and Emotionality
Kuenzel et al. (2009)
Study done in Germany based on a popular car brand
Traits were taken from Hieronimus’ (2003) German brand personality scale and adjectives frequently used in the company’s brand communications
Size: 1170 Panel: Company’s national customer database Subjects: Product owners
Exploratory and confirmatory factor analyses were used to Validate the actual structure of the brand personality scale
Further evidence has been found in Support of Austin et al. (2003)’s warning to apply general brand personality scales to single brands - Brand personality had a two dimensional (security and passion) structure having 5 adjectives each
Heine (2009)
Study in Germany to measure luxury brand personality
Traits were identified by interviewing 31 German millionaires
Size: 31 Panel: Used a snow ball sampling approach Subjects: German millionaires
- Used a qualitative methodology including Repertory Grid Method. -Used content analysis of the data to uncover 5 personality dimensions
Found that consumers perceive 5 different personality dimensions for luxury brands, namely modernity, eccentricity, opulence, elitism and strength
Lee et al. (2010)
Study done among people
Traits were taken from prior studies of Aaker
Size: 209 to 220 in each country
A partially balanced
Found that destinations differed
52
from Australia, Germany, Brazil and South Korea considering three tourist destinations (France, U.S.A and China)
(1997), Phau and Lau (2000), Aaker, Benet-Martinez and Garolera (2001), Austin, Siguaw and Mattila (2003) and Sung and Tinkham (2005)
Panel: Online survey Subjects: People from Australia, Germany, Brazil and South Korea
incomplete block design was used and respondents were asked to pick the most and least descriptive terms for each destination and best-worst ratio score was calculated to produce trade-off scores
in their personality - France was seen as significantly more sophisticated, the USA as significantly more rugged and lively and China as significantly more trendy, lively and genuine than the other destinations
Heere (2010)
Study done in New Zealand and based on a sports brand
Traits were identified by asking the managers of the organization to use the free listing technique to describe their own brand
Size: Stage-1 - 5 Stage-2 - 224 Panel: Online survey and questionnaires administered to spectators of a specific sports event Subjects: Managers and sports spectators
Two stage study – development and evaluation of brand personality associations. Then an exploratory factor analysis was done to examine the schematic associations the consumers had between the different associations
Two factors were found, namely 1) Game-related factor and 2) Event-related factor
[Source: Original]
53
2.4.3.11. Comparison of Brand Personality Scales
The following table shows a comparison of different brand personality scales made on the bases of major theories
which back up the measurement approach used by the researcher, strengths, limitations and cross-cultural validity of
the study.
Table-2.4. Comparison of Brand Personality scales
Author Major theoretical bases
Strengths Limitations/Criticisms Cross-cultural validity
Aaker (1997)
- Based on big five human personality structure - Defined brand personality as the set of human characteristics associated with a brand
-Strong theoretical background -First robust, reliable and valid scale to measure brand personality
- Does not strictly adhere to the conceptual definition employed(Austin et al., 2003) -Loose definition of the brand personality construct - Encompasses unrelated aspects of brand identity -Only positive brand attributes are considered
Not valid; As a result, several researchers have adapted the scale
Ambroise et al. (2005)
- Defines brand personality as the set of human personality traits associated with a brand
-Uses a more strict definition of brand personality than that of Aaker(1997)
-Generalizability of the scale is questioned since the study is based on very few number of brands from two product categories only
Not checked
Sweeney et al. (2006)
-Based on inter-personal circumplex (IPC) model - Defines brand personality as the set
-Enriches understanding of personality. - Offers more relevance when an in-depth inquiry of brand personality is needed.
- Preliminary and largely conceptual study -Analysis is based on a small sample of judges
Not valid
54
of human personality traits that correspond to the interpersonal domain of human personality and are relevant to describing the brand as a relationship partner
- IPC offers detailed information about specific traits, blends of traits, and interpersonal (consumer– brand) interactions
Bosnjak et al.(2007)
-Used a person- centric perspective which entails an exploration of those positive and negative human personality dimensions which are both applicable and relevant to brands. -Uses brand personality definition of Azoulayetal. (2003)
-Addresses various limitations of Aaker’s (1997) scale -Between subjects variability in identifying the dimensionality of brand personality was taken into account -Based on a stricter definition of brand personality -Negative dimensions are also taken in to consideration
-Relatively low sample size -Developed and tested for German cultural context only -Needs to conduct a random sample approach to advance the generalizability of the study
Not checked
Geuens et al. (2009)
- Based on big five dimensions to personality -Uses brand personality definition of Azoulay et al.(2003)
-Restricts brand personality to human personality traits that are relevant for and applicable to brands. -Can be used for a variety of studies such as aggregate level across multiple brands of different product categories, different
-Since the researchers used a data driven method of selecting and retaining items, there is a chance of deleting useful and meaningful items because they were not associated with one of the dimensions. - The validity and reliability were studied extensively in
Found to be cross culturally valid in Europe and U.S
55
competitors within a specific product category, individual brand level, and cross-cultural studies. - Global companies can use the scale toassess the degree to which their brands have a true global personality
Belgium only. In other countries it was done only in a limited manner. -Nomological validity should be further investigated.
Kuenzel et al. (2009)
Mainly based on the studies of Aaker (1997) and Hieronimus (2003)
-Helps to reduce respondent fatigue since it is a short scale
-Limited number of traits were considered for the study
Not valid
Heine (2009)
Based on the Repertory Grid Method (RGM) developed by Kelly (1955)
- Provides more relevant dimensions with respect to a specific product category -RGM allows describing constructs with a group of words, which enables researchers to decode their varying contextual meanings for different constructs - Delivers relevant binary oppositions, which correspond to both the human information processing and the requirements for the design of a brand identity and positioning
-Influenced by subjective interventions of the researcher
Not valid
Lee et al. - Uses Aaker’s (1997) -As there is only one way to - Introduces an approach, not a Not checked
56
(2010) definition of brand personality -Uses aBest-Worst Scaling (BWS) invented by Jordan Louviere.
choose something as best (or worst), the method eliminates response style biases - Provides a context to each answer, by asking people to make choices among relevant options that capture their trade-offs
scale to measure brand personality
Heere (2010)
-Based on Psychological Meaning strategy of Friedmann (1986)
- Helps to capture the brand personality associations as well as provides the researchers with valuable insights on both the organization's ability to represent that particular association, as well as capturing the relative importance of those associations to a particular set of consumers
-Study is on a specific product category and the respondents were mainly women, hence generalizability is limited
Not valid
[Source: Original]
57
The scope of the above review was limited to scales which introduced a
new approach to measure brand personality. Hence, in the following section, a
brief discussion of the most important and noteworthy brand personality
researches is presented. This includes category-specific brand personality scales,
country-specific brand personality scales, brand personality scale for non-profit
sectors, brand personality scale in the online context and so on.
2.4.4. A review of the applications of brand personality concept
A compendium of applications of brand personality concept in the different
realms of knowledge is comprehended and discussed here to portray the
advancement of this concept in the marketing domain. This section of the
literature review focuses on ‘micro’ approaches to brand personality scale
development, which deals with the development of category-specific or product-
specific brand personality scales. An effort is also made to highlight how different
brand personality scales have been used for further research in marketing. The
different categories discussed here consist of destinations, restaurants and so on.
2.4.4. 1. The case of destinations
Research works of Ekinci and Hosany (2006) revealed that brand
personality can be attributed to destinations and it can serve as an effective and
strong tool to differentiate tourist destinations (Murphy, Moscardo and
Benckendorff, 2007). The application of the concept of brand personality to
tourism destinations yielded the concept of destination personality. It is defined as
“the set of personality traits associated with a destination” (Ekinci and Hosany,
2006). A study conducted by Ekinci and Hosany (2006) has given empirical
evidence to prove that tourists attribute human characteristics to destinations.
Destination personality studies have revealed that this concept has a positive
impact on travellers’ intention to recommend (Ekinci and Hosany, 2006), intention
58
to return and word of mouth (Ekinci, Sirakaya-Turk and Baloglu, 2007). Several
destination personality studies (Henderson, 2000; Santos, 2004; Ekinci and
Hosany, 2006; Hosany, Ekinci and Uysal, 2006; Ekinci et al., 2007; Murphy et al.,
2007; Prayag, 2007; Sahin and Baloglu, 2009; Usakli and Baloglu, 2011) have
been done to understand the various aspects of destination personality.
Ekinci and Hosany in 2006 undertook a study on destination personality to
understand the application of brand personality concept in the tourism domain.
This research was conducted among 250 tourists from Britain, based on the last
destination they visited, with the help of an adapted version of Aaker (1997) brand
personality scale with 27 items and using a Likert scale with five points to capture
the descriptiveness of each item in the scale. They found that destination
personality is a three dimensional construct consisting of sincerity, excitement, and
conviviality. Further, they found destination personality can help in creating a
positive destination image in the minds of travellers and can help in increasing
their intention to recommend the destination. This study identified that the
conviviality dimension plays a moderating role in travellers’ intention to
recommend the destination.
Yet another study, done by Hosany et al. (2006) involved 148 British
tourists as respondents, relied on the last destination they visited as a stimuli for
making their response in a 27 item questionnaire using a five-point Likert scale.
As in the previous research, this work also yielded three dimensional destination
personality structure with sincerity, excitement and conviviality being the
dimensions. The findings of this research suggested that, destination image and
destination personality are two related constructs and destination image is more
encompassing while destination personality is a part of the affective component of
destination image. In general branding literature, one cannot find an empirical
study examining the relationship between brand image and brand personality and
that makes Hosany et al. (2006)’s work a novel one.
59
Another destination personality study was the one done by Ekinci et al.
(2007) in Turkey among 365 tourists of German origin. It was carried out using a
20 item destination personality scale with a five point Likert type rating as
recommended by Ekinci and Hosany (2006). The same three dimensions namely
sincerity, excitement and conviviality were identified as destination personality
dimensions. Their study revealed that host image has a positive impact on
destination personality. This research also revealed the role played by destination
personality in influencing tourists’ intention to return and word of mouth
promotion of the destination.
Considering two destinations, namely Cairns and Whitsunday Islands in
Australia, a survey was conducted using a 20-item questionnaire developed based
on Aaker’s (1997) brand personality scale, among 480 tourists who visited
Queensland by Murphy, Moscardo and Benckendorff (2007). This study found
Cairns Islands to possess a three dimensional personality (sincere, sophisticated
and outdoorsy) while Whitsunday was perceived to possess a four dimensional
personality structure consisting of upper class, honest, exciting and tough. Their
study also provided some empirical evidence for the use of brand personality to
differentiate destinations.
Murphy et al. (2007a), further conducted a study among 277 visitors to
Whitsunday Islands in Queensland in Australia using a 20 item scale developed
from Aaker’s (1997) brand personality scale, and identified four dimensions of
destination personality, namely sincerity, excitement, sophistication and
competence, and ruggedness. This study provided evidences for relationship
between destination personality, travel motivation and self congruity. The findings
of the study suggested that tourists’ actual visitation and intention to visit did not
have any relationship with the personality of the destination.
60
Murphy, Benckendorff and Moscardo (2007b) did a study among 277
tourists to Whitsunday Islands in Australia replicated a four dimensional
personality of the Island as identified by Murphy et al. (2007a). This study also
identified associations between high self congruity levels and destination
personality.
Lee and Kaplanidou (2011) conducted a study to understand the influence
of Beijing Olympics in changing the destination personality perception of China
among leisure travellers from the USA. The study found no substantial change in
the collective perception of destination personality of China among the sample.
Usakli and Baloglu (2011) conducted a study among 382 tourists to Las
Vegas. Their research revealed the tendency of tourists to ascribe personality
characteristics to destinations. Also, this study identified a five dimensional
personality structure for the destination personality of Las Vegas and it consisted
of vibrancy, sophistication, competence, contemporary and sincerity. The study
also found that the above mentioned dimensions had a positive impact on tourists’
intention to recommend and intention to return. Another valuable contribution of
this research is that: it gives support to self congruity theory in the tourism context
and identifies that visitors’ behavioural intentions are influenced by actual and
ideal congruity.
Upadhyaya and Makarand (2012) studied the relationship between
destination image and destination personality in an empirical manner. This
research was based on 200 respondents and a scale consisting of items from
different past researches. The results found that the image of a destination is
influenced by the personality perceptions of the tourist regarding that destination.
Xie and Lee (2013) conducted a research in Beijing among 500 foreign
tourists and identified four dimensions, namely competence, excitement,
sophistication and ruggedness as the building blocks to create a strong destination
61
personality. This study also revealed that these personality dimensions play a key
role in projecting the personality of the destination and personality dimensions
such as sophistication, excitement and competence play an important role in
determining the visitors’ behavioural intentions.
Kim and Lehto (2013) tried to compare the difference in the projected and
perceived personality dimensions of South Korea. The projected personality was
measured by conducting a content analysis study of the official Korean tourism
website and the perceived personality dimensions were extracted by analysing the
responses obtained in a survey conducted among the Americans who visited South
Korea. They identified seven personality dimensions namely Family orientation,
Sincerity, Competence, Uniqueness, Excitement, Ruggedness and Sophistication.
2.4.4. 2. Online destination personality studies
Pitt, Opoku, Hultman, Abratt and Spyropoulou (2007) conducted a study to
understand how 10 select African destinations project their personality online. For
this study they developed an inventory of synonyms to Aaker’s (1997) scale and
identified personalities of each country with the help of a computer aided content
analysis. This was the first study which demonstrated a method to measure online
destination personality. The results of theresearch revealed that some of the
African destinations considered for the study communicated a few of the
dimensions in Aaker’s (1997) brand personality scale, but many of them failed to
do the same.
Prayag in 2007 conducted a research based on two destinations namely,
South Africa and Cape Town, based on a sample of 85 international tourists who
visited Cape Town. Unlike other destination personality researches in which a
structured questionnaire was employed to understand the personality of the
destination under consideration, he used a qualitative research method relying on
62
projective techniques and in-depth interview method. Though this study did not
yield any personality dimensions, it found that projective technique could be used
to elicit the destination specific personality traits.
2.4.4. 3. Country personality studies
D’Astous and Boujbel (2007) conducted a study to create a country
personality scale based on responses from Canadians who speak French. They
used a scale developed by incorporating traits from various previous researches in
the field and employed a five-point bipolar scale. They defined country personality
as “the mental representation of a country on dimensions that typically capture an
individual's personality”. The study unveiled six dimensions of country personality
namely, agreeableness, wickedness, snobbism, assiduousness, conformity and
unobtrusiveness. They developed two versions of the scale, a full version which
consisted of 37 items and a short version consisting of 24 items.
Sahin and Baloglu (2009) identified five dimensions of destination
personality, namely sincerity, originality and vibrancy, cool and trendy,
competence and modernity and conviviality. This study was done at Istanbul,
Turkey among 272 international tourists who visited this destination using a
structured questionnaire developed by taking 23 items from Aaker’s (1997) brand
personality scale and five items identified through content analysis of travel
brochures and web sites on Istanbul. This study revealed a difference in the
perceptions of different nationals in perceiving the same destination.
Stokburger-Sauer (2011) conducted a study among 421 potential German
tourists to the holiday destination of the Republic of Ireland. This research
revealed that national brand identification and congruence between the tourists’
personality and that of the nation, has a strong influence on the visit intentions.
This work also found that brand advocacy is strongly influenced by nation brand
embeddedness and that it serves as a predictor of brand advocacy.
63
2.4.4. 4. City personality studies
Kaplan, Yurt, Guneri and Kurtulus (2010) tried to understand the use of
brand personality dimensions in differentiating cities. They developed a city
personality scale for this purpose. This study was conducted in Turkey and was a
two-stage study in which the first phase consisted of generation of 410 items to
describe city personality (by means of a qualitative study and a literature survey).
And, in the second phase, the initial item pool was reduced to 87 items based on
judgement by the researchers. The questionnaire was administered among 898
students to identify the city personality dimensions. This study employed a five-
point Likert scale for rating. The new scale consisted of six dimensions, namely
excitement, malignancy, peacefulness, competence, conservatism and ruggedness.
A detailed presentation of this scale is given in the diagram below.
Figure-2.12. Brand personality dimensions for cities
[Source: Kaplan et al. (2010)]
2.4.4. 5. Geographic personality studies
Aiken, Campbell and Koch (2013) conducted a study to understand how the
personality of a team and personality of the home destination of the team
(geographic personality) are related. This research was done in the USA and
considered 10 teams which participate in the national football league. Responses
Brand Personality Dimensions for Cities
Excitementpassionateoutgoingfemininesympathetic
Malignancyunreliablearrogantself-seeking
Peacefulnesscalmdomestic
CompetenceAuthoritariansophisticated
Coservatismreligiousuneducated
Ruggedness
64
were collected from 434 people. For the purpose of this study, they relied on the
definition of geographic personality put forward by Aiken, Koch and Madrigal
(2000). They defined the construct as “the inferred set of human attributes
identified with a geographic location”. The findings of the study revealed a high
level of congruity between team personality and geographic personality.
2.4.4. 6. The case of restaurants
Austin, Siguaw, and Mattila (2003) conducted a study to check the
generalizability of Aaker’s (1997) brand personality framework. For this study
they considered nine restaurant brands as stimuli and a set of 247 student
respondents were considered. Respondent evaluation of restaurant brands were
collected using Aaker’s (1997) brand personality scale and using a five-point
Likert scale. A confirmatory factor analysis of the data indicated a poor fit with
Aaker’s (1997) framework. Based on their findings, they urged researchers to be
cautious while employing Aaker’s scale, as it may not be appropriate in the case of
individual brands or in the case of different product categories.
Murase and Bojanic (2004) conducted a study with an intention to
understand how the restaurant brand personality varies across cultures. This study
was done in the USA and Japan using Aaker’s (1997) brand personality scale and
considering three popular restaurant chains. The findings of this study exhibited
little difference across cultures in the personality of the three restaurants
considered.
Opoku, Abratt, Bendixen and Pitt (2007) tried to demonstrate a means by
which small and medium sized enterprises in the restaurant industry can evaluate
their projected personality on their websites. This study helped different
restaurants compare their positioning strategy with that of the competitors.
Aaker’s (1997) brand personality scale was used for this study and the results of
the work revealed that the Small and Medium enterprises used their websites
effectively to communicate their personality.
65
Ramaseshan and Tsao (2007) investigated how the brand concept played a
moderating effect in the relationship between perceived quality and brand
personality. The study conducted employing Aaker’s (1997) brand personality
scale revealed that excitement and sophistication dimensions indicated a higher
association with perceived quality than the rest of the dimensions, namely
sincerity, competence and ruggedness. The study helped to prove brand
personality as a viable substitute to price as a cue for quality.
A laboratory experiment research done by Magnini and Thelen (2008)
using a set of business school students tried to examine the impact that music
creates in the consumer perception of brand personality. The findings of the
experiment indicated that music can influence the brand personality perceptions of
customers in a restaurant environment.
Musante, Bojanic and Zhang (2008) conducted a study to develop a
modified version of the brand personality scale that will be more relevant in the
restaurant context. The research employed a student sample to reduce the Aaker’s
(1997) brand personality scale to be more relevant in the restaurant context,
considering three brands as stimuli. Analysis done at this phase yielded a seven
factor structure with 67.73% of the variance being explained. However, instead of
proceeding with this structure of brand personality, they set brand personality
dimensions at five, citing Aaker’s (1997) scale and obtained a five-factor structure
which explained around 62.07% of variance. The researchers conducted several
iterations of factor analysis to reach to a final set of 18 item scale of restaurant
personality, which performed better than the first version of 42 item scale.
Lee, Back and Kim (2009) conducted a study among 475 diners at a chain
restaurant to find that Aaker’s (1997) brand personality scale was applicable in the
case of restaurants.
66
Lee and Back (2010) conducted a study relying on Aaker’s (1997) brand
personality dimensions to understand the relationship between trust, brand
personality and brand loyalty. Investigation in to these aspects was done
considering the upper upscale hotel segment and the findings of the study revealed
the mediating role of trust in the relation between brand loyalty and brand
personality. This study also indicated user imagery as the strongest predictor of
brand personality.
Kim, Magnini and Singal (2011) conducted a study with 336 respondents to
understand whether brand personality has any positive impact on restaurant
chains. The results of their work revealed that diners’ personality perceptions on
restaurant chains have a positive impact on attitudinal loyalty and brand
preference. The study also provided empirical evidence for the positive influence
of brand preference on attitudinal loyalty and positive influence of attitudinal
loyalty on word of mouth.
2.4.4. 7. Sports brand personality
Carlson, Donavan and Cumiskey (2009) conducted a study among 162
university students who were participants of a sports marketing class. An adapted
version of Aaker’s (1997) scale was employed to collect their responses. The
study indicated that respondents tend to attribute sport brands with unique
personalities and this had a positive influence on their identification of that sports
team.
Kang (2013) developed a five dimensional scale to measure sport brand
personality. The dimensions in this scale included Conscientiousness, Honesty,
Openness, Agreeableness, and Emotionality/Extraversion. This study was
restricted to sport brands in professional leagues in the USA.
67
2.4.4. 8. The case of Sports organization Brands
Smith, Graetz and Westerbeek (2006) employed Aaker’s (1997) brand
personality framework to evaluate its applicability in the case of a sports
organization in Australia. This research gathered responses from 413 respondents
and the factor analysis of their responses revealed that Aaker’s (1997) scale was
found to be less appropriate in the case of sports organization brands. They
suggested a six dimensional frame work to measure sports organization
personality, which consisted of an addition of Innovation dimension to Aaker’s
(1997) brand personality framework.
2.4.4. 9. The case of non-profit organizations
Venable, Rose, and Gilbert (2003) developed a scale to measure brand
personality of non-profit organizations. A mix of qualitative and quantitative
approaches was adopted to develop the scale. The qualitative study revealed that
people attribute human personality traits to non-profit organizations. They
proposed a four dimensional frame work consisting of nurturance, sophistication,
integrity and ruggedness as dimensions. A study by Venable, Rose, Bush, and
Gilbert (2005) found that current and potential contributors to non-profit
organizations attribute personality traits to them and their tendency to contribute
relies to some extent on the personality of the brand.
2.4.4.10. The case of media brands
Kim, Baek and Martin (2010) created a news media brand personality scale
by considering a set of news paper brands and relying on a sample of 444 college
students. The study identified a five dimensional structure with factors such as
trustworthiness, dynamism, sincerity, sophistication and toughness.
68
Valette-Florence and de Barnier (2011) created a print media brand
personality scale considering 24 print media brands and a total of 780 respondents.
An eight dimensional structure (with 27 items) consisting of wisdom, conventional
character, misleading character, natural, agreeable, assertive character,
seduction and elegance formed the scale.
2.4.4.11. Corporate brand personality
Keller and Richey (2006) developed a corporate brand personality scale to
measure the personality dimensions of corporate brands. They defined corporate
brand personality as “the characteristics or traits of the employees of an
organization as a whole”. The six traits and three dimensions identified were
Passionate and Compassionate which constituted the dimension Heart, Creative
and Disciplined constituted the Mind dimension and Agile and Collaborative
formed the Body dimension.
2.4.4.12. Online brand personality
Okazaki (2006) tried to capture the personality dimensions which American
firms intended to communicate to their customers with the help of content
analysis. This study considered 270 websites of 64 American brands which were
developed for countries like the U.S., France, Germany, UK, and Spain. A five
dimensional framework with dimensions affection, popularity, excitement,
sophistication and competence were identified as the online personality
dimensions of American brands with the help of a principal component analysis.
2.4.4.13. Website Personality
Poddar, Donthu and Wei (2009) tried to extend the general marketing
concept of customer orientation of sales people to the online context or more
specifically to the internet marketing context. For the study, they employed an
69
adapted version of D'Astous and Levesque (2003) store personality scale namely
website personality scale and an online version of Saxe and Weitz (1982) scale
was used for measuring customer orientation. The study indicated that the impact
of a websites’ personality on the purchase intention of a consumer is mediated by
the perceived quality of the website.
2.4.4.14. Retailer personality
Retail brand personality studies can be marked as the beginning of brand
personality concept. It is supposed to be introduced by Martineau (1958) who
wrote an article titled ‘The personality of the retail store’ in the Harvard Business
Review journal. However, the term was used in a much broader sense and
captured many aspects of store image as well.
d’Astous and Le´vesque (2003) conducted a research to develop a measure
that captures the personality of a store. An initial version of the store personality
scale was constructed with the help of responses from 226 adult respondents. It
consisted of 34 items falling under five different dimensions namely genuineness,
enthusiasm, sophistication, solidity and unpleasantness. A short version of the
scale was then created which consisted of 20 items and five dimensions, which
helped in increasing the respondent friendliness of the tool.
Beldona and Wysong (2007) tried to explore the role of the personality of a
store in influencing the consumer perception regarding store brands and the
variance occurring in their response after experiencing the product. The findings
indicated that as compared to store brands, national brands had an upper hand;
however the real experience of the products diminished those differences.
Zentes et al. (2008) conducted a study in Germany with a sample of 1337
respondents and found Aaker’s (1997) scale to be applicable in the German
70
retailer store context and also found a relationship between the brand personality
construct and customers loyalty towards the store.
Brengman and Willems (2009) explored shoppers’ perceptions of fashion
store personality. For achieving this objective, they conducted a qualitative
exploratory research among 70 Belgium adult fashion store shoppers with the help
of semi-structured in-depth interviews. A convenience sampling method was
adopted for sampling selection in this research. They relied on d'Astous and
Lévesque (2003)scale for data collection. The findings of this study indicated that,
in the case of fashion stores, design of the store and store environment play a key
role in determining the fashion store personality. This study also explored the
attributes of the store that induces each of the five store personality perceptions
namely unpleasantness, solidity, enthusiasm, sophistication and genuineness. The
major attributes identified includes; consumer perception of pricing in the store,
quality of the service and product in the store, reputation of the store, corporate
social responsibility activities of the store, level of service offered by the retailer,
product assortment in the store and staff in the fashion store.
Willems, Swinnen, Janssens and Brengman (2011) argued that a general
brand personality scale like Aaker (1997) or a general store personality scale
similar to d'Astous and Levesque (2003) will be less relevant in the case of fashion
stores. For this research, they carried out a three phase study in which the first
phase consisted of a repertory grid analysis based on responses collected through
interview from 51 respondents, with an objective to identify the attributes that
shoppers attribute to a fashion store. In the next phase,a purification was done with
a sample of 481respondents and using an exploratory factor analysis, they
identified a five dimensional fashion store personality frame work with
dimensions named as conspicuousness, chaos, agreeableness, innovativeness and
sophistication. In phase three of this study, they examined self congruity principle
71
in the context of fashion store personality and found supporting evidences for this
concept in the fashion retailing context.
2.4.4.15. Drug brand personality
Leonard and Katsanis (2013) conducted a study to understand the consumer
perception of personality of drug brands and the scale was named as ‘drug brand
personality scale’. About 483 respondents from the USA participated in the study.
A total of 15 drug brands were considered and respondents were required to rate
them based on 22 personality traits. The result of the study unveiled a two
dimensional drug brand personality scale with competence and innovativeness as
the two distinct dimensions.
2.4.4.16. Industrial brand personality
Herbst and Merz (2011) developed a scale to measure Industrial brand
personality. This basically dealt with the personality of business-to-business
brands.
2.4.4.17. Country-specific brand personality scales
Aaker, Benet-Martinez and Garolera (2001) developed scales to measure
brand personality dimensions in the Japanese and Spanish contexts. In Japan, they
followed a multi-stage procedure to develop the scale which started with an
indigenous item generation phase followed by the main study for identifying the
major brand personality dimensions in the Japanese context. In the main study,
they considered 25 brands (both product and service) and 1495 respondents, which
represented the Japanese population in terms of gender, marital status, age,
occupation and education level. Their research identified a five dimensional
Japanese brand personality structure consisting of excitement, competence,
peacefulness, sophistication and sincerity. In their research to identify personality
72
dimensions of Spain, 692 respondents were included and a total of 25 brands were
considered as stimuli. A factor analysis of the responses revealed a five factor
structure with dimensions namely excitement, sophistication, sincerity,
peacefulness and passion.
A study conducted by Muniz and Marchetti (2012) among 1,302 Brazilian
consumers identified a five dimensional brand personality structure in the
Brazilian context. A total of 24 brands (consisting of both product and services)
were selected as stimuli in the main study. The dimensions identified were
credibility, joy, audacity, sophistication and sensitivity.
Sung and Tinkham (2005) tried to identify culture specific dimensions in
the American and Korean contexts. For this they selected well known global
brands that were common in both countries. They tried to create a general brand
personality scale by considering 18 product and service categories. To identify the
U.S. brand personality structure, a survey was conducted among a total of 320
university students, using a list of 80 carefully selected personality traits based on
a previous exploratory study. A seven point Likert type scaling was employed in
the study. An exploratory factor analysis of the data yielded an eight dimensional
structure with dimensions such as likeableness, trendiness, competence,
sophistication, traditionalism, androgyny, white collar and ruggedness. For
creating Korean personality measure, a survey was conducted among 337 Korean
undergraduate students. The analysis procedures in the case of Korea were similar
to that of the U.S. case. The Exploratory Factor Analysis conducted on the data set
identified an eight factor structure which explained around 62.5% variance. The
dimensions found were competence, trendiness, likeableness, passive likeableness,
sophistication, accountability, ruggedness and traditionalism.
73
2.4.4.18. Theory building studies using brand personality
In this section some of the important studies based on different brand
personality scales which explores and establishes relationship between brand
personality and other concepts such as loyalty, trust and so on
A research work of Kim, Han, and Park (2001) found that customers’
identification of a brand is positively affected by the different values of brand
personality, such as its distinctiveness, attractiveness and self expressive value.
Diamantopoulos, Smith and Grime (2005) employed Aaker’s (1997) brand
personality scale to examine the impact of brand extensions on the personality of
the brand. The results of this experimental study found that brand extensions do
not adversely affect the personality of the core brand.
Freling and Forbes (2005) provided empirical evidence for the impact of
brand personality on different brand associations.
Fennis and Pruyn (2007) examined whether the personality perceptions
regarding a brand is carried over and attributed to its owner by consumers. The
research found that the influence of brand personality on impression formation
exists but varies depending on the situation.
Boudreaux and Palmer (2007) studied the impact of the consumer
perceived personality of a brand on the purchase intention, in the case of wine
purchase. For this research, the researchers collected responses of 90 experiment
labels of wine brands which varied on their features such as design layout, colour
and so on. The results revealed that brand personality explained only half of the
variance in purchase intention. Further, the study identified the facets such as
74
charming, spirited, successful and up-to-date to be highly correlated with purchase
intent.
Louis and Lombart (2010) conducted a study to understand how trust,
attachment and commitment to a brand are affected by the consumer perceived
dimensions of brand personality. For this purpose, they carried out a study among
348 young French consumers in the age group 19-23 years, considering Coca cola
brand as a stimuli. They found that all the brand personality dimensions had
impact on at least one of the three, namely trust, attachment and commitment
towards the brand.
Müller and Chandon (2003) conducted an experiment research to
understand the effect of forced exposure to a brand website on the personality of
the brand. The study revealed that two dimensions of personality namely,
sincerity/confidence and youthfulness/modernity were increased significantly with
exposure to the websites. The study also revealed that if a website evokes a
positive feeling in the mind of a visitor, then he/she may perceive that brand to be
a more modern or younger one. A positive feeling regarding the website can make
the customer perceive the brand to be more sincere and trustworthy one.
Cervera-Taulet, Schlesinger, and Guillen (2013) tried to understand how
advertising affects the personality of a brand in the airline industry of Spain by
conducting a research relying on Aaker’s (1997) brand personality scale. The
research found that only the sophistication dimension among the five dimensions
was influenced by advertising.
Huang and Mitchell (2014) tried to understand the function of brand
personification and imagination in the creation of brand relationships through an
experimental study which employed a 2*2 factorial design and the research was
75
conducted among a total of 468 subjects. The study revealed that brand
personification can moderate the effect of self expansion theory.
2.4.4.19. Celebrity endorsement and brand personality
Celebrity endorsement research can be broadly classified into two streams,
namely source credibility stream of research and congruence research (Roy and
Moorthi, 2012). Congruence research is also known as match-up studies. Three
main dimensions of source credibility include trustworthiness, expertise and
attractiveness.
Congruence studies: This stream of research deals with the match between the
brand and celebrity and is known as match-up hypothesis. The match-up
hypothesis postulates that a celebrity advertisement will be effective when there is
congruence between the celebrity and the brand advertised (Roy and Moorthi,
2012).
The meaning transfer model: The meaning transfer model was proposed by
McCracken, who postulated that celebrities carry and convey a set of meanings to
consumers. Further, they offer a set of personality traits and lifestyle (Roy and
Moorthi, 2012). According to McCracken (1989) endorsers can bring in and
transfer symbolic meanings to endorsement process. These symbolic meanings
may include personality, life style types, age, gender, social class and so on. One
of the rationales behind the use of celebrity endorsers by advertisers could be
obtained from the arguments of Fortini-Campbell (1992). According to him, like
human beings, products also possess personality and consumers will prefer to
consume or own products that have a personality similar to theirs or a personality
they aspire to possess, probably those of their family members or celebrities they
like.
76
Celebrity endorsement is considered to play a crucial role in building brand
personality as it can be used to attribute instant personality to brands (Dickenson,
1996).
2.4.5. Brand Personality Researches in the Indian Context
Purkayastha (2009) studied brand personality of four brands namely 7 up,
Samsung, Motorola and Raymond using a 15-item scale based on Aaker’s (1997)
brand personality scale. A total of 100 responses were collected for each brand and
the brand personality of each brand was measured using a factor analysis of the
responses.
Thomas and Sekar (2008) conducted a study to check the validity of
Aaker’s (1997) brand personality scale in the Indian context. Their study found
Aaker’s (1997) scale to be not fully applicable in the Indian context.
Das, Datta and Guin (2012) developed a department store personality scale.
For this purpose they defined department store personality as “a consumer's
perception of the human personality traits attributed to a department store”. The
study consisted of an initial item generation and purification stage using
qualitative methods followed by an empirical study which identified department
store personality dimensions. Out of the three to seven factor structures emerged
as personality dimensions, the researcher opted for a five dimensional framework.
The dimensions of department store personality consisted of dependability,
authenticity, sophistication, empathy and vibrancy. Further, this work also found
each of the above mentioned dimensions had a positive impact on each retailer
equity dimensions (awareness, associations, perceived quality and loyalty with
respect to the retailer). The only exception to this was between empathy and
retailer loyalty.
Das (2013a) tried to understand how store loyalty is influenced by retailer
personality. This study was done in Kolkata, India, using a systematic sampling
77
procedure, and relying on non-food retailers as stimuli. The results of the study
indicated that retailer personality had both direct and indirect (based on purchase
intention and perceived quality) impact on retailer equity.
A research was undertaken by Das (2013b) to find how the personality of a
retailer and match between shoppers’ self and retailer’s personality affect the
loyalty of the shopper towards the store. The findings of this research work
indicated a positive impact of the symbolic benefits on store loyalty.
Das, Guin and Datta (2013) conducted a study to understand the impact of
store personality antecedents on store personality dimensions. The findings of the
research suggested that all the various antecedents of store personality such as
customers’ general attitude regarding the retailer, ambience of the retail store,
product assortment in the store and brand name exhibited a positive impact on the
empathy dimension of store personality. Whereas quality of the product, customer
attitude towards the retailer, ambience of the store, word of mouth and brand name
indicated positive impact on the dependability dimension. In the case of
authenticity dimension, the antecedents that exhibited a positive impact included
ambience of the store, assortment of products, brand name and attitude of
customers towards the retailer, product quality and quality of service in the store.
Finally, for vibrancy dimension, the antecedents with positive impact were
customer attitude towards the retailer, brand name, product variety and style and
store ambience.
2.5. Chapter discussion and conclusion
Further studies are required in brand personality measurement to address
questions like the appropriateness of a general BPS to measure personality of a
specific brand, to identify the product and service categories for which a general
BPS will be suitable and the categories which require category-specific BPS, to
78
test whether the existing scales can measure the brand personality of a new
product, and so on. Some of the scales discussed here have to be tested with
respect to their reliability and validity. The cross cultural validity of most of the
brand personality scales need to be checked. Hence, future researchers can
investigate the applicability of these scales in different cultural contexts and
various validity and reliability tests may also be done with these scales under
various situations.
2.5.1. Gaps considered for the present study
Spurred on by an increasing body of research in marketing and the
importance of brand personality in it, it is apparent that the research related to this
construct will not subside. Driven by the generally accepted importance of the
brand personality construct in marketing and the lacuna in this area in India and
the limitations of the existing scales, the requirement of an India-specific and
product brands -specific brand personality scale is identified.
2.6. Summary of the literature review
This extensive review of existing literature brings to light that there exists
significant evidence to support the necessity of brand personality research in
marketing in general and branding in particular. There is a scarcity of research
which focuses on product brand personality and brand personality researches in
the Indian context. However, many attempts are made by researchers to contribute
to the knowledge of brand personality construct. It is also obvious from the
reviews that there is a necessity to develop a product brand personality scale,
which will be of much use to advertisers and brand managers, since it will be more
precise and will improve their measurement of the personality of their product
brand. Also, the literature reveals the necessity of an India-specific brand
personality scale which will address the validity issues of the current scales to a
great extent.
79
Chapter 3 RESEARCH METHODOLOGY
Designing Scale Development Framework
The objective of this chapter is to review the important approaches and
methodologies adopted by researchers to measure brand personality and to design
a research methodology for developing the product brand personality scale.
3.1. The concept of measurement
Measurement is an important aspect of science and social science research.
People generally acquire knowledge in their everyday life by the process of
observation of various happenings around them, in this process; they tend to
quantify what they observe to make a sense and record of it, which one can refer
as measurement in social science. The scientific method to measure concept in
social science and psychology is termed as psychometrics (DeVellis, 2011).
3.2. Approaches to brand personality measurement
Researchers used both qualitative and quantitative approaches to measure
brand personality. Many researchers like Geuens et al. (2009) used a mixed
method approach, wherein both quantitative and qualitative methods were
employed in determining the personality of the brand. For instance, in the case of
Geuens et al. (2009), the researcher used qualitative tools like free elicitation and
expert opinion for initial item generation and initial item reduction; while a
quantitative approach was employed for determining dimensionality, validity and
reliability. It can be observed that, in the item generation phase most of the
researchers who works on scale development tries to include some qualitative
techniques. A few of the prominent qualitative techniques employed in brand
personality research include content analysis, focus group discussion, interviews,
free elicitation studies and so on.
80
Generally, free elicitation method or word association test are employed at
the item generation phase. The purpose of conducting these techniques is to
generate or identify new items. For instance, when a new scale is created to
measure destination personality, many of the items selected from a general brand
personality scale like Aaker’s (1997) may not be sufficient to accurately represent
the dimensions of a destination’s personality. In such cases, a qualitative study can
help the researcher to identify new traits or adjectives that are relevant for
destination brands. Again, in the above mentioned case some of these techniques
may be used to eliminate irrelevant items chosen from the existing literature that
are not applicable in the case of destination brands. In some cases, the meanings
or connotations of the word used may vary from country to country; many of the
terms used in one country may not be so popularly used in other countries. In such
cases also qualitative techniques may come to the aid of the researcher to identify
or to remove the items based on the country or context under consideration.
3.2.1. Qualitative Vs. Quantitative approaches to Brand Personality Scale development
To gain a better understanding regarding the multitude of approaches to
measuring brand personality, it is required to scan through the existing approaches
in scale development, especially in the field of brand personality. Dobni and
Zinkhan (1990) say that, in psychology, human personality construct research
reached the present state over a long period of time. Hence, the brand personality
research may also have to go a long way to achieve stability. The lack of
consensus in this field related to aspects such as definition of the construct,
dimensions of the construct and so on has resulted in application of multitude of
research techniques and definitions for measuring brand personality. Researchers
have tried to adapt many of the measurement techniques employed in human
personality measurement research, in search of a better brand personality
measurement tool. However, none of them have gained universal acceptance. A
81
discussion of these attempts would be of immense help to gather insights for
designing a new tool for product brand personality measurement.
3.2.1.1. Qualitative Measurement
Qualitative approaches to brand personality have contributed significantly
to the enhancement of measurement techniques in this domain. These methods
often employ indirect ways to capture the brand personality perceptions of
consumers. Qualitative methods are often employed or helpful when direct
approaches fail or, in cases were respondents are unwilling to respond. In some
other cases, direct approaches will be less effective to capture the feelings,
attitudes and thoughts of consumers, which also may prompt the researcher to rely
on indirect methods which could be more result yielding.
One such qualitative technique widely used in marketing is the projective
technique. It is based on the hypothesis that “when people attempt to understand
an ambiguous or vague stimulus, their interpretation and response to that stimulus
reflects a projection of their needs, feelings, attitudes and experiences” (Davis,
1997). Free association test is one of the most widely used projective techniques
in brand personality research. Relying on the projective hypothesis, free
association test is conducted by confronting the respondent with a stimulus (say a
brand name, symbols, product category and so on) and encouraging him/her to
give a set of words or associations that come to his/her mind (Aaker, 1991; Batra
et al., 1996).
A research conducted by Durgee and Stuart (1987) employed projective
techniques to identify the associations with five product categories. They used
Free association techniques to identify that the words associated with symbols of
each selected brand. For instance, their study found the brand Exxon which had a
tiger symbol to be associated with ‘speed’ and ‘power’, another brand name
Floaters was found to be associated with ‘comfort’ and ‘relaxation’; while,
82
‘refreshment’ and ‘coolness’ werefound to be associated with the brand Nestle;
‘power’ and ‘dependability’ were associated with Hercules and ‘ruggedness’ and
‘toughness’ with Dodge’s Ram symbol.
Another projective technique that can be employed in brand personality
research is sentence completion test, in which a respondent may be asked to
complete sentences such as “I feel Fanta to be a _ _ _ _ _ _ _ brand”, for which
he/she may answer with words such as ‘cool’, ‘exciting’ and so on.
Another form of projective technique often employed by researchers is the
use of pictures to evoke and elicit perceptions of consumers. According to Aaker
(1991), pictures can help to make respondents express their real feelings and
attitudes in a vicarious manner and transfer them to the characters in the picture
(Aaker, 1991). Valette-Florence et al. (2011) give an example of a research which
used pictures as stimuli to study about brands. According to them in 1950, a
researcher by name Haire used the Thematic Apperception Test (TAT) to
understand the reasons behind the housewives’ early resistance towards instant
coffee brands. In the study procedure, respondents were given two lists of items to
be shopped. One list contained an instant coffee brand and the other contained a
homemade coffee brand and were then instructed as follows ‘“project yourself into
the situation as far as possible until you can more or less characterize the woman
who bought the groceries…write a brief description of her personality and
character” (Haire, 1950). The study identified adjectives such as ‘bad house
keeper’ and ‘laziness’ to be associated with instant coffee drinkers, while
adjectives like ‘industrious’ and ‘good housekeeper’ to be associated with
homemade coffee.
Repertory Grid: This is a method introduced by Kelly (1955). It is a qualitative
approach used in psychology which has been adopted to brand personality
research (see for instance Heine, 2009). This approach employs an iterative
83
process that requires respondents to compare different triad combinations of
stimuli, to express their perceptions on constructs and to relate the stimuli to the
construct poles. Irrespective of different versions, Repository Grid Method leads
to structured data that facilitate elaborate analysis and interpretation (Heine, 2009).
Content analysis: Content analysis is another widely applied qualitative method
used extensively in social science and in marketing. This technique is applied
extensively in the destination personality research domain to identify the portrayed
personality or, what the destination marketing organizations try to communicate to
potential tourists. The brand personality studies which used this technique include
Douglas and Mills (2006) and Pitt, Opoku, Hultman, Abratt and Spyropoulou
(2007).
Free Elicitation method: This is a widely employed qualitative technique used to
elicit personality traits from respondents. Generally, in brand personality research,
respondents will be introduced to the concept of brand personality and then they
will be asked to imagine a brand as if it were a person and what adjectives they
may use to describe that person. This is similar to free association test (Reilly,
1990).
Focus group discussion: This is another widely employed qualitative technique in
marketing. It is generally conducted among a small group of respondents to
understand their feelings, perceptions, attitudes, opinions and beliefs. This
technique may be employed in brand personality research to identify the adjectives
consumers generally use to describe the personality characteristics of a brand
(Edmunds, 1999).
Though indirect approaches have the advantage of helping to bypass the
inhibitions of the respondent in answering direct questions, the qualitative nature
of the data is considered to be a limitation. Since, most of these qualitative
information are not quantifiable, statistical applicability of these data are less.
84
Another issue according to Aaker (1991) is the non-probability sampling
technique which is often employed in the case of qualitative studies, which
restricts or limits the reliable predictions on the population characteristics. Also,
the interpretation of results of this qualitative study is often criticised as it lacks
objectivity, as the interpretations are made by the practitioner or researcher
himself (Aaker, 1991). Many researchers tend to adopt quantitative methods
because of the limitations of qualitative research methods. Also, quantitative
methods successfully address many of the problems of qualitative research.
3.2.1.2. Quantitative Measurements
According to Aaker (1991), if a brand can be rated based on a set of
dimensions, the objectives of the result could be more quantifiable and objective.
A generally adopted approach in quantitative brand personality research is the
Trait approach which is borrowed from the human personality research studies. A
brief discussion of Trait approach can help us to gain better understanding
regarding this concept. Trait approach is now a commonly employed approach in
the domains of both brand personality and human personality research. Brand
personality research borrowed this concept from the field of Psychology. Traits are
developed as a result of the cognitive processing of an individual. Or, traits refer to
the habitual patterns of behaviour, thought and emotion. Again traits are
considered to be relatively stable over time, changes from person to person and
can influence the behaviour of a person. They are formed as a result of cognitive
processing of an individual, or it is formed as a result of his/ her efforts to
efficiently organize large quantity of information he/ she captures from different
environmental stimuli around them to make use of them (Strausbaugh and Lynn,
1998).
The measurement of brand personality using quantitative techniques is
done with the help of structured questionnaires consisting of a set of items to
measure the performance of the brand with respect to different dimensions. These
85
set of dimensions and adjectives are generally termed as brand personality scales
or instruments to measure brand personality. These instruments are used to collect
responses with some rating scales. Though a variety of rating scales such as Likert
scale, semantic differential scale, best-worst scale and so on are available in
marketing research, most of the researchers in brand personality, like Aaker
(1997), Aaker et al. (2001), Ekinci and Hosany (2006), Hosany et al. (2006),
Ekinci et al. (2007), Murphy et al. (2007a), Murphy et al. (2007b) and soon have
used a five-point Likert scale to collect responses. Only a very few exceptions can
be seen in this regard. One such case is the Best Worst Scaling technique used by
Lee et al. (2010). However, this is not widely accepted in the brand personality
research domain. One of the reasons for the wide acceptance of Likert-type rating
scale (in brand personality research) could be attributed to the purpose of the
brand personality researches. Aaker (1997) argues that a Likert-type scale should
be opted against other rating scales such as semantic differential scale, as the
objective is to determine the extent to which a brand can be described by certain
human characteristics or traits. Whereas a rating scale like semantic differential
scale would be more appropriate when the objective is to examine brand
personality valence of a brand or in other words, when one tries to examine
whether a brand is associated with positive or negative personality characteristics
(Aaker, 1997).
A number of brand personality scales are available to quantitatively assess
the personality of a brand (See for example Aaker, 1997; Aaker et al. 2001;
Geuens et al., 2009). The different scales available are developed with varying
purposes. While some are developed to measure personality of all types of brands
(see for example Aaker, 1997), others focus on specific categories like
destinations (see Ekinci et al., 2007), restaurants (Siguaw et al., 2003) and so on.
Again, the scope of these scales is restricted to specific countries; for instance,
Aaker (1997) scale is suitable in the U.S context; the scale of Sung and Tinkham
86
(2005) is suitable in a Korean context. Also, all these scales differ in the methods
and rigor in the development process. A discussion on this aspect is done in the
following sections.
3.3. A review of methodology of building brand personality scales
A set of articles are selected and reviewed to identify the various
methodologies adopted by researchers in developing brand personality scales.
Popular academic search engines like Google scholar and Scirus were employed to
identify the articles. Also various databases like EBSCO Host, Proquest, Science
direct and so on were also explored to find the most relevant and important
articles. Since the objective of this part was to evaluate the methodological
practices in scale development, only those studies which focused on scale
development were considered. The selected papers were then subjected to an in-
depth and careful review to determine the methodology adopted and the flaws, if
any existed. Finally, considering the inputs from reviews, a viable scale
development strategy was designed to develop a product brand personality scale in
the Indian context. The researcher has mainly looked into the following aspects
while considering the literature. This includes: research methods, sampling
techniques, brands considered, mode of survey, item generation, item purification,
assessment of items, dimensionality, validity and reliability.
3.3.1. Research approaches
One can see micro and macro approaches to brand personality
measurement. Again, Research methods adopted in brand personality research can
be categorized in to quantitative, qualitative, and mixed approaches.
3.3.1. 1. Micro versus Macro approaches to brand personality
Brand personality researches, in general, can be categorized into two: micro
and macro. One of the concerns in brand personality research could be whether to
go for an overly ‘global scale’ (a more holistic one) or to adopt an overly reduced
87
one, developed considering a single brand, a very few brands, or small product or
service category. When considered in this manner, it can be seen that the scope of
brand personality scales ranges from a very narrow, micro versions to very broad,
macro versions. One can also see the domain of brand personality to range from
general (Aaker, 1997; Aaker et al., 2001; Capraraa et al., 2001; Sung and
Tinkham, 2005; Bosnjak et al., 2007; Geuens et al., 2009 and so on) to specific
categories (Opoku et al., 2007; Heere, 2010; Maciel, da Rocha and da Silva, 2013
and so on). It is quite difficult to say which one is better. While a ‘micro’
perspective could give more focus and accuracy in terms of results, it hinders the
scope to compare between products over categories based on their personality. A
macro or holistic perspective on scales addresses the above mentioned problem at
the cost of accuracy, to a certain extent. Also, macro approach has the advantage
of cost and time effectiveness as against micro scales. Hence, there exists a
conflict of interest between the researchers on whether to adopt a micro or macro
perspective on brand personality scale development and researches. While few
researchers like Valette-Florence and Barnier (2011) adopted a micro approach of
scale development, Aaker (1997) and Geuens, Weijters and Wulf (2009) adopted a
macro perspective of scale development.
What is chosen for the study: Deviating from the two extremes of micro or macro
perspectives in brand personality research, this research tries to adopt a middle
path, a meso perspective (scope of the scale with respect to brands) that is neither
general nor specific. This research will be focussed on product brands alone so
that the scale will be more useful than a general brand personality scale and will
have a wider application than a category-specific or brand-specific scale.
Considering the fact that less often one need to compare a product brand with a
service brand (a manager needs to compare his brand with its competing brand,
which will also be a product brand) , the effect of avoidance of service brands will
have a lesser impact on the applicability of this scale. Also, creating category-
88
specific scales in the case of product brands will be a challenging task as many of
these brands will be having presence in multiple product categories and when one
goes for product category specific scales, it is difficult to capture the consumer
perception of the brand solely on the particular product category they are trying to
measure.
One of the rationale for researchers to create separate brand personality
scales in the field of service brands is that, they derive personality through many
other sources. For instance, d’Astous and Lévesque (2003) showed retail brands to
have a unique personality which they derive through their sales personnel. This
may be the case with other service brands like banks, telecommunication services,
destinations and so on. In all these cases, the sources of brand personality vary
significantly. For instance, take the case of destinations, a tourist’s personality
perceptions are not restricted to advertisements or brand ambassidors of the
destination. It is highly influenced by the service providers at the destination, local
people, the typical visitors to the destination and so on. However, in the case of
products, these are restricted to a great extent. For instance, most of the product
brands will be sharing a common retail environment, common service personnel
and so on.
3.3.1. 2. Emic and Etic approaches in research
According to Usunier (1998) emic approach is based on the idea that
behaviours and attitudes of people can be understood exclusively within a
particular cultural context. Pike (1967) who introduced the etic approach proposed
the existence of universal principles that stand independently in different cultural
contexts. Valette-Florence et al., (2011) argues that in the case of brand
personality some dimensions were more associated with American culture. The
works of Aaker et al. (2001); Supphellen and Gronhaug (2003) and Sung and
Tinkham (2005) suggests the same. All these suggest to the emic and etic
89
approaches and serves as a further justification for this research in the Indian
context, considering product brands.
3.3.2. Research methods adopted in Brand Personality scale development
One can observe the use of qualitative, quantitative or mixed method
approaches in developing the brand personality scale. These approaches were used
for the measurement of either brand specific or general brand personality
measurement tool generation. However, a large number of researchers in this field
has started with a qualitative approach initially (mainly for item generation and
initial item reduction) and then moved on to quantitative methods for
identification of brand personality dimensions and validation of the scale.
It is good to employ a qualitative technique at the initial stage. For instance,
one important qualitative technique that would be appropriate is the interview
technique. It can be done for a single individual or for a focus group. This can help
the researcher understand how consumers perceive symbolic aspects of the brand.
Another alternative at this phase could be the free elicitation study, which was
discussed before in this chapter. The advantage of free elicitation test over
interview technique is that more respondents can be contacted at a shorter time,
Review of literature indicates that this technique was widely used by researchers
in the item generation phase. Consumers were generally asked to imagine brand as
if it were a person and were then asked to describe it with as many adjectives as
possible.
The two generally employed methods for item reduction are expert opinion
survey and consumer survey. While the item generation phase tries to generate as
many items as possible, the item reduction phase tries to eliminate the less relevant
items from the list. The items are removed either with the help of experienced
consumers or with judges (generally marketing practitioners or researchers). This
90
is more of qualitative nature and the general purpose of this phase is to reduce the
number of items in the item pool to a manageable level. However, to identify
brand personality dimensions, researchers generally employ exploratory factor
analysis. Confirmatory factor analysis techniques are employed to confirm the
brand personality dimensions identified. Many researchers measure the reliability
and validity of the scale with the help of different statistical techniques such as
Cronbach’s alpha, test re-test reliability, and so on with an intention to establish
the psychometric properties of the scale.
3.3.2.1. Sampling choices
Aaker (1997) and d’Astous and Lévesque (2003) relied on households for
collecting data. While some others like Aaker et al. (2001) and Caprara et al.
(2001) conducted survey among consumers. Some other prominent methods
employed by researchers include mall intercept survey used by Helgeson and
Supphellen (2004); online method adopted by Geuens et al. (2009) and student
sample method used by Sung and Tinkham (2005).
Most of the researchers in brand personality research have used
convenience sampling; for instance Aaker (1997), Aaker et al. (2001), Caprara et
al. (2001), d’Astous and Lévesque (2003), Sung and Tinkham (2005), Geuens et
al. (2009) are some of them. Very few researchers like Helgeson and Supphellen
(2004) has used random sampling technique. Aaker (1997) for instance, relied on a
national consumer panel and collected responses through Federal Express mail
service.
3.3.2.2. Brands Considered
Selection of stimuli or brands based on which respondents evaluate the
candiate traits forms an important part of any brand personality research. Aaker
(1997) for instance while selecting the brands for her research observed the
91
following rules such as consideration of salient and well known brands, as the
study was to be conducted among a set of national sample. Again, with the
objective of enhancing the generalizability of the scale, brands from a wide variety
of product categoeries were considered. Further, additional care was given to
ensure that selected brands consisted of three types of product categories namely,
symbolic, utilitarian and symbolic and utilitarian. Considering these aspects her
research work considered 37 brand s at the dimension identification phase and 20
brands at the dimension confirmation phase.
25 global brands that had presence in the two countries under consideration,
namely Spain and Japan was selected by Aaker et al. (2001) to unveil the brand
personality dimensions in these countries. They also considered some more
aspects such as familiarity of the brands among the consumers and
representativeness of different product categories. Caprara et al. (2001) restricted
their study to 12 mass market brands.
d’Astous and Lévesque’s (2003) relied on four departmental store brands to
generate their store personality scale. They considered two types of stors;
hardware and automobile, in order to to generate their scale.
Helgeson and Supphellen (2004) opted a convenience strategy to select five
clothing brands from Sweden to develop their scale. A total of 13 global brands
that had presence in both Korea and U.S (which were the study areas) were
selected by Sung and Tinkham in their efforts to develop their scale to measure
brand personality dimensions in these countries.
Geuens et al. (2009) adopted a different strategy from other researchers in
selecting the brands for their study. The researchers wanted to ensure that the
selected list of brands represented different purchase motivations of customers.
The different purchase motivations they taken in to account while selecting the
brands includes emotional or symbolic, functional, and experiential. Considering
92
the above mentioned criteria, the reseachers seleced 20 brands for their study to
identify the brand personality dimensions and about 193 brands were considered
to check generalizability.
From the reviews in the above section it is observable that researchers tend
to rely on both small and large sample sizes to develop their scales. However, both
of these approaches have its own advantages as well as limitations. According to
Aaker (1997), one can enhance the generalizability and robustness of a scale by
increasing the number of brands considered in the scale development process, but
at a cost of respondent fatigue and boredom. Where as too short number of brands
(say four as in the case of Ambroise et al., 2005) can reduce respondent fatigue
and boredom at the cost of scale reliability and robustness. Hence, it is advisable
to rely on moderate number of brands (say 20 as in the case of Geuens et al.,
2009) to achieve generalizability and robustness to a great extent at a lowered
respondent fatigue and boredom.
3.3.2.3. Survey Administration
Brand personlaity researchers have used both online and offline modes of
data collection in their research. Offline modes were generallly used ot conduct
qualitative study (see for instance, Aaker, 1997; Helgeson and Supphellen, 2004;
and Geuens et al. 2009). In the case of empirical studies both online and offline
methods were used. For instance, Aaker (1997) used Federal express to deliver
questioonnaire to households that were part of a consumer panel.
d’Astous and Lévesque (2003) took help from a set of intrviwers to
administer their questionnaire. The interviewers visited each house hold and
explained the questionnaire and left the respondent to fill it by themself and
collected the questionnaire at a later point of time as per thrir convenience.
Sung and Tinkham (2005) whose respondents for the research were
students, collected their responses directly in classrooms. A mall intercept
93
technique was used by Helgeson and Supphellen (2004) to collect responses from
their target female shoppers. Geuens et al. (2009) sought the help of an online
survey panel to collect responses from the required type of respondents for their
study.
3.3.3. Generation of Items
A systematic generation and selection of items is required to accurately
represent or create the brand personality scale. The items were generated using
both inductive methods such as literature reviews and deductive methods such as
exploratory research.Many studies on brand personality in the initial stages were
mostly a transposition of theories in human personality to the domain of brands
(Wee, 2004). Adoption of these ‘crude’ scales to measure brand personality can
hinder the validity of the scale developed. Aaker (1997) has clearly put forward
this issue in the measurement of brand personality construct. She emphasized that
not all items in human personality scales are relevant or applicable in the case or
brands. And hence she insisted on selecting those items; that are relevant and
applicable’ in the case of brands. Though, Aaker (1997) relied to a great extent on
traits or adjectives from human personality scales, she ensured that only items that
are relevant for brands were included in the scale by means of a consumer centric
study, which evaluated and purified scales by eliminating items irrelevant in the
case of brands. However, while considering the items, she purposely removed
items with negative connotations, arguing that the purpose of any brand
personality scale is to measure how effective advertisers are in communicating
particular human personality traits to their audience and they generally try to
project a positive image only. But, researchers like Bosnjak et al. (2007) disagree
with this, arguing that consumers need not always perceive a brand to be positive,
they may attribute negative traits as well. Based on this argument, Bosnjak et
al.(2007)in their scale included negative traits and even identified a negative
dimension (superficiality) and facet (boredom).
94
For the proposed scale, items from different types of brand personality
scales were decided to be included. However, items from human personality scales
were not considered because many of the brand personality scales considered
herein, were developed relying on human personality scales and the items they
removed were mainly those that were considered less applicable in the context of
brands. Hence, considering items from human personality scales may not provide
any significant contribution in terms of improving the quality of the new tool
against the effort required to filter the inappropriate ones (which may be added
from human personality scales) out. However, no restriction should be put based
on whether items come from micro or macro scales. Items from scales based on
both these approaches can be included in the study. Of course, many items in the
case of some brand personality scales, especially those based on service industry
brands like destination personality scales or restaurant brand personality scales
might be less useful in the product brand personality context. However, it was
decided to include them to avoid any kind of personal bias and to ensure
maximum inclusion of candidate traits. For example, items like ‘Compassionate’
which is a trait in the case of charitable organisations (Venable et al., 2005), need
not be significantly appropriate for product brands or in other words could be less
relevant than other items. But, these were decided to be included and left with the
experts and consumers to decide on the extent of suitability.
Another, aspect to be considered is whether to include negative items in the
scale. Since many previous researchers like (Bosnjak et al., 2007) included
negative items and their argument that a brand may be positioned with a negative
image, it is decided to include negative items as well, and let the purification stage
decide whether to retain or eliminate these items.
In addition to items from literature, many researchers such as Aaker (1997),
Aaker et al., (2001), Geuens et al., (2009) and so on, have developed new and
original items for their study, with an objective of improving the content validity.
95
It is advisable for any future researcher who attempts to develop a new scale to
rely on some qualitative technique discussed earlier in this chapter, to generate
new context-specific items. This will help in improving the content validity of the
scale developed. For the proposed product Brand Personality scale development,
new and context-specific items may be generated by conducting a free elicitation
study.
3.3.3.1. Sources of items
Researchers have relied on different sources to identify the candidate items
to develop the scale. For instance, Aaker (1997) relied on sources suh as items
from scales used by marketing practitioners, human personality scales, and an
original qualitative study. Aaker et al.(2001) adopted item from personality scales
in psychology, items in scales used by practitioners, and a set of newly genrated
items based on an interview conduted among 50 respondents. While Caprara et al.
(2001) relied solely on literature survey for item generation, d’Astous and
Lévesque (2003) used items from literature survey and individual interview,
Helgeson and Supphellen (2004) conducted a survey among shoppers to identify
the relevant items for developing the measurement tool. Sung and Tinkham (2005)
used items from literature survey and personal interview; Geuens et al. (2009)
used items from literature survey and focus group interview.
3.3.3.2. Initial number of items
The initial number of items ranges from as low as eight (Helgeson and
Supphellen, 2004) to as high as 309 (Aaker, 1997). Some other researches wherein
the initial number of items varies in between these values includes 253 and 266
items respectively in the cases of Japan and Spain by Aaker et al., (2001); 40
items used by Caprara et al.(2001) to develop a general brand personality scale; 99
items by d’Astous and Lévesque (2003) for developing a scale to measure store
96
personality; 91 items by Sung and Tinkham (2005); 244 items by Geuens et al.
(2009).
3.3.4. Initial item reduction
Many researchers in the realm of brand personality research tend to adopt
an initial item reduction phase which is intended to reduce the large number of
initial items generated to a manageable level. They either rely on an expert opinion
or a small consumer survey for this. For example, Aaker (1997) used responses
from 25 consumers to reduce the initial item pool generated in their study to a
manageable number of 114 items; Aaker et al. (2001) conducted consumer survey
among consumers respectively to reduce the initial item pool generated to create
brand personality measures for Japan and Spain. The reduced set consisted of 140
items in the case of Japan and 75 items in the case of Spain. Geuens et al. (2009)
conducted two stages of expert opinion surveys. The first stage which employed
responses from eight experts helped them to reduce the item pool to 108. This was
further reduced to 40 with the help of another expert opinion survey consisting of
20 researchers. However, this step was not done in the case of some researchers
like Caprara et al.(2001), Sung and Tinkham (2005) and so on. One reason for
this could be relatively small number of items in the initial item pool generated (
40 in the case of Caprara et al.(2001) and 80 in the case of Sung and Tinkham
(2005)). Though a small number of items in the initial item pool can help the
researcher save a step in the process of scale development, it may affect the
content validity of the scale .
It is observed from the literature that researchers have employed either
expert opinion and consumer survey to reduce the initial pool of items.Further, it
is identified that very small number of judges may be used to eliminate items at
the initial stage.
97
3.3.5. Assessment o f dimensionality
Researchers have depended mainly on exploratory factor analysis (EFA) to
identify the number of dimensions and the items corresponding to each dimension
(see for instance, Aaker, 1997; Aaker et al., 2001; Sung and Tinkham, 2005 and
Geuens et al., 2009). However, depending on exploratory factor analysis alone for
checking dimensionality is not advisable. As factor analysis possesses several
limitations such as the factor structure obtained based on EFA being only one
among the infinite number of potential solutions (Segars and Grovers, 1993).
Again, Segars and Grovers (1993) argues that in the case of data showing
correlation among factors, a varimax rotation may produce distorted factor
loadings and incorrect conclusions on factor solutions. Again, in many cases items
may load on multiple factors which hinder the interpretability of the factors to a
great extent (Sureshchandar, Rajendran and Anantharaman, 2002). Because of
these limitations, it may not be advisable to rely solely on exploratory factor
analysis to assess dimensionality of a scale.
Many of the limitations of an exploratory factor analysis model can be
addressed with the help of a confirmatory factor analysis (CFA). For instance,
CFA enables the researcher to compare different model specifications and evaluate
the invariance (that is, whether values are varying significantly or not) of specific
parameters in the factor solution (Marsh and Hocevar, 1985). Also, the goodness
of fitness index values provided by CFA analysis results helps the researcher to
examine the fit of the factor structure with data. Many researches in brand
personality have employed CFA along with EFA to confirm the brand personality
dimensions (Aaker, 1997; Aaker et al., 2001; Sung and Tinkham, 2005; and
Geuens et al., 2009).
Two very important aspects to be considered while conducting a study for
evaluating a scale’s dimensionality are sample size considered and number and
type of brands chosen.
98
Sample size for EFA: Aaker (1997) used 631 respondents to identify the five
dimesion stucture of U.S brand personality. Aaker et al. (2001) used
1495respondents in the case of Japan and 692 respondents in the case of Spain to
identify brand personality dimensions. Sung and Tinkham (2005) collected
responses from 320 in the case of U.S and 337 in the case of Korea. Geuens et al.
(2009) used 1235 respondents to identify the personality dimensions in the case of
Belgium. Helgeson and Supphellen (2004), limited their sample size to 424. It was
d’Astous and Lévesque (2003) who used smallest number of samples at this phase.
They considered responses from only 26 participants to identify the brand
personality dimensions.
Sample size for confirming brand personality: Researchers have used relatively
small sample size for confirming brand personality dimensions. Aaker et al.
(2001) used 180 respondents; d’Astous and Lévesque (2003) used 220
respondents. But many researchers like Caprara et al. (2001), Helgeson and
Supphellen (2004), Sung and Tinkham (2005) did not confirmed the brand
personality dimensions they identified.
Number and type of brands for EFA study: The next important aspect is the
number of brands to be considered. Aaker (1997) used 37 brands; Aaker et al.
(2001) employed 25 brands in the case of both Japan and Spain. Number of brands
used by some other researchers at this phase were as follows Caprara et al. (2001)
used 12 brands, Helgeson and Supphellen (2004) used five brands, d’Astous and
Lévesque (2003) used four brands, Sung and Tinkham (2005) used 13 brands,
Geuens et al. (2009) used 20 brands and so on. The type of brand to be selected
depends on the scope of the study Aaker (1997) for instance tried to develop a
general brand personality scale hence she tried to make her brand profile to be
more general while d’Astous and Lévesque (2003) relied on store brands, as their
objective was to create a store personality scale.
99
Number and type of brands used for CFA study: Aaker (1997) used 20 brands to
confirm the dimensionality of the brand personality scale. d’Astous and Lévesque
(2003) used just two brands at this phase. As mentioned before type of brand
opted depended on the scope of the scale.
3.3.6. Reliability and Validity of the Scale
Reliability of a scale is another important aspect to be considered while
developing a measurement tool. Nunnally (2010) suggests Cronbach’s alpha
values to be an efficient indicator of reliability. A value greater than 0.7, is
considered to indicate a good reliability (Nunnally, 2010). In the case of various
factors identified by Aaker (1997) the reliability values ranged from 0.90 for
ruggedness to 0.95 for excitement or in other words one can observe very good
reliability in the case of Aaker’s (1997) brand personality dimensions. For the
scale of d’Astous and Lévesque’s (2003) reliability values of the factors ranged
between 0.75 and 0.91. Where as in the study done by Sung and Tinkham (2005),
the reliability values for U.S brand personality dimensions ranged between 0.80
and 0.95 and for Korean brand personality dimensions, reliability scores were
between 0.72 and 0.95. All the factors identified by Caprara et al. (2001) indicated
very good reliability scores. All the factors were found to have scores higher than
0.96. In the research work of Geuens et al. (2009), the Cronbach’s alpha values
were found to be higher than the recommended value of 0.7, for four dimensions.
However one dimension namely openness was found to possess a low reliability
score of 0.61 which is less than the recommended value of 0.7. The
recommendations by experts and practices in this field indicate that a Cronbach’s
alpha score of 0.7 or higher will be appropriate in the case of reliability of
dimensions of a brand personality scale.
Another indicator for reliability is split half technique. However, no
researches in brand personality were found to use this technique. But, it is
100
justifiable as Cronbach’s alpha value is considered to be a better indicator of
reliability than split half technique. Both, Cronbach’s alpha value and split half
techniques try to measure the internal reliability of a scale.
Another important aspect to be considered in the case of reliability is the
external reliability of the scale. However, most researchers in the brand
personality domain tend to neglect this aspect. Aaker (1997) conducted a test re-
test study and used the correlation values to assess the external reliability of the
scale. The average test re-test correlation value was found to be .80. The
recommended value for test re-test correlation is 0.70.
Convergent Validity: Convergent validity refers to the extent to which a set of
items intended to represent a construct actually converge on the same construct
(Ladhari, 2010). Many brand personality researchers have overlooked the
convergent validity of their scale. A general procedure to check convergent
validity as recommended by Fornell and Larcker (1981) is that the average
variance should be greater than .50. For the scale developed by Sung and
Tinkham (2005), the convergent validity was found to range from 0.75 to 0.94 in
the case of U.S and it ranged between 0.71 and 0.97 in the case of Korea.
Discriminant Validity: Ladhari (2010), describes discriminant validity to be the
extent to which theoretically unrelated constructs practically exhibit less
correlation. Researchers like Sung and Tinkham (2005), Geuens et al. (2009) and
few others verified the discriminant validity of their scale by comparing the
average variance extracted for each factor with the squared correlation between
that construct and other constructs in the model. The average variance extracted
value should be greater than the squared correlation value to suggest discriminate
validity (Gounaris and Dimitriadis, 2003).
101
3.4. Proposed scale development procedure
This is a research done with the objective of developing a tool to measure
product brand personality. Based on the methodology review, the researcher has
adopted a mixed method approach to achieve this objective. This study proposes
to employ both qualitative and quantitative methods. A qualitative approach is
proposed for the initial stages of item generation and initial item reduction. The
empirical or quantitative part of this study will be concerned with identification of
brand personality dimensions, confirming the dimensionality and ensuring the
validity and reliability of the scale. Based on the literature a scale development
model is developed for the proposed product brand personality scale.
The following sections details on the sources of data, brand selection
procedure methodologies and techniques to be adopted in the quantitative part of
the scale development process and so on.
3.4.1. Sources of data
The researcher should rely on primary data for creating the brand
personality scale. Secondary sources can be used as background material. The
variables identified based on literature survey constitutes the background material
that should be used for the study. Another secondary source of information is
related to the selection of brands. To rely on a definite sample frame for the
selection of brands, a list of top brands is required. Since the survey needs to be
conducted with a Pan Indian representation, the brands selected should be familiar
at the national level. Hence, a list of top 200 brands should be created. This needs
to be done relying on secondary data.
The two prominent sources of list of top brands in the Indian context are 1)
brand trust report, which is an annual report published by Trust Research
Advisory, which contains a total of 1000 top brands trusted by Indian consumers.
102
This list was developed based on a survey conducted among 2718 carefully
selected respondents from 15 cities, which generated around 2 million data-points
from 12000 hours of research. This list consists of all types of brands including
global giants like Nokia, Sony, LG, Samsung; Indian multinationals like Tata,
Bajaj; public sector brands like ONGC; celebrity brands like Anna Hazare, Sachin
Tendulkar, Salman Khan and so on. Also, this list covers a wide range of product
and service categories such as consumer products, electronic goods, cosmetics,
fragrances, automobiles, service sector, public sector, and celebrities and so on,
which makes the list a comprehensive one (Trust Advisory, 2012).
Another important and trustworthy list of top brands is the one published by
the Brand Equity survey (from The Economic Times Newspaper). The list is
published every year based on survey conducted in different cities in India. This
project is done with the help of a small team of researchers. They visit households
to collect responses directly from the respondents. The Brand Equity team claim
that the brands they select to be the ones which ‘Indian consumers love and trust
the most’. This list contains top 100 brands including both product and service
brands (ET Bureau, 2011)
.
Though both lists seems to have been developed based on good research
work, it is advisable to develop a new list cosnidering both lists, which will
increase the validity of the brand selection phase. A list of top 200 brands needs to
be considered, so that the list is more general and will cover a wide range of
product categories. The inclusion criteria for the selection of brands is that they
must be product brands, they must belong to either of the two listings. The top
ones in the listing should be considered first. And the exclusion criteria consists of
omission of celebrity brands, service sector brands and so on.
103
The list prepared needs to be used in the two different phases of the
research. The first one is at the time of consumer survey done to identify the
dimensionality of the scale. The second time will be the hold out sample phase
where in the dimensionality of the scale needs to be verified using another set of
brands. However, at the second time, the brands selected at the first phase needs to
be omitted, so that any possible influence of those brands in the result can be
avoided.
3.4.2. Brand selection
It is advisable to select the brands randomly. However, inclusion of brands
from different categories can enhance the generalizability of the scale. Hence, it is
advisable to randomly select brands from different product categories at the
dimension identification stage. Random sampling will be more appropriate at the
hold out sample study phase also. This will help to reduce the bias in brand
selection.
Number of brands: Considering the number of brands opted by previous
researchers and considering the objectives of this study, a total of 21 brands seems
to be appropriate at the dimension identification stage and 12 brands would be
sufficient at the stage of confirming the brand personality dimensions.
Primary data sources proposed for the study includes two groups of
respondents, experts and students. Responses from experts shall be collected for
item purification and item reduction stages. While student responses will form the
primary source of data for stages such as item generation, brand personality
dimension identification and for confirming the brand personality dimesnions.
The following sections describes the methodology proposed at different
phases for carrying out this research.
104
3.4.3. Methodology design for Item Generation
Two sources of items are suggested for the study. The first is items from
existing brand personality scales, and the second is using a qualitative study.
Inclusion criteria for items from literature survey should be that, items should be
from a scale developed for measuring brand personality. However, the items from
human personality scales need not be considered as the important and those
relevant in the case of brands will be present in the existing scales. However, it is
less likely that, all the items that are relevant in the Indian context will be covered
by item pool generated in this manner. Hence, a qualitative study to uncover the
specific traits used by Indian consumers to describe the brands needs to be
undertaken.
3.4.3.1. Procedure for item generation from literature survey: All the important
brand personality scales should be identified and the traits used by researchers
should be pooled together to create the initial item pool. A spreadsheet such as
MS Office Excel 2013 may be used to identify the redundant items in the list as
well as to arrange the items in the alphabetical order.
3.4.3.2. Procedure for item generation from the qualitative study
A free elicitation method is proposed to generate new items. The purpose of
this qualitative study is to identify the major personality attributes that consumers
ascribe to different brands.
Research Instrument: A semi-structured questionnaire may be appropriate to
conduct the study. The questionnaire should detail the concept of brand
personality to the respondents with the help of relevant examples and should ask
them to try to think of their favorite brand as if it were a person and then elicit the
human personality traits they ascribe to brands. Respondents may be given the
option to select the brand of their choice from three categories of products such as
symbolic, utilitarian, and those which are both symbolic and utilitarian.
105
3.4.4. Methodology design for Language expert opinion survey
Language expert opinion survey is not a generally used procedure in brand
personality scale development. However, it seems to be rational to include this
step to the scale development process. This is done as a part of the efforts to
eliminate the irrelevant items that may be generated during literature survey phase.
A language expert opinion survey is intended to deliver a refined pool of items by
eliminating or replacing items that are less used in India. A total of three language
experts may serve the purpose.
The modified pool of items from the language expert opinion phase needs
to be combined together with those generated through qualitative study to form a
master pool. Any redundant items may be removed at this phase. MS Office Excel
software may be used to achieve this.
3.4.5. Methodology design for Expert opinion survey
From the literature it is evident that most of the researchers have relied
either on an expert opinion survey or a small consumer survey to reduce the large
number of items to a manageable level. For this research, an expert opinion survey
is suggested, considering the fact that the item pool will be still crude despite the
refinement at the language expert opinion phase. As the language experts need to
look at the semantic aspects of the items in the Indian context, the filtering process
will not be complete.That is, they may not look in to aspects such as whether an
item is suitable considering the definition of brand personality. The purpose of this
phase is to filter the items more rigourously. The experts are suggested as
respondents at this phase as their expertise in the field will be helpful in refining
the items based on the brand personality definition considered for the study and
the appropriateness of the items to describe the personality of a brand.
A total of 12 experts are suggested for this phase. An equal number of
judges from different fields related to branding such as advertising practitioners,
106
brand managers, marketing academicians, and consumer psychologists can help to
give a balanced representation of experts. Each respondent may be asked to give
their responses based on three broad categories of produts as mentioned before in
the case of qualitative study ie, symbolic, utilitarian, and symbolic and utilitarian.
To identify any missing traits that did not turned up at the qualitative study phase,
free association test may also be conducted for these experts and any new and
relevant items identified may be incorporated to the item pool.
3.4.6. Methodology design for identifying the brand personality dimensions
A consumer survey needs to be conducted to uncover the brand personality
dimensions in the Indian context. This survey is expected to identify the
appropriate number of dimensions and the corresponding items to each dimension
that will represent product brand personality in the Indian context.
Research instrument: A structured questionnaire designed from the results of
Expert opinion survey can be used for the consumer survey. The questionnaire as
discussed before, should introduce the brand personality concept with illustrations
and should contain detailed instructions on how to select brands and rate each
items based on the selected brands. A 5-point Likert scale may be used with 5
indicating the item to be very much appropriate to describe the brand and 1
indicating vary inappropriate to describe the brand. A 5-point scale needs to be
considered against a 3 point or 7 point scales as the former will give inadequate
options to respondents while a the latter requiress high level of sensitivity and
discrimination that mght be difficult considering the concept under consideration
and the expertise of the respondents.
Administration of the questionnaire: A university level (mainly post graduate and
doctoral students) student sample may be opted for this study as they have better
proficiency in English language in which the questionnaire will be designed.
Additionally, student samples have been shown to be acceptable for studies
107
involving theory testing due to the homogeneity of the sample (Calder, Phillips,
and Tybout, 1981; Carlson, Donavan and Cumiskey, 2009). Selected respondents
should be representative of the country in terms of gender, family income, and
needs to be from across the country.
Sampling plan: A cluster sampling may be adopted for sample selection. The
sample size may be restricted to 600, considering the sample sizes used in the case
of previous researches in this field. The respondent’s familiarity with the brand
that he/she rates was considered as an inclusion criteria for the study. Familiarity
may be assertained by instructing the respondents to choose brands that they have
used from a list of popular brands.
Data analysis and statistical tools: Exploratory factor analysis may be employed
to identify the factor structure of the brand personality construct. Principal
component analysis with varimax rotation is expected to yield the results. SPSS
(Version 16) software package may be employed for this purpose. Further, the
reliabilty of the scale should be assessed by considering the Cronbach’s alpha
values and test re-test reliability values. Cronbach’s alpha values for each factor
can be calculated relying on the data already collected for this phase. However, to
evaluate test re-test reliability one needs to collect another set of data from a small
set of respondents who participated in the consumer survey phase. The sample size
for this survey can be restricted to 90. The test re-test values for each variable and
each factor needs to be evaluated by correlating the values obtained for the
variables in the two phases. A value of 0.7 or higher is considered to show good
test re-test reliability.
3.4.7. Methodology design for confirming the brand personality dimensions
A hold out sample survey needs to be conducted to confirm the brand
personality dimensions and the items selected.A structured questionnaire based on
the brand personality dimensions and items identified using exploratory factor
108
analysis in the previous phase can be used at this phase. A similar procedure
adopted in the case of brand personality dimension identification phase needs to be
used at this phase also. To elaborate, the survey may rely on a 5-point Likert type
scale to collect responses.The sample size may be restricted to 200 (or depending
on the number of items, minor variations may be made to satisfy the sample size
requirements to employ confirmatory factor analysis method). For this phase also
university-level students may be adopted as respondents.
Data Analysis techniques: A confirmatory factor analysis needs to be performed
with LISREL (Version 8.72) software package to confirm the brand personality
dimensions. Based on the LISREL ouputs, the researcher should check the
construct validity of the scale by considering convergent and discriminant validity
measures.
3.5. Developing a short version of the scale
Many of the brand personality researchers have developed two versions of their
scales. The first is the full version and the second one is called a short version of
the scale(see for instance, Aaker, 1997; Burisch, 1997; Rammstedt and John,
2007; Geuens et al., 2009). In this case also, if the first version contains relatively
large number of items say greater than 40, a short version of the scale may be
developed. For this, data collected during the previous phases may be used.
However, it is required to verify the dimensionality, different types of reliability
(such as Cronbach’s alpha values, test re-test reliability values), validity and so on
separately for the short version of the scale.
109
3.6. Research framework
Construct definiton: Definition based on Azoulay and Kapferer (2003)
Content Validity: • Item Generation:
• Using literature survey • Qualitative study - Free elicitation method, 33 participants,
• Un-restricted stimuli selection (from product brands)
Measure Purification • Language Expert Opinion Survey: 3 language experts
• Expert Opinion Survey: 12 experts; 3 each from among marketing academicians, advertising practitioners, consumer psychologists and brand managers
• Consumer Survey; Sampling: quota method, respondents: 606, Brands: 21 randomly selected
Dimensionality and Item Selection: • Exploratoryfactor analysis using SPSS software
Reliability Assessment • Cronbach’s Alpha
• Test-retest method: 91 participants of consumer survey.
Analysis using Hold out sample Sample:12 brands,217 student respondents
Confirming brand personality dimensions: Confirmatory Factor Analysis using LISREL
Validity Assessment Content validity: based on construct definition, item generation, item selection
• Discriminant Validity: Using LISREL • Convergent Validity : Using LISREL
Developing a short version of the scale Assessment of the reliability and validity of the short scale
Figure- 3.1. Framework for scale development
[Source: Original]
110
3.7. Chapter conclusion
This chapter has reviewed different approaches to measuring brand
personality, various scale development procedures adopted by brand personality
researchers, different aspects considered by researchers in developing the brand
personality scale and so on. Based on these, a methodology for developing a scale
to measure product brand personality in the Indian context has been designed. The
methodology proposed for this study focuses on a mixed method approach
consisting of multiple phases that facilitate systematic generation and filtering of
items to develop the required scale. The qualitative part of this approach is mainly
intended for initial item generation while the quantitative part is intended at
identifying brand personality dimensions and verifying the reliability and validity
of the scale. Various aspects such as selection of brands, sampling procedures and
so on, have also been discussed in detail in this chapter.
111
Chapter 4
SCALE DEVELOPMENT AND EVALUATION
This chapter broadly deals with the scale development process employed in
this research and describes the construct definition, item generation, measure
purification, reliability and validity assessment, development of a shorter version
of the tool and so on. The major objective of this chapter is to detail the major
steps involved in formulating the product brand personality framework in the
Indian context.
4.1. Introduction
A rigorous scale development procedure as described in Chapter 3 was
adopted for this study. It is a seven phase process, starting with construct
definition.
4.2. Defining the Construct
As discussed in Chapter 2, brand personality has been defined in a number
of ways by different researchers. However, a definition put forward by Azoulay
and Kapferer (2003) is opted for this study as it is considered to be a better one
than the widely used Aaker’s (1997) definition of brand personality, by many
researchers like Bosnjak et al.(2007), Geuens et al. (2009) and so on. The main
advantage of this definition is that, it is consistent with the human personality
definitions in psychology. Azoulay and Kapferer (2003) define brand personality
as the ‘the set of human personality traits that are both applicable to and relevant
for brands’.
4.3. Content Validity
Content validity refers to item sample adequacy. That is, the extent to
which a specific set of items reflects a content domain (DeVeltis, 2011). Aaker
112
(1995) entails two steps to ensure content validity, of which the first is concerned
with item selection and the second with stimuli selection. Firstly, a broad and
representative set of items are to be pooled, which can be used to measure brand
personality. Next, a set of stimuli (brands) are to be selected against which the
appropriateness of the items shall be examined using an empirical study. Numally
(2010) argue that if the construct definition, item selection andstimuli selection are
done in a well defined manner, then the content validity of the scale can be
established before its construction.
4.4. Stimuli Selection
To ensure content validity, it is necessary to choose a broad and
representative set of stimuli (brand). The selection of brands for the study was
guided by the following principle. Brands that were salient and well known at the
national level were considered for the study, since the study was to be conducted
among a pan-Indian sample.This ensured that the selected brands are more
relevant in the national context and the respondents will be familiar with each
brand. Stimuli (brands) were selected randomly to get a better representation as
well as to avoid bias in the selection process. The selected brands belong to a list
of 200 top product brands developed relying on two popular listings of top brands:
one from an agency named Trust Research Advisory, a leading research agency
and the second was Brand Equity listing (from the Economic Times, a reputed
business newspaper from the Times of India Group). Restricting the study to
product brands was based on the notion that the personality of service brands may
differ significantly from that of product brands. For instance, unlike product
brands, the main antecedent of service brand personality is the service provider,
whereas in the case of product brands advertisements, user imagery and product
features play a key role.
113
Another aspect to be considered in stimuli selection is the number and type
of brands. If a large number of brands are used for developing the scale, the
generalizability and robustness of the scale will be better. But the problem with a
large number of brands is response bias resulting out of boredom and respondent
fatigue. Hence, an optimum number of brands are to be selected which can
maximize gneraliability while reducing respondent fatigue. Based on the previous
researches in this area and considering the Indian scenario, the number of brands
to be considered in this stage was fixed to be 21. Further, to increase
generalizability of the scale, brands from different product categories were
considered in the process. Hence the list of top 200 brands was classified in to
different product categories and a representative brand was selected randomly
from each category using simple random sampling technique (lottery method).
The brands considered for the study are listed in the table given below.
Table-4.1. Brands selected for the study
No. Brand Name No. Brand Name No. Brand Name 1 LEVI'S 8 PONDS 15 AXE
2 MRF 9 LG 16 ASIAN PAINTS
3 GARNIER 10 PEPSI 17 RAYBAN
4 AMUL 11 HERO MOTOR CORP.
18 ADIDAS
5 DABUR 12 LUX 19 BRITANNIA
6 DELL 13 CADBURYS 20 NOKIA
MOBILE
7 COLGATE 14 TITAN 21 GODREJ
[Source: Original]
4.5. Item Generation
There are mainly three areas which Aaker (1995) suggests to be considered
while generating the item pool. Firstly, the selection of the item should be based
114
on existing literature or theory. This was to ensure that item selection was based
on a theoretical framework. Secondly, meaningfulness and familiarity of the items
considered for the study. Third, the importance of each item with respect to the
construct under consideration, to be examined in order to establish external
validity of the measurement tool. Hence, to achieve the above mentioned
objectives, a three phase proceedure was adopted in this research
Phase 1: Generation of Items based on literature survey
Phase 2: Language expert opinion survey
Phase 3: Free elicitation study
4.5.1. Qualitative Study: Free Association Test
The purpose of this study was to generate more items that were relevant
and specific to the Indian context. For this study, a set of 34 MBA students from
different parts of the country were selected. The sample consisted of 35.3%
females and 64.7% male. Firstly, respondents were detailed with the concept of
brand personality and two examples were also given in order to make them more
familiar with the concept. They were then required to select two brands each, from
three product categories and describe them with as many traits/ adjectives as
possible (See Appendix I). The Product categories consisted of Symbolic products
(like Jeans, Cosmetics and Perfume/Fragrance), Utilitarian products (like Laptops,
Electronics, Appliances and Personal care products), and Symbolic-and-Utilitarian
products (like Motorbikes, Smart phones and Athletic Shoes). This categorization
was made as per the guidelines of Aaker (1997). This exercise resulted in 119
unique adjectives.
115
Results
The adjectives derived from the qualitative study are presented in the following table
Table-4.2. List of Adjectives from Free Association Test
Sl. No. Adjectives Sl. No. Adjectives Sl. No. Adjectives 1 Accountable 41 Fashionable 81 Posh 2 Aesthetic 42 Fast 82 Powerful 3 Amorous 43 Flawless 83 Prestigious 4 Aristocratic 44 Flyer 84 Pride 5 Attention Seeker 45 Fresh 85 Professional 6 Attractive 46 Fun 86 Pure 7 Awesome 47 Gallant 87 Quality 8 Beautiful 48 Global 88 Quick 9 Believable 49 Good Looking 89 Racy 10 Best 50 Handsome 90 Reliable 11 Bold 51 Happiness 91 Rich 12 Bright 52 Harmless 92 Richness 13 Capacity 53 Healthy 93 Rough 14 Carefree 54 Helpful 94 Royal 15 Catchy 55 Hot 95 Rugged 16 Charming 56 Iconic 96 Simple 17 Cheap 57 Indian 97 Sleek 18 Chic 58 Innovative 98 Smart 19 Class Apart 59 International 99 Sober 20 Classy 60 Joy 100 Solid 21 Classic 61 Killer 101 Soothing 22 Clean 62 Leonine 102 Speedy 23 Colourful 63 Lively 103 Sporty 24 Comfortable 64 Lovable 104 Strong 25 Compatible 65 Loyal 105 Sturdy 26 Cool 66 Luxurious 106 Stylish 27 Cosy 67 Macho 107 Superior 28 Courageous 68 Magnetic 108 Top class 29 Creative 69 Magnificent 109 Tough 30 Dashing 70 Manly 110 Traditional
116
31 Different 71 Masculine 111 Trendy 32 Distinct 72 Middle Class 112 Trust Worthy 33 Eco-Friendly 73 Mild 113 Trustable 34 Effective 74 Modern 114 Ultimate 35 Efficient 75 Multifaceted 115 Unique 36 Elegant 76 Natural 116 Friendly 37 Enviable 77 Outgoing 117 Versatility 38 Ethical 78 Passion 118 Vibrant 39 Excitement 79 Pleasing 119 Youth 40 Faithful 80 Pleasant
[Source: Original]
The result of the study indicates that Indian consumers also have a tendency
to attribute human characteristics to product brands.
4.5.2. Items from Literature Review
Another way used to create item pool was through an extensive literature
survey. Inclusion of items from literature helps to increase the content validity of
the scale. The scales which were used to derive the personality traits for this study
include scales developed by Aaker (1997), Ferrandi, Valette-Florence and
Finefalcy (2000), Aaker et al. (2001), Caprara et al. (2001), Ambroise et
al.(2005), Sung and Tinkham (2005), Sweeney and Brandon (2006), Okazaki
(2006), Van Rekom et al. (2006), d'Astous and Boujbel (2007), Geuenset
al.(2009), Heine (2009), Heere (2010), Lee et al.(2010), and Usakli and Baloglu
(2011).
Table-4.3. List of Adjectives from Literature survey
Sl. No. Items Sl. No. Items Sl. No. Items 1 Accepting 120 Fighter 239 Provocative 2 Accessible 121 Firm 240 Prudent 3 Accommodating 122 Flaunty 241 Puristic 4 Active 123 Flexible 242 Rakish 5 Adult 124 Flourishing 243 Rational 6 Adventurous 125 Forceless 244 Real
117
7 Affectionate 126 Forgiving 245 Realistic 8 Aggressive 127 Formal 246 Recent 9 Agreeable 128 Frank 247 Reflective 10 Alive 129 Freaky 248 Refreshing 11 Aloof 130 Free 249 Regular 12 Altruist 131 Fresh 250 Relaxed 13 Amusing 132 Friendly 251 Reliable 14 Approachable 133 Frugal 252 Religious 15 Aristocratic 134 Fun 253 Reserved 16 Arrogant 135 Funky 254 Resolute 17 Artificial 136 Funny 255 Respectable 18 Assertive 137 Gay 256 Responsible 19 Attractive 138 Generous 257 Reveller 20 Authentic 139 Genial 258 Rich 21 Balanced 140 Gentle 259 Rigorous 22 Beneficial 141 Genuine 260 Robust 23 Big 142 Glamorous 261 Romantic 24 Bigheaded 143 Golden 262 Rugged
25 Bohemian 144 Good-Looking 263 Ruthless
26 Boisterous 145 Good-Natured 264 Satisfying
27 Bold 146 Handy 265 Scrupulous 28 Bon-Vivant 147 Happy 266 Secure
29 Boring 148 Hard To Work 267 Self-contained
30 Bubbly 149 Hard-Working 268 Self-composed
31 Busy 150 Haughty 269 Self-effacing 32 Calculating 151 Healthy 270 Sensual 33 Calm 152 Heavy 271 Sentimental 34 Careless 153 Honest 272 Serene 35 Casual 154 Hopeful 273 Serious 36 Ceremonious 155 Humorous 274 Sexy
37 Charismatic 156 Ill-Mannered 275 Sharp
38 Charming 157 Imaginative 276 Shrewd 39 Chatty 158 Immoral 277 Showy 40 Chauvinist 159 Impersonal 278 Shrill 41 Cheerful 160 Impolite 279 Shy
118
42 Childlike 161 Important 280 Silent 43 Clean 162 Impractical 281 Simple 44 Clear 163 Impulsive 282 Sincere 45 Clumsy 164 Independent 283 Small-town 46 Cocky 165 Industrious 284 Smooth 47 Comfortable 166 Inefficient 285 Snobbish 48 Comforting 167 Informative 286 Sociable 49 Comical 168 Informed 287 Solid 50 Competent Person 169 Innovating 288 Sophisticated 51 Competitive 170 Innovative 289 Spirited 52 Confidence 171 Intelligent 290 Spiritual 53 Confident 172 Intense 291 Sporty 54 Connoisseur 173 Joyful 292 Stable 55 Conscientious 174 Kind 293 Steady 56 Conservative 175 Kitschy 294 Strict 57 Considerate 176 Laid-Back 295 Strong 58 Consistent 177 Lazy 296 Stunning 59 Conspicuous 178 Leader 297 Stylish 60 Constant 179 Leading 298 Successful
61 Contemporary 180 Level-headed 299 Superficial
62 Cool 181 Light-hearted 300 Sweet
63 Cooperative 182 Likeable 301 Sympathetic 64 Cordial 183 Lively 302 Talkative 65 Corporate 184 Logical 303 Technical
66 Countrified 185 Logo oriented 304 Temperamental
67 Courageous 186 Loyal 305 Tenacious 68 Cowardly 187 Mannered 306 Tense 69 Crazy 188 Masculine 307 Thoughtful
70 Creative 189 Mild-mannered 308 Tidy
71 Cruel 190 Minimalist 309 Timid 72 Cunning 191 Moderate 310 Tolerant 73 Curious 192 Modern 311 Tough 74 Cute 193 Modest 312 Traditional 75 Daring 194 Mysterious 313 Traditionalist 76 Decadent 195 Mystical 314 Tranquil 77 Decorated 196 Naïve 315 Trendy
119
78 Delicate 197 Naked 316 Tricky
79 Dependable 198 Narrow-Minded 317 True
80 Dependant 199 Natural 318 Trustworthy 81 Determined 200 Neat 319 Typical 82 Different 201 Neutral 320 Unaggressive 83 Diffident 202 New 321 Unargumentative 84 Dignified 203 Nice 322 Uncheery 85 Diligent 204 Normal 323 Uncomplex 86 Discreet 205 Offender 324 Unconceited 87 Discrete 206 Offish 325 Uncordial f88 Disrespectful 207 Old 326 Understated
89 Docile 208 Old-Fashioned 327 Undevious
90 Dominant 209 Openhearted 328 Unique
91 Donnish 210 Open-Minded 329 Unknown
92 Down-To-Earth 211 Optimistic 330 Unobtrusive 93 Dubious 212 Orderly 331 Unrefined 94 Dynamic 213 Ordinary 332 Unreliable 95 Easy 214 Organized 333 Unremarkable 96 Easygoing 215 Original 334 Unrevealing 97 Efficient 216 Ostentatious 335 Unselfconscious 98 Egocentric 217 Outdoorsy 336 Unsly 99 Elegant 218 Outgoing 337 Untidy
100 Emotional 219 Over forward 338 Unvain
101 Energetic 220 Overstated 339 Upper-class 102 Engaging 221 Passionate 340 Up-to-date 103 Enthusiastic 222 Patient 341 Urban
104 Environmentally-Friendly 223 Peaceful 342 Versatile
105 Exciting 224 Persistent 343 Vibrant 106 Exotic 225 Playful 344 Violent 107 Expensive 226 Pleasant 345 Voluptuous 108 Exploitative 227 Poised 346 Vulgar 109 Expressive 228 Popular 347 Warm 110 Extravagant 229 Positive 348 Warmth less 111 Faithful 230 Practical 349 Well-made 112 Familiar 231 Precise 380 Well-mannered
120
113 Family-Oriented 232 Prestigious 351 Western 114 Famous 233 Pretenseless 352 Wholesome 115 Fanciful 234 Pretentious 353 Wimpy 116 Fast 235 Productive 354 Young 117 Feminine 236 Professional 355 Youthful 118 Fervent 237 Progressive
119 Feisty 238 Proud [Source: Original]
4.6. Language Expert Opinion Survey
Since, the list of items derived from literature review were based on
different countries, many of the items could be inappropriate and irrelevant in the
Indian context. This necessitated the need for screening the appropriateness of
items in the Indian context. A group of three English language teaching experts
scrutinized the appropriateness of the items in the Indian context. They were
instructed to give alternative words, if existed, which will give a similar meaning
in the Indian context (See Appendix II for the questionnaire). Based on their
opinion, the scale items were modified and the items for which there were no
appropriate Indian words were removed from the item pool. The main objective of
this step was to make the items more relevant in the Indian context.
4.7. Expert Opinion Survey
With the intention to further reduce the personality inventory to a
manageable level, an expert opinion survey was conducted. For this, a total of 12
experts: three Marketing professors, three Consumer psychologists, three
professionals from the Advertising Industry and three brand managers served as
experts. An expert opinion survey was opted against a consumer survey at this
stage for three reasons: experts can better rate the items based on the construct
definition considered in this study; as against consumers, experts will have a better
understanding of the concept and hence they will be in a position to rate the items
more appropriately than the ordinary consumers; a small sample of experts can be
as reliable as a large sample of consumer respondents. Owing to the huge response
121
time required for this study, relying on a large sample of consumers will be
inappropriate and impractical. This large questionnaire can lead to respondent
fatigue and bias.
The experts were introduced to the brand personality definition based on
which they had to rate the items. They were required to rate one brand of their
choice from three types of product categories, namely, symbolic, utilitarian and
symbolic-and-utilitarian. The average time taken for this exercise was around 50
minutes. Both online and offline questionnaires were employed for this survey
(See Appendix III for the questionnaire used). The respondents were required to
rate the items on a 5-point scale with 1 indicating ‘Very inappropriate’ to describe
the brand and 5 indicating ‘Very appropriate’ to describe the brand. Items which
scored 4 or more were selected to the next level of item reduction.
4.8. Consumer Survey
In this phase, items from the Expert opinion survey were further reduced
based on a consumer survey. A student sample was used for this study. A total of
21 brands were considered. The development of the scale based on a large number
of brands has the advantage of increasing the generalizability of the scale. But,
relying on a large number of brands can lead to respondent fatigue and bias.
Hence, it is required to adopt an optimum number of brands which will not cause
respondent fatigue and at the same time will be generalizable. Any branding or
advertising strategy in India is developed for the entire country and hence any
scale developed relying on any single state would be inadequate and inappropriate
in the Indian context. Hence, a pan-Indian sample was required to develop the
scale. The sample was made representative of Indian population in terms home
state, income and gender. For example among the respondents included for the
study, nearly 53 percent were males, around 16 percent were from Uttar Pradesh
(Uttar Pradesh is the most populated state in India). About 78 percent of
respondents were Hindus.
122
Respondents who participated in the survey were introduced with the
concept of brand personality and were given the following instruction “This is a
study to understand how people perceive different brands. To elaborate, if I ask
you to give me your impression about a particular person, you may describe him
as a smart, sincere and competent person. Similarly think of a brand as if it were a
person. For instance you may think Raymond to be a respectable, responsible, and
a caring brand or Thums Up to be a daring, spirited and adventurous brand. The
purpose of this study is to understand which of the adjectives given below are most
appropriate to describe a brand. I request you to spare a few minutes of your
valuable time to take part in this study.”
The respondents were required to rate one brand each from three groups.
The purpose of grouping was twofold; one, to give choice to the respondents so
that they will have enough choices to opt from and second, to control the
responses in such a manner as to get fairly equal number of responses for each
brand. The respondents were required to rate the personality traits on a 5-point
Likert scale (where 1 represented Very Inappropriate, 2 – Moderately
Inappropriate, 3 – Neutral, 4 – Moderately Appropriate and 5 – Very
Appropriate). They were requested to rate the appropriateness of each personality
trait in describing the brand they opted (See Appendix IV for questionnaire).
4.8.1. Measure Purification: Exploratory Factor Analysis
A total of 606 respondents rated a total of 96 personality traits based on 21
brands. Since the objective of this research is to develop a set of brand personality
factors, which can be used to profile brands, it is necessary to analyze the inter
correlation of personality traits across brands (Aaker, 1995). The 96*96
correlation matrix was subjected to Exploratory Factor Analysis (EFA) with
principal components method (see Nunally, 2010). A varimax rotation was
employed during the proceedure with the objective of identifying appropriate
123
factors, rather than ending up with a single large affective factor (Aaker, 1995).
The process of EFA resulted in a 9-factor strucure which explained around
61.42% of the variance.The EFA results are presented in the form of a table
below.
Table-4.4. Exploratory Factor Analysis Results
Factor Variables Variance explained
(%)
Cumulative Percentage
of Variance
F1 Happiness, Honest, Joyful, Good Natured, Positive, Sincere, Pleasant, Enthusiastic, Optimistic, Well-Mannered, Independent, Passionate, Friendly, Clear, Wholesome, Engaging, Lively, Precise, Practical, Outgoing, Loyal, Rational
21.652
21.652
F2 Stylish, Sexy, Youthful, Trendy 7.239 28.892
F3 Reliable, Trust Worthy, Respectable, Believable
6.682 35.574
F4 Aggressive, Bold, Adventurous, Daring, Courageous,
6.43 42.00
F5 Competent, Competitive, Compatible, Comfortable
4.839 46.844
F6 Accessible, Accountable, Accepting 4.44 51.286 F7 Charming, Charismatic, Appealing 4.165 55.451 F8 Innovative, Creative, Professional 4.164 55.451 F9 Dominant 2.343 61.426
[Source: Original]
4.8.2. Scale Reliability
The next objective was to check the reliability of the scale or to evaluate the
extent to which the scale is free from errors and would show consistency in the
results (Peter, 1981). Reliability can be measured mainly in two ways namely,
Cronbach’s alpha and Test re-test correlations.
124
The Cronbach’s alpha values of the first eight dimensions were 0.9565, 0.774,
0.769, 0.808, 0.799, 0.765, 0.743 and 0.567 respectively. The recommended
value for Cronbach’s alpha is 0.7 (Hair, Black, Babin, and Anderson, 2012). After
this stage of the research two dimensions, factors 8 and 9 were dropped from
further analysis. Factor 8 was dropped because of a poor Cronbach’s alpha value
whereas factor 9 was dropped because it contained only one item.
Test re-test reliability: A vital step to prove a scale to be reliable is to prove it is
stable over time. For this a survey was conducted among the respondents chosen
from those who participated in the first measurement purification phase. This was
conducted one month after their first response was collected. The data was
collected from 91 respondents. Although the reliabilities were calculated only for
the items selected to be included in the final structure, all the items which were
included in the questionnaire for first phase item purification, were included here
also as per Aaker’s (1995) recommendations. The rationale of this was to avoid
systematic bias which may occur in the test re-test analysis. The one month time
interval was given with the intention of avoiding any potential memory effects
resulting from their previous responses. The test re-test correlation values ranged
from 0.7 for Precise to 0.93 for Happiness. The test re-test values for the seven
factors and each traits are presented in the following tables.
Table-4.5. Factor wise test re-test correlation values
Sl. No. Factor Test re-test correlation 1 Factor 1 0.87 2 Factor 2 0.91 3 Factor 3 0.9 4 Factor 4 0.83 5 Factor 5 0.81 6 Factor 6 0.84 7 Factor 7 0.82
[Source: Original]
125
Table-4.6. Variable-wise test re-test correlation values
Sl. No. Traits Correlation Sl. No. Traits Correlation Sl. No. Traits Correlation
1 Happiness 0.93 16 Engaging 0.74 31 Aggressive 0.72 2 Honest 0.82 17 Lively 0.74 32 Bold 0.74 3 Joyful 0.78 18 Precise 0.7 33 Adventurous 0.71 4 Good-natured 0.76 19 Practical 0.76 34 Daring 0.75 5 Positive 0.72 20 Outgoing 0.74 35 Courageous 0.76 6 Sincere 0.8 21 Loyal 0.73 36 Competent 0.78 7 Pleasant 0.81 22 Rational 0.7 37 Competitive 0.77 8 Enthusiastic 0.76 23 Stylish 0.93 38 Compatible 0.8 9 Optimistic 0.73 24 Sexy 0.87 39 Comfortable 0.82 10 Well-mannered 0.71 25 Youthful 0.84 40 Accessible 0.82 11 Independent 0.81 26 Trendy 0.76 41 Accountable 0.74 12 Passionate 0.74 27 Reliable 0.82 42 Accepting 0.71 13 Friendly 0.75 28 Trustworthy 0.9 43 Charming 0.79 14 Clear 0.76 29 Respectable 0.71 44 Charismatic 0.72 15 Wholesome 0.81 30 Believable 0.85 45 Appealing 0.82
[Source: Original]
126
4.9. Confirming Product Brand Personality Scale (PBPS) dimensions
Aaker (1995) and Numally (2010) suggest a hold out sample study
mainly for two reasons; first is to check out the sufficiency of the first set of
brands in determining the factorial structure. By reproducing the same factorial
structure with another set of brands, it can be proven that there was no bias in
the structure because of brand selection. Second, it could check whether the
factor structure was a function of subject sample, or in other words, this hold
out sample will help to check whether the factor structure was the result of the
particular set of respondents participated in the study. The stimuli or the brands
considered in this phase were also drawn randomly from the same source, the
list developed by Trust Research Advisory and Brand Equity listing. A total of
12 brands were considered in this stage. The selected brands were Vicks, Rin,
Whirlpool, Dettol, Lee, Horlicks, Bata, Hyundai, Lenovo, Nescafe, Mirinda
and Parle-G.
For this study also a student sample was used. Responses were collected
from 217 students. As done in the earlier study, respondents were required to
rate three brands each of their choice. They were instructed to give their
responses on a 5-point Likert scale and had to rate 45 personality traits based
on the specific brands they opted (See Appendix IV for questionnaire). Since
the objective of this stage was to determine the robustness of the factor
structure across a new set of brands and respondents, a confirmatory factor
analysis was carried out using the LISREL software package.
127
Figure-4.1. Confirmatory factor analysis model of PBPS
[Source: Original]
128
The Chi Square value was significant (2238.09 with 924 degrees of
freedom; P=0.0).This is a common phenomenon in models with large sample
sizes. Because the Chi Square test is sensitive to sample size and can lead to a
rejection of a model differing in a trivial way from the data for large sample
sizes. It is prudent to examine other measures of fit (Bagozzi and Heatherton,
1994). On the other hand, where small samples are used, the Chi-Square
statistic lacks power and because of this, it may not discriminate between good
fitting models and poor fitting models (Kenny and McCoach, 2003). Due to the
restrictiveness of the Model Chi-Square, researchers have sought alternative
indices to assess model fit. One example of a statistic that minimizes the impact
of sample size on the Model Chi-Square is Wheaton’s (1977) relative/ normed
chi-square (χ2/df) [df denotes degrees of freedom]. Although there is no
consensus regarding an acceptable ratio for this statistic, recommendations
range from as high as 5.0 (Wheaton, 1977) to as low as 2.0 (Tabachnick and
Fidell, 2007). Hence, it is required to look in to the other fit indices. Four main
fit indices are reported here considering the recommendations of Hu and
Bentler’s Two-Index Presentation Strategy (1999).
Table-4.7. Evaluation of model fit
Fit Index Estimated Value
Acceptable Threshold Levels
Relative χ2 (χ2/df )
2.423 3:1 (Kline, 2011)
Non-Normed Fit Index (NNFI) 0.96 Values greater than 0.95
Standardized root mean square residual (SRMR)
0.056 SRMR less than 0.08 (Hu and Bentler, 1999)
Comparative Fit Index (CFI) 0.96 Values greater than 0.95
[Source: Original]
The relative Chi square value, SRMR, NNFI and CFI, presented above
indicate a good model fit. In conclusion, the results of the above confirmatory
129
factor analysis indicated that the seven-factor structure is robust and stable over
different set of stimuli and sample.
4.9.1. Construct Validity
Validity refers to the extent to which research is accurate (Hair et al.,
2012). A most popular and widely accepted tool to check construct validity of
the proposed measurement model is Confirmatory Factor Analysis (CFA).
Construct Validity refers to the extent to which the items used to measure the
construct actually reflect the theoretical latent dimensions those items are
designed to measure (Hair et al., 2012). For this study, Construct Validity was
evaluated in terms of Convergent validity and Divergent validity. Convergent
validity implies that “Items that are indicators of a specific construct should
converge or share a high proportion of variance in common” (Hair et al., 2012).
This is referred to as convergent validity. Hair et al., (2012) suggest different
ways to calculate the relative amount of convergent validity amid item
measures. They point that high loadings on a dimension indicates that the items
unite on a common point called latent construct. The factor loading values in
the standardized estimates of the CFA model ranges from 0.68 for outgoing to
0.90 for daring, which is higher than the recommended minimum value of 0.5
by Hair et al., (2012). Next, the average variance extracted was calculated and
the result was as follows:
Table-4.8. Average Variance Extracted
Factor Average Variance Extracted
Factor 1 0.603914
Factor 2 0.677775
Factor 3 0.64254
Factor 4 0.64685
Factor 5 0.69975
Factor 6 0.651533
Factor 7 0.714167
[Source: Original]
130
The Average Variance Extracted for all factors were greater than 0.5
which indicates that on an average less error remains in the item than what is
explained by the latent construct. In sum, the factors exhibit good convergent
validity.
Another indicator of convergent validity is construct reliability (Hair et
al., 2012). Construct reliability is generally used in conjunction with Structural
Equation Models (SEM). It can be calculated by using the formula
𝐶𝐶𝐶𝐶 =(∑ 𝐿𝐿𝐿𝐿𝑛𝑛
𝐿𝐿=1 )2 (∑ 𝐿𝐿𝐿𝐿𝑛𝑛
𝐿𝐿=1 )2 + (∑ 𝑒𝑒𝐿𝐿𝑛𝑛𝐿𝐿=1 )
where Li represents factor loadings and ei represents error variance. The
construct reliability calculated for each construct is given in the following table
Table-4.9. Construct Reliability values
Sl. No. Factors Construct Reliability 1 Factor 1 0.969064862
2 Factor 2 0.894250708
3 Factor 3 0.899435596
4 Factor 4 0.878969957
5 Factor 5 0.899239053
6 Factor 6 0.847967103
7 Factor 7 0.881557383
[Source: Original]
The construct reliability of the factors ranges from 0.84 to 0.96, which is
much higher than the recommended value of 0.7. This indicates that there is a
high internal consistency within the factors, which in turn implies that all the
measures consistently represent the same latent construct.
131
Discriminant Validity
It refers to the extent to which a construct is truly distinct from other
constructs (Hair et al., 2012). To check discriminant validity, the Average
Variance Extracted (AVE) scores of the two constructs needs to be compared
with the square of the correlation estimates between these two constructs. From
the table given below, it is clear that the variance extracted estimates (Diagonal
values) are greater than the squared correlation estimate. This implies that the
latent constructs explains more of the variance in its item measures than it
shares with another construct. This also implies that the individual indicators
measure only one latent measure, which also implies that there are no cross
loadings for the indicator variables.
Table-4.10. Evaluation of discriminant validity
F1 F2 F3 F4 F5 F6 F7 F1 (0.6039)
F2 0.2025 (0.677775)
F3 0.1369 0.3249 (0.64254)
F4 0.3844 0.1024 0.16 (0.64685)
F5 0.3364 0.1521 0.1296 0.4225 (0.699)
F6 0.2809 0.09 0.0625 0.3969 0.2601 (0.6515)
F7 0.3136 0.2704 0.3364 0.16 0.2809 0.2601 (0.714167)
Notes: Diagonal elements (values in parentheses) are the Average Variance Extracted (AVE); off-diagonal elements are the square correlations among constructs.
[Source: Original]
132
4.10. A short version of the PBPS
The objective of this section is to create a short version of the PBPS.
Since brand personality is often measured along with other concepts such as
self concept, celebrity personality and so on; a large number of items in the
scale will make it practically insignificant. Hence many researchers have
developed a short version of their scales (For example, Aaker, 1997; Burisch,
1997; Rammstedt and John, 2007; Geuens et al., 2009). To meet this purpose
as well as to satisfy structural model validity, the maximum number of items to
be included under a particular factor was decided to be 4 and minimum as 3.
Item selection was done with the help of factor analysis results. Since the first
factor comprised of 22 items, they were factor analyzed and four surrogate
variables were identified to represent them. They include Honest, Enthusiastic,
Precise and Outgoing. Again, Courageous was dropped from factor 4 (which
contained five items initially) being the lowest loading item. All the other
factors contained three or four items only, so they were kept intact. This
resulted in the final version of the scale with 7 factors and 26 items. The
psychometric properties of this short version were also evaluated by means of
evaluating test re-test reliability, Cronbach’s alpha values, Convergent validity
and Discriminant validity. The final set of factors and corresponding items are
given in the following table.
Table-4.11. Short version of the PBPS
Factor Variables
F1 Honest, Enthusiastic, Precise, Outgoing
F2 Stylish, Sexy, Youthful, Trendy
F3 Reliable, Trust Worthy, Respectable, Believable
F4 Aggressive, Bold, Adventurous, Daring
F5 Competent, Competitive, Compatible, Comfortable
F6 Accessible, Accountable, Accepting
F7 Charming, Charismatic, Appealing
[Source: Original]
133
4.10.1. Checking the reliability of the short version of PBPS
The reliability of the short PBPS was evaluated based on Cronbach’s
alpha values and test re-test reliability values.
Cronbach’s Alpha Values
The Cronbach’s alpha values of the seven dimensions were calculated
using SPSS software package and the result is presented in the following table.
Table-4.12. Cronbach’s Alpha Values of the short scale
Sl. No. Factors Alpha Values
1 Factor 1 0.795
2 Factor 2 0.774
3 Factor 3 0.769
4 Factor 4 0.771
5 Factor 5 0.799
6 Factor 6 0.765
7 Factor 7 0.743
[Source: Original]
Cronbach’s alpha values for all the variables are found to be greater than the
recommended value of 0.7.
134
Test re-test reliability
Test re-test values were also examined for the seven dimensions of the
short version of the product brand personality scale and the results are
presented below.
Table-4.13. Test re-test correlation values of the factors in the short scale
No. Trait Test re-test correlation
1 Factor 1 0.805
2 Factor 2 0.91
3 Factor 3 0.90
4 Factor 4 0.813
5 Factor 5 0.81
6 Factor 6 0.84
7 Factor 7 0.82
[Source: Original]
All the values were greater than 0.7 and hence it is concluded that the scale
possesses test re-test reliability.
4.10.2. Confirmatory Factor Analysis (CFA) for short version of PBPS
To confirm the seven-factor structure of the short version of product
brand personality scale, a confirmatory factor analysis was done using LISREL
software package. The data collected for the hold out sample study for the full
version of the scale was used in this phase.
135
Figure-4.2. Confirmatory factor analysis model for short version of PBPS
[Source: Original]
136
As discussed before in this chapter, four main fit indices are reported
here considering the recommendations of Hu and Bentler’s Two-Index
Presentation Strategy (1999).
Table-4.14. Fit indices for the evaluation of CFA model of the short PBPS
Fit Index Estimated Value
Acceptable Threshold Levels
Relative χ2 (χ2/df )
2.169 3:1 (Kline, 2011)
Non-Normed Fit Index (NNFI)
.96 Values greater than 0.95
Standardized root mean square residual (SRMR)
0.054 SRMR less than 0.08 (Hu and Bentler, 1999)
Comparative Fit Index (CFI)
0.97 Values greater than 0.95
[Source: Original]
Since all the fit indices possess recommended values, it can be
concluded that the confirmatory factor analysis confirms the model validity of
the short PBPS.
4.10.3. Construct validity
For the short version of the PBPS scale, Construct Validity was
evaluated in terms of Convergent validity and Discriminant validity (employed
a similar procedure adopted for the validation of the full version of PBPS scale
developed in the previous section).
Convergent validity
Convergent validity of the scale was evaluated based on factor loadings,
average variance extracted and construct reliability. To check the convergent
validity of the scale, the factor loading values in the standardized estimates of
the CFA model was checked. Its values ranged from .65 for Outgoing to .91 for
Charismatic, which is higher than the recommended minimum value of 0.5 by
137
Hair et al., (2012). Next, the average variance extracted was calculated and the
result was as follows:
Table-4.15. Average Variance Extracted value in the case of short PBPS
Sl. No. Factor
Ave. Variance Extracted
1 Factor 1 0.533475 2 Factor 2 0.68165 3 Factor 3 0.64455 4 Factor 4 0.6512
5 Factor 5 0.686775 6 Factor 6 0.651533 7 Factor 7 0.7145
[Source: Original]
The Average Variance Extracted for all factors were greater than 0.5
which indicates that on an average, less error remains in the item than what is
explained by the latent construct. In sum, the factors exhibit good convergent
validity.
Construct reliability
Next, indicator to be checked to ensure convergent validity is construct
reliability (Hair et al., 2012). It was calculated using the formula
𝐶𝐶𝐶𝐶 =(∑ 𝐿𝐿𝐿𝐿𝑛𝑛
𝐿𝐿=1 )2 (∑ 𝐿𝐿𝐿𝐿𝑛𝑛
𝐿𝐿=1 )2 + (∑ 𝑒𝑒𝐿𝐿𝑛𝑛𝐿𝐿=1 )
where Li represents factor loadings and ei represents error variance. The
construct reliability calculated for each construct is given in the following
table.
138
Table-4.16. Construct Reliability values of short PBPS dimensions
No. Factors Construct Reliability
1 Factor 1 0.818324137
2 Factor 2 0.895559211
3 Factor 3 0.876712329
4 Factor 4 0.878881961
5 Factor 5 0.897592188
6 Factor 6 0.847967103
7 Factor 7 0.880344937
[Source: Original]
The construct reliability of the factors ranges from 0.818 to 0.897, which
is much higher than the recommended value of 0.7. This indicates that there is
a high internal consistency within the factors, which in turn implies that all the
measures consistently represent the same latent construct.
Discriminant Validity
As discussed before in the previous section on Discriminant validity, the
Average Variance Extracted (AVE) scores of the two constructs were
compared with the square of the correlation estimates between those two
constructs. From the table given below, it is clear that the variance extracted
estimates (diagonal values) are greater than the squared correlation estimate.
This implies that the latent constructs in the short version of the PBPS explains
more of the variance in its item measures than it shares with another construct.
This also implies that the individual indicators measure only one latent
measure, which also implies that there are no cross loadings for the indicator
variables.
139
Table-4.17. Evaluation of the discriminant validity of the short PBPS
Factor 1
Factor 2
Factor 3
Factor 4 Factor 5
Factor 6
Factor 7
Factor 1
(0.533475)
Factor 2
.25 (0.68165)
Factor 3
.1521 .3364 (0.64455)
Factor 4
.36 .1024 .1369 (0.6512)
Factor 5
.3136 .16 .1156 .42 (0.686775)
Factor 6
.2209 .09 .0529 .3969 .2601 (0.651533)
Factor 7
.36 .2704 .3249 .16 .2809 .2601 (0.7145)
Notes: Diagonal elements (values in parentheses) are the Average Variance Extracted (AVE); off-diagonal elements are the square correlations among constructs
[Source: Original]
4.11. Naming of seven factors
Next, a name was chosen for each dimension which reflected the general
characteristics of that dimension. This naming was designed considering both
the full and short version of the PBPS.
Table-4.18. Names of seven factors identified
Factor Name Factor 1 Happy Factor 2 Youthful Factor 3 Reliable Factor 4 Adventurous Factor 5 Competent Factor 6 Accountable Factor 7 Appeal
[Source: Original]
140
4.12. Chapter conclusion
This chapter has discussed the scale development procedure adopted in
this research. The various sections herein, have elaborated on the item
generation procedure, stimuli selection procedure, item purification methods
adopted, procedures to check the reliability and validity of the scale and so on.
A seven dimensional framework for measuring the product brand personality in
the Indian context was designed. The psychometric properties of this scale
were established with the help of various statistical analysis techniques.
141
Chapter 5 FINDINGS, DISCUSSION OF RESULTS AND
CONCLUSION
This chapter includes a discourse on the conclusions, limitations and
implications of this thesis. An effort is made to identify and suggest the various
important research opportunities in this domain.
5.1. Introduction
In the competitive market scenario characterized by proliferation of
brands, brand managers are striving hard to understand and manage the
symbolic benefits their products offer to target consumers. Brand personality is
one among those important benefits. Though, the brand personality concept
originated long ago in branding, unlike its counterpart in psychology (that is,
human personality), the concept is still in its nascent stage. Hence, more
researches are required to build this field. Relying on a stricter definition put
forward by Azoulay and Kapferer (2003) and considering product brands in the
Indian context, a new product brand personality scale has been developed. The
new scale consists of seven dimensions. Through different phases in the scale
development process the researcher has established the content validity,
reliability, construct validity and dimensionality of the scale. With a relatively
less number of items, the new tool proves to be a practical instrument for
branding research. The new tool will be helpful for both academicians and
practitioners. The new scale is expected to simplify theorizing and hypothesis
generation since one scale can be used for any product category.
The following sections in this chapter will elaborate on the major
findings and their implications, besides providing a comparison of the present
scale with the existing ones.
142
5.2. Major findings
The major findings and implications of this research is summarised as
follows. Though all the phases adopted in this research were important, not
every phase provide direct implications for research. However, the most
important ones are discussed here.
The free elicitation study which was conducted as a part of the item
generation phase indicates that Indian consumers do attribute personality
characteristics to brands. This in a way adds support to the previous researches
in this field which suggests that consumers have a tendency to attribute
personality characteristics to brands and talk about them.
The language expert opinion study in this research indicated that, many
items that are relevant in foreign contexts were identified to be irrelevant or
less appropriate in the Indian context. The qualitative study done as a part of
the expert opinion survey did not yielded any extra items to the initial pool
generated, which is an indication of the exhaustiveness of the initial item pool.
Another important result that needs to be mentioned is findings from
exploratory factor analysis done with the data obtained through consumer
survey. The results indicate that, a total of nine dimensions can explain around
61.43% of the total variance. The contribution to variance by each factors were
in the order 21.65, 7.23, 6.68, 6.43, 4.83, 4.44, 4.16, 3.63, and 2.34. However,
the reliability measure results (Cronbach’s alpha values) suggested the removal
of last two dimensions. A test re-test correlation values for the remaining seven
dimensions indicated a good test re-test reliability. Finally, this seven
dimension structure was proved to be a reliable structure to represent product
brand personality in the Indian context. The next important result worth
mentioning is the hold out sample survey conducted to check the validity of the
seven factor structure identified in the previous consumer survey phase. The
results of this study confirmed the seven dimensional structure of product
brand personality. The results of this study also provided proof for the
143
construct validity. The results of the analysis such as factor loadings, average
variance extracted, and construct reliability provided support for convergent
validity of the scale. Whereas a comparison of the average variance extracted
and square correlation among constructs provided support for discriminant
validity. Further, a study done to develop a short version of the scale indicated
that a seven factor structure with 26 items to be adequate enough to measure
product brand personality. The following diagram presents the final version of
the scale.
144
Figure-5.1. Product brand personality scale
[Source: Original]
Product brand personality scale
Happy
Honest, Enthusiastic, Precise, Outgoing
Youthful
Stylish, Sexy, Youthful, Trendy
Reliable
Reliable, Trust Worthy, Respectable, Believable
Adventurous
Aggressive, Bold, Adventurous, Daring
Competent
Competent, Competitive, Compatible, Comfortable
Accountable
Accessible, Accountable, Accepting
Appeal
Charming, Charismatic, Appealing
145
5.3. Originality and novelty
Next, it is worth discussing the originality and contributions of this
research. Though much study exists on the personality of brands in general, no
studies have identified the personality of product brands in particular. This
research addresses this lacuna by following a methodical process in bringing out
the dimensions of product brand personality. Also this is a scale developed in the
Indian context and will be of great use for Indian researchers and marketing
practitioners.
This research has significant implications for theorizing product brand
personality and contributes to marketing theory in terms of a tool to measure
product brand personality. Preliminary studies on the existing brand personality
scales indicate the necessity of a product specific brand personality scale. The
PBPS is parsimoniously presented in terms of a seven dimensional 26 item
measure. For a scale with dimensions, Mowen and Voss (2008) recommends three
to five items per dimension. The PBPS is in line with this recommendation. From
a pragmatic point of view, this scale is relatively short and will help to reduce
respondent fatigue issue, which arise in the case of most of the existing brand
personality scales.
When compared with a five dimensional structure, the seven dimensional
structure of the construct provides more options for the branding practitioners to
position their brands.
5.4. Implications
This tool will be a boon for both academicians and practitioners. For
practitioners this PBPS will help them to contrast between the perceived and
portrayed personality of their brands which will in turn help them in managing
their brands in a better manner. Further, it will help marketers to impart a
146
personality to their brands as desired by their consumers. . The new product brand
personality scale may also help the brand manager to ensure that the personality of
the proposed brand extensions is in line with that of the existing one.
5.4.1. How a brand manager may use the PBPS
A brand manager may use the product brand personality scale to evaluate
the personality of his brand. This can help him/her to understand whether the
target consumers of the brand perceive the same personality which the company
tries to communicate to them. The scale may be administered in a questionnaire
form to target consumers and their perceptions may be gathered. A summated
score of a representative set of target consumers can help to identify how the
personality of the brand is perceived. Based on the results, the practitioner may
modify or manage the personality of the brand. He may also make comparison
with rival brands to formulate new strategies.
The product brand personality scale may be administered to the target
consumers of the brand and they may be asked to rate the brand based on the traits
in the scale. The average of the consumer perception will indicate the performance
of the brand on different personality dimensions. Brand manager may use a radar
chart to better represent the personality of the brand.
Also, a longitudinal data collected using this scale (that is, data collected
over a period of time) may help one to understand how the personality of a brand
moves or changes over time. Another use of this data could be to understand the
impact of advertising on creating and changing brand personality.
5.4.2. Implications for Academics
For academics, this scale can serve as a building block for new theories in
brand management and consumer psychology. For instance, they can identify the
sources of various product brand personality dimensions or they may conduct
147
research to identify the relationship between various product personality
dimensions and consumer behaviour, and so on.
From a theoretical perspective this study reveals that consumers attribute
personality characteristics to product brands. This is in line with previous studies
on brand personality (for example Aaker, 1997; Ambroise et al., 2005; Bosnjak et
al., 2007). This study reveals that the Indian product brand personality framework
consists of seven factors. Further, this research supports the argument that Aaker’s
(1997) brand personality scale may not be applicable in the case of contexts other
than United States. This research informs academicians that the use of Aaker’s
(1997) scale in brand personality research in India would be less meaningful. It
equips them with a better tool for measuring brand personality in the Indian
context.
Another theoretical implication of this study is methodological. This
research indicates that both qualitative and quantitative methods can be employed
together to develop a brand personality scale.
5.5. A critical discussion on PBPS
The scale shares many commonalities and differences with the existing
scales in terms of the methodology adopted and the dimensions identified.
General approach: The mixed method approach adopted in this study is similar
to the works of Aaker (1997), Aaker et al. (2001), Geuens et al.(2009). However
one cannot find this approach in the works of researchers like Caprara et al.
(2001).
Adoption of a ‘meso’ approach against micro and macro approaches: The
introduction of the new approach of a ‘meso’ perception in brand personality
research adds to the novelty of this work. Unlike ‘micro’ the scope of this scale
will not be too narrow, which may reduce the applicability of the scale to compare
148
between different products personality dimensions. Also, this approach can be
considered more powerful than a ‘macro’ approach, since the reduced scope of the
scale will increase the accuracy in measuring the concept.
Focus/domain: The focus of this brand personality scale is product brands in the
Indian context whereas the previous researches in this field tried to focus on
brands in general or brands in a specific product category, with reference to
specific countries. For instance, consider Aaker (1997)’s general brand personality
scale in the U.S. context, Aaker et al. (2001)’s general brand personality scales for
Japan and Spain, Geuens et al. (2009)’s general brand personality scale in the
Belgium context, or retail store personality scale like the one by d’Astous and
Lévesque (2003), Helgeson and Supphellen (2004). In this regard the new scale is
different from all other scales, as it is neither too specific like scales that are
developed with a micto perspective (like a sports personality scale, destination
personality scale or a restaurant personality scale) nor too broad (like those
developed for capturing the personality dimensions of brands in general).
Rigor: Unlike Aaker (1997) this research relies on a stricter definition of brand
personality put forward by Azoulay and Kapferer (2003). Researchers who have
employed this definition in their scale development process includes: Bosnjak et
al. (2007)and Geuens et al. (2009). Methodologically, the scale development
process has adopted many of the best practices in brand personality research.
Initial item pool: The study has relied on a large item pool to narrow down to a
small set of representative items in a systematic manner, which ensures the content
validity of the scale. Other researchers like Aaker (1997), Aaker et al. (2001),
Geuens et al. (2009) have also adopted a similar procedure. Whereas scales
developed by Caprara, Barbaranellia and Guidob (2001), d’Astous and Lévesque
(2003), Helgeson and Supphellen (2004), Sung and Tinkham (2005) relied on
relatively small initial item sets. This research has started with an exploratory
149
study and then moved on to quantitative studies to identify and confirm brand
personality dimensions. This methodology is also observable in the studies done
by Aaker (1997), Aaker et al. (2001) and Geuens et al. (2009).
Sampling: The researchers who relied on student sample to create their brand
personality scale or to study the brand personality construct includes Austin et
al.(2003), Sung and Tinkham (2005), Ambroise et al. (2005), Magnini and Thelen
(2008), Musante et al.(2008), Carlson et al.(2009), Kim et al. (2010), Kaplan et
al.(2010). Other researchers differ with this product brand personality scale with
respect to the sample they adopted.
Dimensionality of the scale: This study has identified seven dimensions. A study
by Kim and Lehto (2013) also identified a seven factor structure consisting of
dimensions such as Family orientation, Sincerity, Competence, Uniqueness,
Excitement, Ruggedness and Sophistication. Some other researchers like Musante
et al. (2008), Das et al. (2012) mentioned in their work that they also identified a
seven factor structure, but judged a five factor structure to be more appropriate for
their study. Though many researchers tend to opt for a five factor structure, in
general, an inconsistency in the number of brand personality dimensions can be
observed. It is advisable to conduct further studies to reach a consensus on the
number of brand personality dimensions. A large number of studies have indicated
a difference of opinion or difference in results with the five dimensional structure
of Aaker (1997). See for instance, three dimensional brand personality structure
identified by Ekinci and Hosany (2006), Murphy et al., (2007); four dimensional
structure by Bosnjak et al. (2007), Murphy et al. (2007a); Six factor structure
proposed by D’Astous and Boujbel (2007), Kaplan et al. (2010), Smith et al.
(2006); eight factor structure by Sweeney and Brandon (2006); 12 factor structure
by Ambroise et al. (2005). Hence, the findings of this research indicate that, unlike
human personality a consensus on a five factor structure in the case of brand
personality dimensions may not be possible to reach.
150
5.6. Comparison of PBPS dimensions with other scales
The various dimensions of the new product brand personality scale are
compared against a few selected brand personality scales in the following sections.
Some of the brand personality dimensions identified by this India-specific
study have similarities with those of Aaker’s (1997) scale. For instance, the
Appeal dimension from this study and Sophistication dimension of Aaker’s scale
try to capture similar aspects of brand personality. However, Charming is the only
item shared in common between the two. The Competent dimension from this
study and Competence dimension of Aaker’s scale try to measure how well a
brand appears to be successful and efficient to its consumers. The Excitement and
Sincerity dimensions of Aaker’s scale seem to be captured by the Happy
dimension in the scale developed through this study. The Ruggedness dimension
of Aaker’s scale and Adventure dimension from this study are distantly similar.
However, the dimensions Reliable, Youthful and Accountable emerge as India-
specific dimensions.
In the case of Ambroise, Ferrandi and Merunka’s scale (2005), the
dimension Cheerful and Happy dimension of this new scale share some
commonality. Both scales have a dimension named Reliable, but the items used to
capture these dimensions differ. The exciting, outgoing dimensions in the case of
Ambroise et al.(2005) shows similarity to the adventurous dimension of PBPS.
Secure and accountable dimensions from the two scales also seem to capture
similar perceptions. Yet other two dimensions namely glamorous and elegant
from Ambroise et al. (2005) can be matched with the Appeal dimension of the
new scale. Other dimensions such as natural, sweet, mature, mischievous seem to
be unique to the scale developed by Ambroise et al. (2005). But, a very large
number of factors in the case of Ambroise et al. (2005) hiders its scope in practical
applications.
151
In the case of Sweeney and Brandon’s scale (2006), the similarities and
dissimilarities are same as discussed before in the case of Aaker (1997), as it is an
extension of Aaker’s (1997) scale. The three dimensions they added to Aaker
(1997), namely Agreeableness, Extroversion and Conscientiousness seems to be
less similar to any items in the newly developed PBPS. However, those three
dimensions were adapted from a human personality scale and the reliability and
validity of those items are questionable.
The scale by Bosnjak et al. (2007) is quite different from the new PBPS in
terms of both number and type of dimensions. This scale has a four dimension
structure against the seven-factor structure of the product brand personality scale.
Again, this scale contains a lot of negative items and facets, like Superficiality
(with items such as Hypocritical, Obtrusive, Arrogant and Selfish) and also other
items such as Boredom, Small-minded, Boring and Old-fashioned. The items that
share some similarity to items in the new scale are: Competent, Responsible,
Reliable, Adventurous and Exciting. However, this scale tends to concentrate
more on the negative aspects of a brand rather than positive aspects.
Another important scale worth compared to is the one by Geuens et al.
(2009), which is an ultra short scale with 12 items and five factors. Though, the
scale is claimed to have good reliability and validity scores, the ability of just 12
items to represent a large set of more than 300 initial items is questionable. The
new PBPS with 26 items would be more representative than this ultra short
version. The dimension Emotionality of the Geuens et al. (2009) scale has some
resemblance with the Youthful dimension of the new scale, Activity dimension
weakly resembles Adventurous dimension, Responsibility factor may be
comparable with the Accountable dimension and Aggressiveness with Adventurous
dimension. However, the Simplicity dimension was not found in the case of
Product brand personality scale. The dimensions such as Happy, Reliable,
152
Competent and Appeal seem to be different and unique to product brand
personality.
5.7. Limitations
No research can be free from limitations. Time and budget constraints
necessitated to limit the scope of the study. The use of a student sample to balance
the limits of time and budget hinders the generalizability of this study. This
practice in consumer research has been widely debated over the years, largely
without resolution (Beltramini, 1983). But from the literature review, it is apparent
that many researchers in the domain of brand personality research rely on student
sample to conduct their research (for example Austin et al, 2003; Sung and
Tinkham, 2005; Ambroise et al., 2005; Magnini and Thelen, 2008; Musante et al.,
2008; Carlson et al., 2009; Kim et al., 2010; Kaplan et al., 2010 and so on). This
practice might have arisen since a study by Aaker (1997) showed that there is no
significant difference in the brand personality dimensions based on demographic
features. Also, none of the existing brand personality research has reported any
variation in the factor structure based on demographic characteristics. Hence, it is
reasonable to use the new PBPS scale in the entire Indian context, until and unless
proved otherwise. India has a large youth demographic: 65 per cent of its
population is 35 and under. Also, there are nearly 150 million 18-23 year olds.
This number equals the population of several European countries put
together (Virmani, 2014). All these indicate that, though the study was based on a
student sample, it is applicable and also generalizable to a great extent in the
Indian context.
But, it is worth examining whether a significant difference in personality
perception based on demographic characteristics exists. Since a factor analytical
approach was employed for this study, some of the meaningful items might be
deleted since they were less related with the prominent product brand personality
153
dimensions or in other words there is a high chance that many meaningful items
were eliminated as they were less associated with any of the identified dimensions.
Another limitation of this scale is related to the context or the vibrant nature
of culture. To elaborate this further, according to Schiffman and Kanuk (2000), the
cultural dimensions may evolve over a period of time (as cited by Muniz and
Marchetti, 2012).This may be because of the changes that may occur in the values,
beliefs and customs of a society over a period of time resulting from various
factors that occur in the socio-cultural environment of the society. Any of these
alterations and developments may affect the way in which consumers perceive
brands. Though major validity checks which were usually done by other
researchers in this field were performed, some types of validity (like predictive
validity) were left out for further studies considering the time and budget
constraints.
5.8. Future research
The outcomes as well as limitations of this research throws open many
avenues for future researchers. This study is limited to product brands, hence it is
advisable to develop brand personality scales to measure categories like political
brands; for instance one can study whether a congruity in the personality of a voter
and a political brand has influence on determining the voters’ intention to vote for
the representative of the political brand. Similar scales and studies are
recommended in areas such as media brands, online stores, web sites and so on
wherein this scale may not applicable. Further studies are required to check the
cross cultural validity of the scale. Since most of the existing brand personality
scales do not have cross cultural validity. It is also worthwhile to check whether
the dimensions vary across different consumer groups. Though, Aaker (1997) has
checked it and proved to be true in the case of her scale, it may not be the case
with product BPS. Also with the introduction of more sophisticated tools like
Exploratory structural equation modeling, this can be done in a better manner.
154
Further studies may also be done to validate the scope of product brand
personality scale in other contexts. Though it is less likely that this scale may find
validity in the case of service brands, as it is developed relying solely on product
brands, it is worth examining the validity of this scale in the case of service
brands. Even though, Aaker’s (1997) scale was proved to be less applicable in the
case of some service brands such as destination brands, retail brands and so on;
this need not be the case with the PBPS scale. Since it is a new scale which is
developed relying on a new definition of brand personality, one can always check
the validity of this scale in other categories such as destination brands, sports
brands (teams/organizations) and so on. Cross industry validation of a brand
personality scale is not new; see for instance the works of Venable et al. (2005),
Smith et al. (2006) and Hosany et al. (2006).
Cross country validation may also be done in the case of the newly
developed product brand personality scale also. The research works which tried to
check the cross country validity of Aaker (1997) include Ferrandi et al. (2000),
Aaker et al. (2001), Supphellen and Grønhaug (2003) and Thomas and Sekar
(2008) .
Another fruitful avenue for further research is to identify the antecedents
and consequences of different product brand personality dimensions. For instance,
researches can be done to identify how a brand can position itself as a youthful
brand. Or, what kind of leverage a brand can acquire from its youthful image.
Further studies are required to identify the role of brand personality
dimensions in determining the consumer’s behaviour in terms of their loyalty,
satisfaction and so on.
It is also worthwhile, investigating the impact of various marketing
programs on different product brand personality dimensions. For example, one can
evaluate the impact of an ad campaign portraying boldness, aggressiveness and
155
adventure on increasing the score of a brand on the adventure dimension or how
the introduction of a particular celebrity endorser change the performance of the
brand in different personality dimensions. Similar studies may also be done in
identifying the impact of various activities related to the brand such as change in
logo, new positioning strategy, brand extension, Corporate Social Responsibility
activities and so on change the perceived personality of the brand.
Though items with negative connotations were included in the initial stage
of the study all of them were eliminated during the various item reduction stages.
Hence this scale may be less effective to capture the personality of those brands
that tries to position themselves with a negative type of personality. However, in
the Indian context, it is uncommon to find any such brands and hence the absence
of these negative traits or dimensions may not affect the scope or applicability of
this scale.
Checking the nomological validity of the tool is another important direction
for future research. The purpose of ensuring nomological validity is to show that,
the structural relationship among different dimensions is consistent with other
studies that have been measured with validated instruments and tested against a
variety of settings, time and persons (Straub, Boudreau and Gefen, 2004). As
suggested by Geuens et al. (2009) a study may be conducted to check the
nomological validity by selecting two groups of respondents who have different
value systems and then the researcher may check how each of the brand
personality dimensions differentially contributes to the attitude of these two group
members towards the brand (Geuens et al., 2009). Previous research works in this
field indicates that consumer values have a great impact on their behaviour and
they will prefer to own products that have a personality similar to that of their own
personality (For an elaborate discussion of the methodology see Geuens et al.,
2009).
156
Another fruitful avenue for future research could be to investigate how
various aspects such as goals, demography, personality and so on of a consumer
affect his / her attitude to different dimensions of the product brand personality
scale. In other words, one can examine whether a male has more affinity to the
adventure dimension, or a professional shows more affinity for competence
dimension and so on. Researchers in future may also examine how a new
advertising or marketing campaign influence the various dimensions of the
product brand personality scale. For instance, one can evaluate whether the
adoption of a youthful, sexy celebrity increases a brand’s score on the Youthful
dimension. It is also worth examining the impact of various corporate social
responsibility activities on the different dimensions of the brand personality scale.
Another research in future could be to examine how the change in
consumer perception regarding the celebrity endorser of the brand affects the
various personality dimensions of the brand. For instance, a brand that employs a
sports star or celebrity as an endorser may suffer if the performance of the
celebrity declines.
It is also worth examining how a scam, quality issues, and other related
aspects of a product affect various personality dimensions of the brand. Again, it
will also be interesting to check the vulnerability of different positioning
techniques that rely on the different dimensions of the product brand personality
scale. Or, which type of positioning will be more affected if some adverse events
occur (that is, which among the seven dimensions will be drastically affected by
the occurrence of an adverse event).Another domain for research is to check the
impact of various product brand personality dimensions in determining brand
choice.
157
5.9. Chapter conclusion
This chapter has summarised the important findings of the research,
implications for academics and practitioners and future research directions in this
domain.
158
REFERENCES
1. Aaker, D. A. (2009). Managing brand equity. Simon and Schuster.
2. Aaker, D. A. (2012). Building strong brands. Simon and Schuster.
3. Aaker, D. A., & Biel, A. L. (2013). Brand Equity & Advertising:
Advertising's role in building strong brands. New Jersey: Lawrence
Erlbaum Associates.
4. Aaker, D.A. & Joachimsthaler, E. (2000). Brand leadership. New York:
Free Press.
5. Aaker, J.L. (1995). Brand Personality: Conceptualization, measurement
and underlying psychological mechanisms. Ann Arbor: UMI.
6. Aaker, J. L. (1997). Dimensions of brand personality. Journal of Marketing
research, 34(3), 347-356.
7. Aaker, J. L., Benet-Martinez, V., & Garolera, J. (2001). Consumption
symbols as carriers of culture: A study of Japanese and Spanish brand
personality constructs. Journal of Personality and Social Psychology,
81(3), 492-508.
8. Aaker, J.L, & Fournier, S. (1995). A brand as a character, a partner and a
person: three perspectives on the question of brand personality. Advances in
consumer research, 22, 391-395.
9. Aiken, K. D., Campbell, R. M., & Koch, E. C. (2013). Exploring the
relationship between team (as brand) personality and geographic
personality: linking consumer perceptions of sports teams and
cities. International Journal of Sports Marketing & Sponsorship, 15(1).
10. Aiken, K. D., Koch, E. C., & Madrigal, R. (2000). What’s in a name?
Explorations in geographic equity and geographic personality. In 2000
AMA Winter Educators' Conference, 11, 301-308.
11. Alt, M., & Griggs, S. (1988). Can a brand be cheeky? Marketing
Intelligence & Planning, 6(4), 9-16.
159
12. Ambroise, L., Ferrandi, J. M., & Merunka, D. (2005). How well does brand
personality predict brand choice? A measurement scale and analysis using
binary regression models. Asia Pacific Advances in Consumer Research, 6,
30-38.
13. American Marketing Association. (1960). Marketing definitions: A
glossary of marketing terms. AMA, Chicago, IL.
14. Austin, J.R., Siguaw, J.A., & Mattila, A.S. (2003). A re-examination of the
generalizability of the Aaker brand personality measurement frame work.
Journal of Strategic Marketing, 11(June), 77-92.
15. Azoulay, A., & Kapferer, J.-N. (2003). Do brand personality scales really
measure brand personality? The Journal of Brand Management, 11(2), 143-
155
16. Bagozzi, R. P., & Heatherton, T. F. (1994). A general approach to
representing multifaceted personality constructs: Application to state
self‐esteem. Structural Equation Modeling: A Multidisciplinary
Journal, 1(1), 35-67.
17. Batra, R., Myers, J.G., & Aaker, D.A. (1996). Advertising Management (4th
Edn.). Upper Saddle River, New Jersey: Prentice-Hall.
18. Batra, R., Lehmann, D. R., & Singh, D. (1993). The brand personality
component of brand goodwill: some antecedents and consequences. Brand
equity and advertising, 83-96.
19. Beldona, S., & Wysong, S. (2007). Putting the “brand” back into store
brands: an exploratory examination of store brands and brand
personality. Journal of Product & Brand Management, 16(4), 226-235.
20. Belk, R. (1988). Possessions and Self: Wiley Online Library.
21. Beltramini, R. F. (1983). Student surrogates in consumer research. Journal of the Academy of Marketing Science, 11(4), 438-443.
160
22. Biel, A. L. (1992). How brand image drives brand equity. Journal of
Advertising Research, 32(6), 6-12.
23. Birdwell, A. E. (1968). A study of the influence of image congruence on
consumer choice. The Journal of Business, 41(1), 76-88.
24. Bosnjak, M., Bochmann, V., & Hufschmidt, T. (2007). Dimensions of
Brand Personality Atributions: A Person Centric Approach in the German
Cultural Context. Social Behavior & Personality: An International Journal,
35(3), 303-316.
25. Boudreaux, C. A., & Palmer, S. E. (2007). A charming little Cabernet:
Effects of wine label design on purchase intent and brand
personality. International Journal of Wine Business Research, 19(3), 170-
186.
26. Brengman, M., & Willems, K. (2009). Determinants of fashion store
personality: a consumer perspective. Journal of Product & Brand
Management, 18(5), 346-355.
27. Burisch, M. (1997). Test length and validity revisited. European Journal of
Personality, 11, 303-315.
28. Calder, B. J., Phillips, L. W., & Tybout, A. M. (1981). Designing research
for application. Journal of Consumer Research, 8(2), 197-207.
29. Caprara, G. V., Barbaranelli, C., & Guido, G. (2001). Brand personality:
how to make the metaphor fit? Journal of economic psychology, 22(3),
377-395.
30. Carlson, B. D., Donavan, D. T., & Cumiskey, K. J. (2009). Consumer-
brand relationships in sport: brand personality and identification.
International Journal of Retail & Distribution Management, 37(4), 370-
384.
31. Cervera-Taulet, A., Schlesinger, M. W., & Yagüe-Guillen, M. J. (2013).
Influence of advertising on brand personality in the airline sector: The case
of Spain. Journal of Travel & Tourism Marketing, 30(5), 445-454.
161
32. Churchill Jr, G. A. (1979). A paradigm for developing better measures of
marketing constructs. Journal of Marketing research, 16(1), 64-73.
33. Costa, P. T., & McCrae, R. R. (1992). Revised neo personality inventory
(neo pi-r) and neo five-factor inventory (neo-ffi), 101, Odessa, FL:
Psychological Assessment Resources.
34. D’Astous, A., & Lévesque, M. (2003). A scale for measuring store
personality. Psychology and Marketing, 20(5), 455-469.
35. Das, G. (2013a). Impacts of retail brand personality and self-congruity on
store loyalty: The moderating role of gender. Journal of Retailing and
Consumer Services, 21(2), 130-138.
36. Das, G. (2013b). Linkages of retailer personality, perceived quality and
purchase intention with retailer loyalty: A study of Indian non-food
retailing. Journal of Retailing and Consumer Services, In Press, Corrected
Proof, Available online 22 November 2013.
37. Das, G., Datta, B., & Guin, K. K. (2012). Impact of retailer personality on
consumer-based retailer equity: An empirical study of retail brands. Asia
Pacific Journal of Marketing and Logistics, 24(4), 619-639.
38. Das, G., Guin, K. K., & Datta, B. (2013). Impact of Store Personality
Antecedents on Store Personality Dimensions: An Empirical Study of
Department Retail Brands. Global Business Review, 14(3), 471-486.
39. D'Astous, A., & Boujbel, L. (2007). Positioning countries on personality
dimensions: Scale development and implications for country
marketing. Journal of Business Research, 60(3), 231-239.
40. Davis, Joel B. (1997). Advertising Research: Theory and Practice. Upper
Saddle River, New Jersey: Prentice-Hall.
41. De Pelsmacker, P., Geuens, M., & Van den Bergh, J. (2007). Marketing
communications: a European perspective. Pearson Education.
42. DeVellis, R. F. (2011). Scale development: Theory and applications, 26,
Sage Publications.
162
43. Diamantopoulos, A., Smith, G., & Grime, I. (2005). The impact of brand
extensions on brand personality: experimental evidence, European Journal
of Marketing, 39, 1-2.
44. Dickenson, N. (1996). Can celebrities ruin a launch, Campaign, 3, 12-24.
45. Dobni, D. & Zinkhan, G.M. (1990). In search of brand image: A foundation
analysis.In M.E. Goldberg, G. Gorn, & R.W. Pollay (Eds.), Advances in
Consumer Research, 17, 110-119. Provo, UT: Association for Consumer
Research.
46. Dolich, I. J. (1969). Congruence relationships between self images and
product brands. Journal of Marketing research, 6(1), 80-84.
47. Douglas, A., & Mills, J. E. (2006). Logging Brand Personality Online:
Website content analysis of Middle Eastern and North African
Destinations. In Information and communication technologies in tourism
2006, edited by M. Hitz, M. Sigala and J. Murphy, Vienna: Springer
Verlag, 345.
48. Durgee, J. F., & Stuart, R. W. (1987). Advertising symbols and brand
names that best represent key product meanings. Journal of Consumer
Marketing, 4(3), 15-24.
49. Edmunds, H. (1999). The focus group research handbook. Chicago: NTC
Business Books.
50. Ekinci, Y., & Hosany, S. (2006). Destination Personality: An application of
brand personality to tourism destinations. Journal of Travel Research,
45(2), 127-139.
51. Ekinci, Y., Sirakaya-Turk, E., & Baloglu, S. (2007). Host image and
destination personality. Tourism Analysis, 12(5-6), 433-446.
52. Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: a three-
factor theory of anthropomorphism. Psychological review, 114(4), 864.
163
53. ET Bureau. (2011, September 28). Brand Equity Brand List. Retrieved
February 23, 2012, from The Economic Times:
http://economictimes.indiatimes.com/topbrandlisting.cms
54. Fennis, B. M., & Pruyn, A. T. H. (2007). You are what you wear: Brand
personality influences on consumer impression formation. Journal of
Business Research, 60(6), 634-639.
55. Ferrandi, J. M., & Valette-Florence, P. (2002). First test and validate the
implementation of a human scale brands personality. Research and
Applications in Marketing, 17(3), 21-40.
56. Ferrandi, J. M., Valette-Florence, P., & Fine-Falcy, S. (2000). Aaker’s
brand personality scale in a French context: a replication and a preliminary
test of its validity. Developments in Marketing Science, 23, 7-13.
57. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models
with unobservable variables and measurement error. Journal of Marketing
Research, 18(1), 39-50.
58. Fortini-Campbell, L. (1992, January). Hitting the sweet spot. Copy
Workshop.
59. Freling, T. H., & Forbes, L. P. (2005). An empirical analysis of the brand
personality effect. Journal of Product & Brand Management, 14(7), 404-
413.
60. Friedmann, R. (1986). Psychological meaning of products: Identification
and marketing applications. Psychology & Marketing, 3(1), 1-15.
61. Geuens, a. M., Weijters, B., & Wulf, K. D. (2009). A new measure of brand
personality. International Journal of Research in Marketing, 26, 97-107.
62. Goldberg, L. R. (1993). The structure of phenotypic personality
traits. American psychologist, 48(1), 26-34.
63. Gounaris, S., & Dimitriadis, S. (2003). Assessing service quality on the
web: evidence from business-to-consumer portals. Journal of Services
Marketing,17(5), 529-548.
164
64. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Talthem, R. L.
(2012). Multivariate data analysis. New Delhi: Pearson.
65. Haire, M. (1950). Projective Techniques in Marketing Research. Journal of
Marketing, 14 (5), 649-656.
66. Heere, B. (2010). A new approach to measure perceived brand personality
associations among consumers. Sport Marketing Quarterly, 19 (1), 17-28.
67. Heine, K. (2009). Using personal and online repertory grid methods for the
development of a luxury brand personality. In 8th European Conference on
Research Methodology for Business and Management Studies: University
of Malta, Valletta, Malta, 22-23 June 2009:[proceedings], p.160.
Academic Conferences Limited.
68. Helgeson, J. G., & Supphellen, M. (2004). A conceptual and measurement
comparison of self-congruity and brand personality-The impact of socially
desirable responding. International Journal of Market Research, 46(2),
205-233.
69. Henderson, J. C. (2000). Selling places: the new Asia-Singapore brand. The
Journal of Tourism Studies, 11(1), 36-44.
70. Herbst, U., & Merz, M. A. (2011). The industrial brand personality scale:
Building strong business-to-business brands. Industrial Marketing
Management, 40(7), 1072-1081.
71. Hieronimus. F. (2003). Personality-oriented brand management: an
empirical study to measure, perception and impact of brand personality,
Lang.
72. Hosany, S., Ekinci, Y., & Uysal, M. (2006). Destination image and
destination personality: An application of branding theories to tourism
places. Journal of Business Research, 59(5), 638-642.
73. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in
covariance structure analysis: Conventional criteria versus new
165
alternatives. Structural Equation Modeling: A Multidisciplinary
Journal, 6(1), 1-55.
74. Huang, H. H., & Mitchell, V. W. (2014). The role of imagination and brand
personification in brand relationships. Psychology & Marketing, 31(1), 38-
47.
75. Hussey, M., & Duncombe, N. (1999). Projecting the right image: using
projective techniques to measure brand image. Qualitative Market
Research: An International Journal, 2(1), 22-30.
76. John, OP. (1990). The "Big Five" factor taxonomy: Dimensions of
personality in the natural language and in questionnaires. In LA Pervin
(Ed.), Handbook of personality: Theory and research. New York: Guilford.
77. Kang, C. (2013). Dimensions of Brand Personality: A New Measure of
Brand Personality in Sport (Doctoral dissertation, Doctoral dissertation,
Texas A & M University. Available electronically from http://hdl.handle.
net/1969. 1/151299).
78. Kapferer, J. N. (2012). The new strategic brand management: Advanced
insights and strategic thinking. 18, Kogan page publishers.
79. Kaplan, M. D., Yurt, O., Guneri, B., & Kurtulus, K. (2010). Branding
places: applying brand personality concept to cities. European Journal of
Marketing, 44(9/10), 1286-1304.
80. Karande, K., Zinkhan, G. M., & Lum, A. B. (1997). Brand personality and
self concept: a replication and extension. Paper presented at the American
Marketing Association, Summer Conference.
81. Kassarjian, H. H. (1971). Personality and Consumer Behaviour: A Review.
Journal of Consumer Research, 14, 409-418.
82. Keller, K. L. (1993). Conceptualizing, measuring, and managing customer-
based brand equity. The Journal of Marketing, 1-22.
166
83. Keller, K. L., & Richey, K. (2006). The importance of corporate brand
personality traits to a successful 21st century business. Journal of Brand
Management, 14(1), 74-81.
84. Keller, K. L., Parameswaran, M., & Jacob, I. (2011). Strategic brand
management: Building, measuring, and managing brand equity: Pearson
Education India.
85. Kelly, G. A. (1955). Psychology of personal constructs 2: clinical diagnosis
and psychotherapy.
86. Pike K.L. (1967). Language in relation to a unified theory of the structure
of human behavior (2nd Edn), 24(24), The Hague: Mouton and Co
87. Kenny, D. A., & McCoach, D. B. (2003). Effect of the number of variables
on measures of fit in structural equation modeling. Structural Equation
Modeling, 10(3), 333-351.
88. Kim, C. K., Han, D., & Park, S. B. (2001). The effect of brand personality
and brand identification on brand loyalty: Applying the theory of social
identification. Japanese Psychological Research, 43(4), 195-206.
89. Kim, D., Magnini, V. P., & Singal, M. (2011). The effects of customers’
perceptions of brand personality in casual theme restaurants. International
Journal of Hospitality Management, 30(2), 448-458.
90. Kim, J., Baek, T. H., & Martin, H. J. (2010). Dimensions of news media
brand personality. Journalism & Mass Communication Quarterly, 87(1),
117-134.
91. Kim, S., & Lehto, X. Y. (2013). Projected and Perceived Destination Brand
Personalities The Case of South Korea. Journal of Travel Research, 52(1),
117-130.
92. King, S. (1970). What is a Brand? London: J.Walter Thompson Company
Limited.
93. Kline, R. B. (2011). Principles and practice of structural equation
modeling. Guilford press.
167
94. Koebel, M. N., & Ladwein, R. (1999). The scale of brand personality of
Jennifer Aaker: Adaptation to French context. Marketing decisions, 81-88.
95. Kuenzel, S., Phairor, K.H. (2009). A Short Scale for Measuring Brand
Personality. Proceedings of Australian and New Zealand Marketing
Academy conference, 1-9.
96. Ladhari, R. (2010). Developing e-service quality scales: A literature
review.Journal of Retailing and Consumer Services, 17(6), 464-477.
97. Lee, J. S., & Back, K. J. (2010). Examining antecedents and consequences
of brand personality in the upper-upscale business hotel segment. Journal
of Travel & Tourism Marketing, 27(2), 132-145.
98. Lee, Y. K., Back, K. J., & Kim, J. Y. (2009). Family restaurant brand
personality and its impact on customer's emotion, satisfaction, and brand
loyalty. Journal of Hospitality & Tourism Research, 33(3), 305-328.
99. Lee,J., Soutar, G., Quintal, V. (2010). Destination personality: cross-
country comparisons. Proceedings of Australian and New Zealand
Marketing Academy conference, edited by David Fortin, Lucie Ozanne,
New Zealand.
100. Leonard, E., & Katsanis, L. P. (2013). The dimensions of prescription
drug brand personality as identified by consumers. Journal of Consumer
Marketing, 30(7), 583-596.
101. Levy, Sidney J. (1959). Symbols for Sales. Harvard Business Review, 37
(4), 117-24.
102. Li, X. R., & Kaplanidou, K. K. (2013). The Impact of the 2008 Beijing
Olympic Games on China’s Destination Brand A US-Based
Examination. Journal of Hospitality & Tourism Research, 37(2), 237-261.
103. Louis, D., & Lombart, C. (2010). Impact of brand personality on three
major relational consequences (trust, attachment, and commitment to the
brand).Journal of Product & Brand Management, 19(2), 114-130.
168
104. Maciel, F., da Rocha, A., & da Silva, J. (2013). Brand Personality of
Global Quick-Service Restaurants in Emerging and Developed Markets: A
Comparative Study in Brazil and the U.S. Latin American Business Review,
14(2), 139-161.
105. Magnini, V. P., & Thelen, S. T. (2008). The influence of music on
perceptions of brand personality, décor, and service quality: The case of
classical music in a fine-dining restaurant. Journal of Hospitality & Leisure
Marketing, 16(3), 286-300.
106. Malhotra, N. K. (1981). A scale to measure self-concepts, person
concepts, and product concepts. Journal of Marketing Research, 18(4).
107. Marsh, H. W., & Hocevar, D. (1985). Application of confirmatory factor
analysis to the study of self-concept: First-and higher order factor models
and their invariance across groups. Psychological bulletin, 97(3), 562.
108. Martineau, P. (1958). The personality of the retail store. Harvard business
review, 36(1), 47-55.
109. McCracken, G. (1989). Who Is the Celebrity Endorser? Cultural
Foundations of the Endorsement Process. Journal of Consumer
Research, 16(3), 310-321.
110. McCrae, R. R., & Costa, P. T. (1989). The structure of interpersonal
traits: Wiggins's circumplex and the five-factor model. Journal of
personality and social psychology, 56(4), 586-595.
111. Mervielde, I. (1992). The B5BBS-25: A Flemish set of bipolar markers
for the" Big-Five" personality factors. Psychologica Belgica, 32: 195-210.
112. Moorthi, Y. (2003). Brand Management: The Indian Context . New
Delhi: Vikas Publishing House PVT LTD.
113. Mowen, J. C., & Voss, K. E. (2008). On building better construct
measures: Implications of a general hierarchical model. Psychology and
Marketing, 25(6), 485-505.
169
114. Müller, B., & Chandon, J. L. (2003). The impact of visiting a brand
website on brand personality. Electronic Markets, 13(3), 210-221.
115. Muniz, K. M., & Marchetti, R. Z. (2012). Brand personality dimensions
in the Brazilian context. BAR. Brazilian Administration Review, 9(2), 168-
188.
116. Murase, H., & Bojanic, D. (2004). An examination of the differences in
restaurant brand personality across cultures. Journal of Hospitality &
Leisure Marketing, 11(2-3), 97-113.
117. Murphy, L., Benckendorff, P., & Moscardo, G. (2007a). Destination
brand personality: visitor perceptions of a regional tourism destination.
Tourism Analysis, 12, 419-432.
118. Murphy, L., Benckendorff, P., & Moscardo, G. (2007b). Linking travel
motivation, tourist self-image and destination brand personality. Journal of
Travel & Tourism Marketing, 22(2), 45-59.
119. Murphy, L., Moscardo, G., & Benckendorff, P. (2007). Using brand
personality to differentiate regional tourism destinations. Journal of travel
research, 46(1), 5-14.
120. Musante, M. D., Bojanic, D. C., & Zhang, J. (2008). A modified brand
personality scale for the restaurant industry. Journal of Hospitality &
Leisure Marketing, 16(4), 303-323.
121. Norman, Warren T. (1963). Toward an Adequate Taxonomy of
Personality Attribute: Replicated Factor Structure in Peer Nomination
Personality Ratings, Journal of Abnormal and Social Psychology, 66, 574-
583.
122. Nunnally, J. C. (2010). Psychometric Theory 3E: Tata McGraw-Hill
Education.
123. Okazaki, S. (2006). Excitement or sophistication? A preliminary
exploration of online brand personality. International Marketing
Review, 23(3), 279-303.
170
124. Opoku, R. A. (2005). Communication of brand personality by some top
business schools online. Licentiate Thesis, Lulea University of Technology,
ISSN, 1402-1757.
125. Opoku, R. A., Abratt, R., Bendixen, M., & Pitt, L. (2007).
Communicating brand personality: are the web sites doing the talking for
food SMEs? Qualitative Market Research: An International Journal , 362 -
374.
126. Park, C. W., Jaworski, B. J., & Maclnnis, D. J. (1986). Strategic brand
concept-image management. The Journal of Marketing, 135-145.
127. Peter, J. P. (1981). Construct validity: A review of basic issues and
marketing practices. Journal of Marketing Research, 18(2), 133-145.
128. Phau, I., & Lau, K. C. (2000). Conceptualizing brand personality: a
review and research propositions. Journal of Targeting, Measurement and
Analysis for Marketing, 9(1), 52-69.
129. Piedmont, R. L., McCrae, R. R., & Costa, P. T. (1991). Adjective Check
List scales and the five-factor model. Journal of Personality and Social
Psychology, 60(4), 630-637.
130. Pitt, L. F., Opoku, R., Hultman, M., Abratt, R., & Spyropoulou, S. (2007).
What I say about myself: Communication of brand personality by African
countries.Tourism Management, 28(3), 835-844.
131. Plummer, J. T. (1984, December/January). How personality makes a
difference. Journal of Advertising Research, 24, 27-31.
132. Plummer, J. T. (2000). How personality makes a difference. Journal of
Advertising Research, 40(06), 79-84.
133. Poddar, A., Donthu, N., & Wei, Y. (2009). Web site customer
orientations, Web site quality, and purchase intentions: The role of Web
site personality. Journal of Business Research, 62(4), 441-450.
171
134. Prayag, G. (2007). Exploring the Relationship between Destination Image
& Brand Personality of a Tourist Destination-An Application of Projective
Techniques. Journal of Travel & Tourism Research, 7(2), 22-30.
135. Purkayastha, S. (2009). Brand Personality: An Empirical Study of Four
Brands in India. The Icfaian Journal of Management Research, 8(4), 7-20.
136. Ramaseshan, B., & Tsao, H. Y. (2007). Moderating effects of the brand
concept on the relationship between brand personality and perceived
quality. Journal of Brand Management, 14(6), 458-466.
137. Rammstedt, B., and John, O. P. (2007). Measuring personality in one
minute or less: A 10- item short version of the Big Five Inventory in
English and German. Journal of Research in Personality, 41, 203-212.
138. Ranjbar, S. (2010). Application of Brand Personality Scale in Automobile
Industry- the study of SAMAND's brand personality dimensions. LUlea
University of Technology.
139. Reilly, M. D. (1990). Free elicitation of descriptive adjectives for tourism
image assessment. Journal of Travel Research, 28(4), 21-26.
140. Roy, S., & Moorthi, Y. L. R. (2012). Investigating endorser personality
effects on brand personality: Causation and reverse causation in
India. Journal of Brand Strategy, 1(2), 164-179.
141. Sahin, S., & Baloglu, S. (2009, January). Brand personality and
destination image of Istanbul: A comparison across nationalities. Paper
presented at the 14th annual graduate student research conference in
hospitality and tourism, Las Vegas, NV.
142. Santos, C. A. (2004). Framing Portugal: representational
dynamics. Annals of Tourism Research, 31(1), 122-138.
143. Saucier, G. (1994). Mini-markers: A brief version of Goldberg's unipolar
Big-Five markers. Journal of personality assessment, 63(3), 506-516.
172
144. Saxe, R., & Weitz, B. A. (1982). The SOCO scale: a measure of the
customer orientation of salespeople. Journal of marketing research, 343-
351.
145. Schiffman, L. G, & Kanuk L. L. (2000). Consumer behavior. Rio de
Janeiro: LTC.
146. Segars, A. H., & Grover, V. (1993). Re-examining perceived ease of use
and usefulness. MIS quarterly, 17(4), 517-525.
147. Sirgy, M. J. (1982). Self-concept in consumer behaviour: a critical
review. Journal of consumer research, 287-300.
148. Sirgy, M. J., Grewal, D., Mangleburg, T. F., Park, J. O., Chon, K. S.,
Claiborne, C. B. & Berkman, H. (1997). Assessing the predictive validity
of two methods of measuring self-image congruence. Journal of the
academy of marketing science, 25(3), 229-241.
149. Smith, A. C., Graetz, B. R., & Westerbeek, H. M. (2006). Brand
personality in a membership‐based organisation. International Journal of
Nonprofit and Voluntary Sector Marketing, 11(3), 251-266.
150. Stein, D. (2004). Testing the reliability and validity of a brand personality
measurement tool. Florida: University of Florida.
151. Stokburger-Sauer, N. E. (2011). The relevance of visitors’ nation brand
embeddedness and personality congruence for nation brand identification,
visit intentions and advocacy. Tourism Management, 32(6), 1282-1289.
152. Straub, D., Boudreau, M. C., & Gefen, D. (2004). Validation Guidelines
For Is Positivist Research. Communications of the Association for
Information systems, 13.
153. Strausbaugh, K. L. (1998). " Miss Congeniality" Or" No More Mr. Nice
Guy?": On a Method for Assessing Brand Personality and Building Brand
Personality Profiles (Doctoral dissertation, University of Florida).
173
154. Sung, Y, & Tinkham, S. F. (2005). Brand personality structures in the
United States and Korea: Common and culture-specific factors. Journal of
Consumer Psychology, 15(4), 334-350.
155. Supphellen, M., & Gronhaug, K. (2003). Building foreign brand
personalities in Russia: the moderating effect of consumer
ethnocentrism. International Journal of Advertising, 22(2), 203-226.
156. Sureshchandar, G. S., Rajendran, C., & Anantharaman, R. N. (2002).
Determinants of customer-perceived service quality: a confirmatory factor
analysis approach. Journal of services Marketing, 16(1), 9-34.
157. Sweeney, J. C., & Brandon, C. (2006). Brand personality: Exploring the
potential to move from factor analytical to circumplex models. Psychology
& Marketing, 23(8), 639-663.
158. Tabachnick, B. G. & Fidell, L. S. (2007). Using multivariate statistics.
New York: Allyn and Bacon.
159. Thomas, B. J., & Sekar, P. C. (2008). Measurement and Validity of
Jennifer Aaker's Brand Personality Scale for Colgate Brand. Vikalpa: The
Journal for Decision Makers, 33(3), 49-61.
160. Trapnell, P. D., & Wiggins, J. S. (1990). Extension of the Interpersonal
Adjective Scales to include the Big Five dimensions of personality. Journal
of Personality and Social Psychology, 59(4), 781-790.
161. Trust Advisory. (2012, January). All India Brand Trust Ranking 2012
(Top 1000 Brands). Retrieved February 22, 2012, from Trust Advisory:
http://trustadvisory.info/allindia_1000.html
162. Tupes, E. C., & Christal, R. C. (1958). Stability of personality trait rating
factors obtained under diverse conditions (No. Wadc-Tn-58-61). Wright
Air Development Center Wright-Patterson Afb Oh.
163. Upadhyaya, M. (2012). Influence of Destination Image and Destination
Personality: An Empirical Analysis. Journal of Marketing &
Communication, 7(3).
174
164. Usakli, A., & Baloglu, S. (2011). Brand personality of tourist
destinations: An application of self-congruity theory. Tourism
Management, 32(1), 114-127.
165. Usunier, J. C. (1998). International and cross-cultural management
research. Sage.
166. Valette-Florence, R., & de Barnier, V. (2011). Do Print Media Brand
Exhibit Specific Personnality Traits? A Micro Perspective in a French
Setting. In The 38th International Research Conference in Marketing.
167. Valette-Florence, R., & De Barnier, V. (2013). Towards a micro
conception of brand personality: An application for print media brands in a
French context. Journal of Business Research, 66(7), 897-903.
168. Van Rekom, J., Jacobs, G., &Verlegh, P. W. J. (2006, July). Measuring
and managing the essence of a brand personality. Marketing Letters, 17,
181-192.
169. Venable, B. T., Rose, G. M., & Gilbert, F. W. (2003). Measuring the
brand personality of nonprofit organizations. Advances in Consumer
Research, 30, 379-380.
170. Venable, B. T., Rose, G. M., Bush, V. D., & Gilbert, F. W. (2005). The
role of brand personality in charitable giving: an assessment and validation.
Journal of the Academy of Marketing Science, 33(3), 295-312.
171. Virmani, P. (2014, April 08). India elections 2014. Retrieved April 08,
2014, from the guardian:
http://www.theguardian.com/commentisfree/2014/apr/08/india-leaders-
young-people-change-2014-elections
172. Wee, T. T. T. (2004). Extending human personality to brands: the
stability factor. The Journal of Brand Management, 11(4), 317-330.
173. Wells, W. D., Andriuli, F. J., Goi, F. J., & Seader, S. (1957). An adjective
check list for the study of" product personality." Journal of Applied
Psychology, 41(5), 317-325.
175
174. Wheaton, D. E. (1977). Assessing reliability and stability in panel models.
In D.R. Heise (Ed.), Sociological Methodology, 84-136.
175. Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptive
terms: The interpersonal domain. Journal of Personality and Social
Psychology, 37(3), 395-412.
176. Willems, K., Swinnen, G., Janssens, W., & Brengman, M. (2011).
Fashion store personality: scale development and relation to self-congruity
theory. Journal of Global Fashion Marketing, 2(2), 55-65.
177. Wylie, R. C. (1979). The self-concept: Theory and research on selected
topics, 2. University of Nebraska Press.
178. Xie, K. L., & Lee, J. S. (2013). Toward the perspective of cognitive
destination image and destination personality: The case of Beijing. Journal
of Travel & Tourism Marketing, 30(6), 538-556.
179. Zentes, J., Morschett, D., & Schramm-Klein, H. (2008). Brand
personality of retailers–an analysis of its applicability and its effect on store
loyalty. The International Review of Retail Distribution and Consumer
Research, 18(2), 167-184.
176
APPENDIX I: Free elicitation questionnaire
Hi,
We are conducting a research on developing a scale to measure brand personality, whichrefers to the ‘set of human personality traits that are both applicable to and relevant forbrands’. We would like you to describe the brand personality (in terms oftraits/adjectives that you feel appropriate or representative) of two brands each, from the3 product categories given below.
1) Symbolic products (e.g. Jeans, Cosmetics andPerfume/Fragrance)2) Utilitarian products (e.g. Laptops, Electronics, Appliances and Personal care
products)3) Symbolic and Utilitarian products (like Bikes, Smartphones and Athletic Shoes)
Write the names of 2 brands that first come to your mind when you think about theseproduct categories and describe each brand with as many traits/adjectives as possible.
Thanks for your valuable participation.Yours truly,Jijo George (Doctoral research scholar)Dr.S.Victor Anandkumar (Research Supervisor)Department of Management StudiesPondicherry University, Puducherry – 605 014.
Gender: Male / FemaleHome State:
ProductCategory
Brand name Descriptors
Symbolic
Utilitarian
Symbolic andUtilitarian
APPENDIX II: Language Expert Opinion Survey
Dear Sir/ Ma’am,
We are conducting a research on developing a scale to measure brand personality in theIndian context. For this purpose a list of adjectives has been created based on theprevious researches done in this field in different countries. We acknowledge yourexpertise in the field of English language and request for your expert opinion on theappropriateness of these words in conveying meaning in an Indian context. If you feelthat a particular word is not common in India, suggest an alternative word which willconvey the same meaning as the original one.
Thanks for your valuable participation.Yours truly,Jijo George (Doctoral research scholar)Dr.S.Victor Anandkumar (Research Supervisor)Department of Management StudiesPondicherry University, Puducherry – 605 014.
Sl. No. Items Alternative word
1 Accepting2 Accessible3 Accommodating4 Active5 Adult6 Adventurous7 Affectionate8 Aggressive9 Agreeable10 Alive11 Aloof12 Altruist13 Amusing14 Approachable15 Aristocratic16 Arrogant17 Artificial18 Assertive19 Attractive20 Authentic
21 Balanced22 Beneficial23 Big24 Bigheaded25 Bohemian
26 Boisterous27 Bold28 Bon-Vivant29 Boring30 Bubbly31 Busy
32 Calculating33 Calm34 Careless35 Casual36 Ceremonious37 Charismatic38 Charming39 Chatty40 Chauvinist41 Cheerful42 Childlike43 Clean44 Clear45 Clumsy46 Cocky47 Comfortable48 Comforting49 Comical50 Competent Person51 Competitive52 Confidence53 Confident54 Connoisseur55 Conscientious56 Conservative57 Considerate58 Consistent59 Conspicuous60 Constant61 Contemporary62 Cool63 Cooperative64 Cordial65 Corporate66 Countrified67 Courageous68 Cowardly69 Crazy70 Creative
71 Cruel72 Cunning73 Curious74 Cute
75 Daring76 Decadent77 Decorated78 Delicate79 Dependable80 Dependant81 Determined82 Different83 Diffident84 Dignified85 Diligent86 Discreet87 Discrete88 Disrespectful89 Docile90 Dominant91 Donnish92 Down-To-Earth93 Dubious94 Dynamic
95 Easy96 Easygoing97 Efficient98 Egocentric99 Elegant100 Emotional101 Energetic102 Engaging103 Enthusiastic104 Environmentally-Friendly105 Exciting106 Exotic107 Expensive108 Exploitative109 Expressive110 Extravagant
111 Faithful112 Familiar113 Family-Oriented
114 Famous115 Fanciful116 Fast117 Feminine118 Fervent119 Feisty120 Fighter121 Firm122 Flaunty123 Flexible124 Flourishing125 Forceless126 Forgiving127 Formal128 Frank129 Freaky130 Free131 Fresh132 Friendly133 Frugal134 Fun135 Funky136 Funny
137 Gay138 Generous139 Genial140 Gentle141 Genuine142 Glamorous143 Golden144 Good-Looking145 Good-Natured
146 Handy147 Happy148 Hard To Work149 Hard-Working150 Haughty151 Healthy152 Heavy153 Honest154 Hopeful155 Humorous
156 Ill-Mannered
157 Imaginative158 Immoral159 Impersonal160 Impolite161 Important162 Impractical163 Impulsive164 Independent165 Industrious166 Inefficient167 Informative168 Informed169 Innovating170 Innovative171 Intelligent172 Intense
173 Joyful
174 Kind175 Kitschy
176 Laid-Back177 Lazy178 Leader179 Leading180 Level-headed181 Light-hearted182 Likeable183 Lively184 Logical185 Logo oriented186 Loyal
187 Mannered188 Masculine189 Mild-mannered190 Minimalist191 Moderate192 Modern193 Modest194 Mysterious195 Mystical
196 Naïve197 Naked
198 Narrow-Minded199 Natural200 Neat201 Neutral202 New203 Nice204 Normal
205 Offender206 Offish207 Old208 Old-Fashioned209 Openhearted210 Open-Minded211 Optimistic212 Orderly213 Ordinary214 Organized215 Original216 Ostentatious217 Outdoorsy218 Outgoing219 Over forward220 Overstated
221 Passionate222 Patient223 Peaceful224 Persistent225 Playful226 Pleasant227 Poised228 Popular229 Positive230 Practical231 Precise232 Prestigious233 Pretenseless234 Pretentious235 Productive236 Professional237 Progressive238 Proud239 Provocative240 Prudent241 Puristic
242 Rakish243 Rational244 Real245 Realistic246 Recent247 Reflective248 Refreshing249 Regular250 Relaxed251 Reliable252 Religious253 Reserved254 Resolute255 Respectable256 Responsible257 Reveler258 Rich259 Rigorous260 Robust261 Romantic262 Rugged263 Ruthless
264 Satisfying265 Scrupulous266 Secure267 Self-contained268 Self-composed269 Self-effacing270 Sensual271 Sentimental272 Serene273 Serious274 Sexy275 Sharp276 Shrewd277 Showy278 Shrill279 Shy280 Silent281 Simple282 Sincere283 Small-town284 Smooth285 Snobbish
286 Sociable287 Solid288 Sophisticated289 Spirited290 Spiritual291 Sporty292 Stable293 Steady294 Strict295 Strong296 Stunning297 Stylish298 Successful299 Superficial300 Sweet301 Sympathetic
302 Talkative303 Technical304 Temperamental305 Tenacious306 Tense307 Thoughtful308 Tidy309 Timid310 Tolerant311 Tough312 Traditional313 Traditionalist314 Tranquil315 Trendy316 Tricky317 True318 Trustworthy319 Typical
320 Unaggressive321 Unargumentative322 Uncheery323 Uncomplex324 Unconceited325 Uncordial326 Understated327 Undevious328 Unique329 Unknown
330 Unobtrusive331 Unrefined332 Unreliable333 Unremarkable334 Unrevealing335 Unselfconscious336 Unsly337 Untidy338 Unvain339 Upper-class340 Up-to-date341 Urban
342 Versatile343 Vibrant344 Violent345 Voluptuous346 Vulgar
347 Warm348 Warmthless349 Well-made380 Well-mannered351 Western352 Wholesome353 Wimpy
354 Young355 Youthful
APPENDIX III: Expert Opinion Survey
Dear Sir/ Ma’am,
We are conducting a research on developing a scale to measure brand personality in theIndian context. We acknowledge your expertise in this field and request for your expertopinion in this survey. For the purpose of this research we rely on the definition of brandpersonality as the ‘set of human personality traits that are both applicable to andrelevant for brands’. Considering this definition, we would appreciate your freeelicitation of brand personality traits and critical assessment of appropriateness ofcommonly considered traits.
For your convenience we have split this survey in to three parts, each part may takeapproximately 15 to 20 minutes.
Thanks for your valuable participation.
Yours truly,
Jijo George (Doctoral research scholar)
Dr.S.Victor Anandkumar (Research Supervisor)
Department of Management Studies
Pondicherry University, Puducherry – 605 014.
Part-1: Think of a brand from symbolic products such as Jeans, Cosmetics andPerfume/Fragrance; that you admire as a branding/marketing expert. Freely elicitpersonality traits that describe this brand very appropriately, in your opinion.
Brand Name (Symbolic Brand):
Traits:
Part-2: Here is a comprehensive compilation of Brand personality traits as identifiedfrom the existing research literature and using a qualitative study. Rate each of the itemsgiven below on a 5 point scale (1: Very Inappropriate, 2: Somewhat inappropriate, 3:Neutral, 4: Somewhat appropriate, 5: Very Appropriate) according to its appropriatenessto describe the symbolic brand you have mentioned above.
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
AcceptingAbnormalAccessibleAccountableActiveAdultAdventurousAestheticAffectionateAggressive
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
AgreeableAliveAmusingAppealingApproachableAristocraticArrogantArtificialAssertiveAttentionSeeker
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
AttractiveAuthenticAwesomeBalancedBeautifulBelievableBeneficialBestBig
BoldVeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
BrightBusyCalmCapabilityCarefreeCarefulCarelessCaringCasualCatchy
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
CeremoniousCharismaticCharitableCharmingChattyChauvinistCheapCheerfulChicChildlike
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
Class ApartClassicClassyCleanClearClumsyColdColorfulComfortableComforting
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
ComicalCompatibleCompetentPersonCompetitiveConfident
ConservativeConsiderate
ConsistentConstantContemporary
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
CoolCooperativeCordialCorporateCosyCourageousCowardlyCrazyCreativeCunning
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
CuriousCuteDaringDashingDecoratedDelicateDependableDependantDetachedDetermined
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
DifferentDignifiedDisrespectfulDistinctDominantDown-To-EarthDullDynamicEasyEasygoing
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
Eco Friendly
EconomicalEffectiveEfficientElegantEmotionalEnergeticEngagingEnthusiasticEnviable
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
EthicalExcitementExcitingExoticExpensiveExpressiveExtravagantEye CatchingFaithfulFamiliar
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
Family-OrientedFamousFancifulFashionableFastFeminineFighterFirmFlawlessFlexible
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
FlourishingForcelessForgivingFormalFrankFreeFreshFriendlyFun
FunnyVeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
GallantGenerousGentleGenuineGlamorousGlobalGoldenGood-LookingGood-NaturedHandsome
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
HandyHappinessHappyHardHard-WorkingHarmlessHealthyHelpfulHonestHopeful
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
HotHumorousIconicImaginativeImmoralImportantImpracticalImpressiveIndependentIndian
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
IndustriousInefficientInformative
InformedInnocentInnovativeIntelligentIntenseInternationalJoyful
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
KillerKindLawbreakerLeaderLeadingLevel-HeadedLight-HeartedLikeableLivelyLogical
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
LovableLovelyLoyalLuxuriousMachoMagneticMagnificentManageableManipulativeManly
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
ManneredMasculineMerry MakerMiddle ClassMildModerateModernModestMultifacetedMysterious
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
NaturalNeatNeutralNewNiceNominalNormalObligingOffensiveOld
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
Old-FashionedOpenheartedOpen-MindedOptimisticOrderlyOrdinaryOrganizedOriginalOutgoingPassion
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
PassionatePatientPeacefulPersistentPlayfulPleasantPleasingPopularPoshPositive
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
PowerfulPracticalPrecisePrestigiousPrideProductiveProfessionalProgressiveProud
PureVeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
QualityQuickRationalRealRealisticRecentReflectiveRefreshingRegularRelaxed
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
ReliableReligiousReservedRespectableResponsibleRetiringRichRichnessRobustRomantic
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
RoughRoyalRudeRuggedRuralRuthlessSatisfyingScholarlySecureSelf-Composed
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
Self-ContainedSensualSentimentalSereneSerious
SexySharpShowyShySilent
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
SimpleSincereSleekSlowSmartSmoothSoberSociableSolidSoothing
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
SophisticatedSpecialistSpeedySpiritualSportySteadyStrongStunningSturdyStylish
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
SuccessfulSuperiorSweetSympatheticTechnicalThoughtfulTidyTimidTolerantTop Class
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
TouchyTough
TraditionalTraditionalistTrendyTrust WorthyTrustableTypicalUltimateUnConventional
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
UnfriendlyUnimpressiveUniqueRareUnreliableUnrevealingUntidyUpper-ClassUp-To-DateUrban
VeryInappropriate
SomewhatInappropriate
Not Sure SomewhatAppropriate
VeryAppropriate
VersatileVibrantVoluptuousWarmWell-MadeWell-ManneredWesternWholesomeYoungYouthfulTrue
Part-3: Think of a brand from utilitarian products such as Laptops, Electronics,Appliances and Personal care products and freely elicit personality traits that describethis brand very appropriately, in your opinion.
Brand Name (Utilitarian Brand):
Traits:
Part-4: Now, rate each of the items given below on a 5 point scale (1: VeryInappropriate, 2: Somewhat inappropriate, 3: Neutral, 4: Somewhat appropriate,5: Very Appropriate) according to its appropriateness to describe the utilitarianbrand you have mentioned above.
(The same set of items used in part 2 were used here)
Part-5: Think of a brand from Symbolic and Utilitarian products such as Bikes,Smartphones and Athletic Shoes ; and freely elicit personality traits that describethis brand very appropriately, in your opinion.
Brand Name (Symbolic and Utilitarian Brand):
Traits:
Part-6: Now, rate each of the items given below on a 5 point scale (1: VeryInappropriate, 2: Somewhat inappropriate, 3: Neutral, 4: Somewhat appropriate,5: Very Appropriate) according to its appropriateness to describe the symbolicand utilitarian brand you have mentioned above.
(The same set of items used in part 2 were used here)
Personal Details
Name:
Organization:
APPENDIX IV: Questionnaire for consumer survey
Dear Sir/ Ma’am,
I am Jijo George, a full time research scholar in Pondicherry University. As part of mydoctoral research work, I am conducting a study to understand how people perceivedifferent brands. To elaborate, if I ask you to give me your impression about a particularperson, you may describe him as a smart, sincere and competent person. Similarly thinkof a brand as if it were a person. For instance you may think Raymond to be arespectable, responsible, and a caring brand or ThumsUp to be a daring, spirited andadventurous brand. The purpose of this study is to understand which of the adjectivesgiven below are most appropriate to describe a brand. I request you to spare a fewminutes of your valuable time to take part in this study.
Group 1 Group 2 Group 3LEVI'S PONDS AXE
MRF LG ASIANPAINTS
GARNIER
PEPSI RAYBAN
AMUL HEROMOTORCORP.
ADIDAS
DABUR LUX BRITANNIA
DELL CADBURY’S
NOKIAMOBILE
COLGATE
TITAN GODREJ
1) Select a brand which you have used from the brands given in Group 1
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
SlNo.
Items Ratings SlNo.
Items Ratings SlNo.
Items Ratings
1 Accepting 34 Dominant 67 Professional2 Accessible 35 Dynamic 68 Progressive3 Accountable 36 Efficient 69 Proud4 Active 37 Energetic 70 Quick5 Adventurous 38 Engaging 71 Rational6 Affectionate 39 Enthusiastic 72 Realistic7 Aggressive 40 Exciting 73 Reliable8 Alive 41 Expressive 74 Respectable9 Appealing 42 Faithful 75 Responsible10 Approachable 43 Friendly 76 Romantic11 Authentic 44 Fun 77 Sexy12 Balanced 45 Genuine 78 Sharp13 Believable 46 Good-
Natured79 Sincere
14 Beneficial 47 Happiness 80 Smart15 Bold 48 Honest 81 Sophisticated16 Capable 49 Independent 82 Sporty17 Charismatic 50 Innovative 83 Strong18 Charming 51 Intelligent 84 Stylish19 Cheerful 52 Joyful 85 Successful20 Clear 53 Leading 86 Superior21 Comfortable 54 Lively 87 Tough22 Compatible 55 Loyal 88 Trendy23 Competent 56 Modern 89 True24 Competitive 57 Optimistic 90 Trust Worthy25 Confident 58 Outgoing 91 Unique26 Consistent 59 Passionate 92 Up-To-Date27 Courageous 60 Pleasant 93 Versatile28 Creative 61 Positive 94 Well-
Mannered29 Daring 62 Powerful 95 Wholesome30 Dependable 63 Practical 96 Youthful31 Determined 64 Precise32 Different 65 Prestigious33 Dignified 66 Productive
2) Select a brand which you have used from the brands given in Group 2
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
( The same set of items used in the case of group 1 was employed here)
3) Select a brand which you have used from the brands given in Group 3
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
(The same set of items used in the case of group 1 was employed here)
4) Demographic Details
Age:
Gender:
Home State:
Approximate Family Income (p.a):
Less than Rs 2 lakhRs 2 lakh to 5 lakhRs 5 lakh to 10 lakhAbove Rs 10 lakh
Religion:
HinduMuslimChristianJainBuddhistOthers
APPENDIX V: Questionnaire for holdout sample survey
Dear Sir/ Madam,
I am Jijo George, a full time research scholar in Pondicherry University. As part of mydoctoral research work, I am conducting a study to understand how people perceivedifferent brands. To elaborate, if I ask you to give me your impression about a particularperson, you may describe him as a smart, sincere and competent person. Similarly thinkof a brand as if it were a person. For instance you may think Raymond to be arespectable, responsible, and a caring brand, Coca Cola to be a friendly, creative,charming brand or MRF to be a sporty, powerful and adventurous brand. The purpose ofthis study is to understand which of the adjectives given below are most appropriate todescribe a brand. I request you to spare a few minutes of your valuable time to take partin this study.
Group 1 Group 2 Group 3Fastrack Rin Whirlpo
ol
Dettol Lee Horlicks
Bata Hyundai Lenovo
Vicks Mirinda Parle - G
APPENDIX V: Questionnaire for holdout sample survey
Dear Sir/ Madam,
I am Jijo George, a full time research scholar in Pondicherry University. As part of mydoctoral research work, I am conducting a study to understand how people perceivedifferent brands. To elaborate, if I ask you to give me your impression about a particularperson, you may describe him as a smart, sincere and competent person. Similarly thinkof a brand as if it were a person. For instance you may think Raymond to be arespectable, responsible, and a caring brand, Coca Cola to be a friendly, creative,charming brand or MRF to be a sporty, powerful and adventurous brand. The purpose ofthis study is to understand which of the adjectives given below are most appropriate todescribe a brand. I request you to spare a few minutes of your valuable time to take partin this study.
Group 1 Group 2 Group 3Fastrack Rin Whirlpo
ol
Dettol Lee Horlicks
Bata Hyundai Lenovo
Vicks Mirinda Parle - G
APPENDIX V: Questionnaire for holdout sample survey
Dear Sir/ Madam,
I am Jijo George, a full time research scholar in Pondicherry University. As part of mydoctoral research work, I am conducting a study to understand how people perceivedifferent brands. To elaborate, if I ask you to give me your impression about a particularperson, you may describe him as a smart, sincere and competent person. Similarly thinkof a brand as if it were a person. For instance you may think Raymond to be arespectable, responsible, and a caring brand, Coca Cola to be a friendly, creative,charming brand or MRF to be a sporty, powerful and adventurous brand. The purpose ofthis study is to understand which of the adjectives given below are most appropriate todescribe a brand. I request you to spare a few minutes of your valuable time to take partin this study.
Group 1 Group 2 Group 3Fastrack Rin Whirlpo
ol
Dettol Lee Horlicks
Bata Hyundai Lenovo
Vicks Mirinda Parle - G
1) Select a brand which you have used from the brands given in Group 1
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
1 Items Ratings SlNo.
Items Ratings SlNo.
Items Ratings
1 Happiness 20 Clear 39 Compatible
2 Joyful 21 Loyal 40 Charming
3 Cheerful 22 Outgoing 41 Appealing
4 Pleasant 23 Lively 42 Charismatic
5 Friendly 24 Precise 43 Stylish
6 Accepting 25 Intelligent 44 Sexy
7 Enthusiastic 26 Practical 45 Youthful
8 Engaging 27 Rational 46 Romantic
9 Comfortable 28 Competent 47 Trendy
10 Positive 29 Professional 48 Sporty
11 Independent 30 Accessible 49 Aggressive
12 Optimistic 31 Innovative 50 Bold
13 Passionate 32 Creative 51 Adventurous
14 Honest 33 Competitive 52 Daring
15Good-Natured 34 Accountable 53 Courageous
16Well-Mannered 35
TrustWorthy 54 Dominant
17 Sincere 36 Believable
18 True 37 Respectable
19 Wholesome 38 Reliable
2) Select a brand which you have used from the brands given in Group 2
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
(The same set of items used in the case of group 1 was employed here)
3) Select a brand which you have used from the brands given in Group 3
Brand Name:
Now, rate each of the items given below on a 5-point scale (1 – Very Inappropriate, 2 –Moderately Inappropriate, 3 – Neutral, 4 – Moderately Appropriate, 5 – VeryAppropriate) according to its appropriateness to describe the brand you have indicatedabove.
(The same set of items used in the case of group 1 was employed here)
4) Demographic Details
Age:
Gender:
Home State:
Approximate Family Income (p.a):
Less than Rs 2 lakhRs 2 lakh to 5 lakhRs 5 lakh to 10 lakhAbove Rs 10 lakh
Religion:
HinduMuslimChristianJainBuddhistOthers
APPENDIX VI: Statistical Outputs
A) Exploratory factor Analysis Results
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .966
Bartlett's Test of Sphericity Approx. Chi-Square 5.844E4
df 1485
Sig. .000
CommunalitiesInitial Extraction
Honest 1 0.751492Happiness 1 0.720678Good-Natured 1 0.642251Joyful 1 0.727262Sincere 1 0.651693Positive 1 0.608061Optimistic 1 0.589512Precise 1 0.567758Cheerful 1 0.745918TRUE 1 0.695297Well-Mannered 1 0.627263Pleasant 1 0.589069Independent 1 0.568404Clear 1 0.601172Practical 1 0.549707Wholesome 1 0.546055Intelligent 1 0.654115Passionate 1 0.577917Outgoing 1 0.582023Enthusiastic 1 0.610099Friendly 1 0.536427Lively 1 0.538462Responsible 1 0.664817Loyal 1 0.490238Engaging 1 0.54471Rational 1 0.490018Sexy 1 0.650946Stylish 1 0.656569Sporty 1 0.597625Trendy 1 0.541567
Romantic 1 0.680776Youthful 1 0.616003Trust Worthy 1 0.563733Reliable 1 0.651801Believable 1 0.586042Respectable 1 0.651395Aggressive 1 0.591732Adventurous 1 0.556696Bold 1 0.666233Courageous 1 0.55985Daring 1 0.569447Accessible 1 0.763259Accountable 1 0.662237Accepting 1 0.62144Dominant 1 0.519854Compatible 1 0.668839Competitive 1 0.663173Competent 1 0.760407Innovative 1 0.753041Creative 1 0.589635Appealing 1 0.509567Comfortable 1 0.578013Professional 1 0.326319Charismatic 1 0.650886Charming 1 0.706559Extraction Method: Principal ComponentAnalysis.
Total Variance ExplainedComponent Initial Eigenvalues Extr
action
Sums ofSquaredLoadings
RotationSums ofSquaredLoading
sTota
l% ofVariance
Cumulative %
Total % ofVariance
Cumulative %
Total
% ofVariance
Cumulative%
1 20.7808
4
37.7833
5
37.78335
20.78084
37.7833
5
37.78335
11.9087
3
21.6522
3
21.652
2 3.32117
2
6.03849
4
43.82184
3.321172
6.03849
4
43.82184
3.98177
7
7.23959
5
28.892
3 2.33437
4.24432
48.06616
2.334377
4.24432
48.06616
3.67496
6.68175
35.574
7 2 2 4 34 1.60
8439
2.92443
5
50.9906
1.608439
2.92443
5
50.9906
3.53682
3
6.43058
7
42.004
5 1.31458
4
2.39015
2
53.38075
1.314584
2.39015
2
53.38075
2.66195
1
4.83991
2
46.844
6 1.22434
2.22607
2
55.60682
1.22434
2.22607
2
55.60682
2.44314
3
4.44207
7
51.286
7 1.11152
8
2.02096
57.62778
1.111528
2.02096
57.62778
2.29073
4
4.16497
1
55.451
8 1.06723
5
1.94042
7
59.56821
1.067235
1.94042
7
59.56821
1.99713
3
3.63115
1
59.082
9 1.02154
9
1.85736
1
61.42557
1.021549
1.85736
1
61.42557
1.28881
2.34329
1
61.426
10 0.95459
4
1.73562
6
63.1612
11 0.89816
5
1.63302
6
64.79422
12 0.86252
6
1.56822
9
66.36245
13 0.80328
6
1.46052
1
67.82297
14 0.78557
6
1.42832
69.25129
15 0.74570
2
1.35582
2
70.60711
16 0.70875
1.28863
7
71.89575
17 0.69140
1
1.25709
2
73.15284
18 0.63553
5
1.15551
8
74.30836
19 0.63422
2
1.15313
2
75.46149
20 0.61301
1
1.11456
6
76.57606
21 0.60296
3
1.09629
7
77.67236
22 0.58408
1
1.06196
5
78.73432
23 0.57 1.04 79.77
4578
4687
901
24 0.55526
8
1.00957
8
80.78859
25 0.55123
8
1.00225
81.79084
26 0.51449
6
0.93544
6
82.72628
27 0.48963
3
0.89024
2
83.61652
28 0.47823
5
0.86951
8
84.48604
29 0.45560
4
0.82837
1
85.31441
30 0.44253
9
0.80461
6
86.11903
31 0.43586
4
0.79247
9
86.91151
32 0.42864
9
0.77936
3
87.69087
33 0.41943
7
0.76261
4
88.45348
34 0.40441
4
0.73529
9
89.18878
35 0.39291
8
0.71439
7
89.90318
36 0.38120
1
0.69309
2
90.59627
37 0.36648
2
0.66633
2
91.2626
38 0.35221
4
0.64038
9
91.90299
39 0.34554
7
0.62826
7
92.53126
40 0.33986
1
0.61792
8
93.14919
41 0.32749
3
0.59544
1
93.74463
42 0.31925
3
0.58046
1
94.32509
43 0.30070
6
0.54673
8
94.87183
44 0.29820
3
0.54218
8
95.41402
45 0.29057
2
0.52831
4
95.94233
46 0.26714
5
0.48571
8
96.42805
47 0.26613
9
0.48388
8
96.91194
48 0.25297
8
0.45996
1
97.3719
49 0.24110
8
0.43837
9
97.81028
50 0.23216
7
0.42212
1
98.2324
51 0.21738
0.39523
7
98.62763
52 0.20358
4
0.37015
3
98.99779
53 0.19246
5
0.34993
6
99.34772
54 0.18426
5
0.33502
8
99.68275
55 0.17448
7
0.31725
100
Extraction Method:Principal ComponentAnalysis.
Rotated Component Matrix
Component
1 2 3 4 5 6 7 8 9Happiness 0.7852
57Honest 0.7752
09Joyful 0.7277
49Good-Natured
0.719811
Positive 0.66641
Cheerful 0.650478
0.476344
Sincere 0.649413
Pleasant 0.6414
42Enthusiastic
0.631044
Optimistic 0.620384
Well-Mannered
0.616102
Independent
0.615333
Passionate 0.608873
TRUE 0.606323
0.472679
Friendly 0.602971
Clear 0.592987
Wholesome 0.590498
Intelligent 0.589082
0.440141
Engaging 0.58825
Lively 0.585019
Precise 0.583589
Practical 0.554616
Outgoing 0.547604
Loyal 0.543292
Rational 0.462261
Stylish 0.708435
Sexy 0.698898
Youthful 0.672062
Trendy 0.646876
Romantic 0.412879
0.643486
Sporty 0.592406
0.450307
Reliable 0.710464
TrustWorthy
0.640095
Respectable
0.639469
Responsible
0.519077
0.587831
Believable 0.512246
Aggressive 0.698058
Bold 0.640274
Adventurous
0.637102
Daring 0.583174
Courageous
0.541115
Competent 0.780911
Competitive 0.663801
Compatible 0.648082
Comfortable
0.510445
Accessible 0.82543
Accountable
0.713912
Accepting 0.667728
Charming 0.650112
Charismatic 0.618478
Appealing 0.519772
Innovative 0.733176
Creative 0.514885
Professional
0.431811
Dominant 0.517
Extraction Method: Principal Component Analysis.Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 9 iterations.
B) Cronbach’s Alpha value Calculations of full version of the scale
Factor 1Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.957 .956 22
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Happiness 70.0496 379.031 .779 .691 .953
Honest 70.1278 377.965 .786 .707 .953
Joyful 70.2375 382.094 .741 .640 .954
Good-Natured 70.1324 381.028 .717 .582 .954
Positive 69.9924 381.353 .741 .609 .954
Sincere 70.2713 380.501 .727 .583 .954
Pleasant 70.0951 383.581 .711 .586 .954
Enthusiastic 70.1779 383.876 .680 .560 .955
Optimistic 70.1826 383.341 .698 .579 .954
Well-Mannered 70.2001 382.517 .708 .570 .954
Independent 70.2911 383.223 .671 .495 .955
Passionate 70.1849 382.584 .690 .538 .955
Friendly 70.0111 384.537 .670 .482 .955
Clear 70.0910 385.717 .652 .454 .955
Wholesome 70.2736 386.440 .631 .463 .955
Engaging 70.2631 386.215 .663 .541 .955
Lively 70.1441 383.850 .674 .511 .955
Precise 70.3390 383.997 .695 .588 .955
Practical 70.1394 384.589 .684 .572 .955
Outgoing 70.2806 385.348 .646 .528 .955
Loyal 69.9376 388.477 .613 .456 .955
Rational 70.4452 388.896 .621 .424 .955
Factor 2
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.774 .781 4
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Stylish 10.2228 10.237 .639 .424 .689
Sexy 10.9978 9.790 .514 .267 .763
Youthful 10.2184 10.365 .584 .347 .716
Trendy 10.3049 10.890 .594 .382 .714
Factor 3
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.769 .770 4
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Trust Worthy 11.2594 7.839 .556 .315 .721
Respectable 11.3980 7.434 .580 .365 .709
Believable 11.1851 7.798 .536 .295 .732
Reliable 11.3071 7.632 .609 .390 .694
Factor 4
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.809 .809 5
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Aggressive 12.8087 16.282 .569 .333 .780
Bold 12.3293 16.036 .636 .413 .759
Adventurous 12.3960 16.304 .564 .332 .781
Daring 12.6352 16.245 .610 .383 .767
Courageous 12.3938 16.408 .597 .383 .771
Factor 5
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.800 .800 4
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Competitive 11.3437 7.979 .571 .431 .769
Competent 11.4767 7.513 .675 .525 .719
Compatible 11.6619 7.161 .655 .478 .728
Comfortable 11.3592 7.929 .552 .371 .778
Factor 6
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.765 .767 3
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Accepting 7.7220 3.441 .561 .320 .726
Accessible 7.7276 3.454 .640 .411 .642
Accountable 8.1340 3.245 .597 .369 .687
Factor 7
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.744 .740 3
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Charming 7.0301 4.103 .632 .475 .582
Charismatic 7.0797 4.035 .670 .497 .535
Appealing 6.7921 5.357 .424 .183 .812
Factor 8
Reliability Statistics
Cronbach's
Alpha
Cronbach's
Alpha Based on
Standardized
Items N of Items
.568 .626 3
Item-Total Statistics
Scale Mean if
Item Deleted
Scale Variance if
Item Deleted
Corrected Item-
Total Correlation
Squared Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Innovative 7.3296 6.343 .502 .338 .335
Creative 7.2937 6.554 .424 .307 .424
Professional 7.3117 4.583 .296 .095 .710
C) Test Re-test Results of Full version PBPS
Factor wise test re-test correlations
Factor 1
Correlations
Factor1T1 Factor1T2
Factor1T1 Pearson Correlation 1 .867**
Sig. (2-tailed) .000
N 258 257Factor1T2 Pearson Correlation .867** 1
Sig. (2-tailed) .000
N 257 262**. Correlation is significant at the 0.01 level (2-tailed).
Factor 2
Correlations
Factor2T1 Factor2T2Factor2T1 Pearson Correlation 1 .909**
Sig. (2-tailed) .000
N 269 269Factor2T2 Pearson Correlation .909** 1
Sig. (2-tailed) .000
N 269 272**. Correlation is significant at the 0.01 level (2-tailed).
Factor 3
Correlations
Factor3T1 Factor3T2
Factor3T1 Pearson Correlation 1 .901**
Sig. (2-tailed) .000
N 270 270Factor3T2 Pearson Correlation .901** 1
Sig. (2-tailed) .000
N 270 271**. Correlation is significant at the 0.01 level (2-tailed).
Factor 4
Correlations
Factor4T1 Factor4T2
Factor4T1 Pearson Correlation 1 .830**
Sig. (2-tailed) .000
N 269 269Factor4T2 Pearson Correlation .830** 1
Sig. (2-tailed) .000
N 269 272**. Correlation is significant at the 0.01 level (2-tailed).
Factor 5
Correlations
Factor5T1 Factor5T2
Factor5T1 Pearson Correlation 1 .807**
Sig. (2-tailed) .000
N 272 272Factor5T2 Pearson Correlation .807** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
Factor 6
Correlations
Factor6T1 Factor6T2
Factor6T1 Pearson Correlation 1 .844**
Sig. (2-tailed) .000
N 272 272Factor6T2 Pearson Correlation .844** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
Factor 7
Correlations
Factor7T1 Factor7T2
Factor7T1 Pearson Correlation 1 .818**
Sig. (2-tailed) .000
N 270 270Factor7T2 Pearson Correlation .818** 1
Sig. (2-tailed) .000
N 270 272**. Correlation is significant at the 0.01 level (2-tailed).
Item wise test re-test correlations
1) Test re-test correlation value for the variable Appealing
Correlations
Appealing Appealing
Appealing Pearson Correlation 1 .762**
Sig. (2-tailed) .000
N 272 272Appealing Pearson Correlation .762** 1
Sig. (2-tailed) .000N 272 272
**. Correlation is significant at the 0.01 level (2-tailed).
2) Test re-test correlation value for the variable Charismatic
Correlations
Charismatic Charismatic2
Charismatic Pearson Correlation 1 .723**
Sig. (2-tailed) .000
N 271 271Charismatic2 Pearson Correlation .723** 1
Sig. (2-tailed) .000
N 271 272**. Correlation is significant at the 0.01 level (2-tailed).
3) Test re-test correlation value for the variable CharmingCorrelations
Charming2 Charming
Charming2 Pearson Correlation 1 .788**
Sig. (2-tailed) .000
N 273 271Charming Pearson Correlation .788** 1
Sig. (2-tailed) .000
N 271 271**. Correlation is significant at the 0.01 level (2-tailed).
4) Test re-test correlation value for the variable Accepting
Correlations
Accepting Accepting2
Accepting Pearson Correlation 1 .712**
Sig. (2-tailed) .000
N 272 272Accepting2 Pearson Correlation .712** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
5) Test re-test correlation value for the variable Accountable
Correlations
Accountable2 Accountable
Accountable2 Pearson Correlation 1 .742**
Sig. (2-tailed) .000
N 273 273Accountable Pearson Correlation .742** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
6) Test re-test correlation value for the variable AccessibleCorrelations
Accessible Accessible
Accessible Pearson Correlation 1 .820**
Sig. (2-tailed) .000
N 273 273Accessible Pearson Correlation .820** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
7) Test re-test correlation value for the variable ComfortableCorrelations
Comfortable2 Comfortable
Comfortable2 Pearson Correlation 1 .818**
Sig. (2-tailed) .000
N 273 272Comfortable Pearson Correlation .818** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed).
8) Test re-test correlation value for the variable CompatibleCorrelations
Compatible Compatible2
Compatible Pearson Correlation 1 .796**
Sig. (2-tailed) .000
N 272 272Compatible2 Pearson Correlation .796** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
9) Test re-test correlation value for the variable CompetitiveCorrelations
Competitive Competitive2
Competitive Pearson Correlation 1 .770**
Sig. (2-tailed) .000
N 273 273Competitive2 Pearson Correlation .770** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
10)Test re-test correlation value for the variable Competent
Correlations
Competent Competent2
Competent Pearson Correlation 1 .780**
Sig. (2-tailed) .000
N 272 272Competent2 Pearson Correlation .780** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
11)Test re-test correlation value for the variable Courageous
Correlations
Courageous Courageous2
Courageous Pearson Correlation 1 .763**
Sig. (2-tailed) .000
N 272 272Courageous2 Pearson Correlation .763** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
12)Test re-test correlation value for the variable Daring
Correlations
Daring Daring2
Daring Pearson Correlation 1 .754**
Sig. (2-tailed) .000
N 271 271Daring2 Pearson Correlation .754** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
13)Test re-test correlation value for the variable Adventurous
Correlations
Adventurous Adventurous2
Adventurous Pearson Correlation 1 .713**
Sig. (2-tailed) .000
N 272 272Adventurous2 Pearson Correlation .713** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
14)Test re-test correlation value for the variable Bold
Correlations
Bold Bold2
Bold Pearson Correlation 1 .742**
Sig. (2-tailed) .000
N 271 271Bold2 Pearson Correlation .742** 1
Sig. (2-tailed) .000
N 271 272**. Correlation is significant at the 0.01 level (2-tailed).
15)Test re-test correlation value for the variable Aggressive
Correlations
Aggressive Aggressive2
Aggressive Pearson Correlation 1 .718**
Sig. (2-tailed) .000
N 273 273Aggressive2 Pearson Correlation .718** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
16)Test re-test correlation value for the variable Believable
Correlations
Believable Believable
Believable Pearson Correlation 1 .850**
Sig. (2-tailed) .000
N 272 272Believable Pearson Correlation .850** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed)
17)Test re-test correlation value for the variable Respectable
Correlations
Respectable Respectable2
Respectable Pearson Correlation 1 .708**
Sig. (2-tailed) .000
N 272 272Respectable2 Pearson Correlation .708** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed).
18)Test re-test correlation value for the variable Trustworthy
Correlations
Trust Worthy Trust Worthy2
Trust Worthy Pearson Correlation 1 .900**
Sig. (2-tailed) .000
N 272 272Trust Worthy2 Pearson Correlation .900** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
19)Test re-test correlation value for the variable Reliable
Correlations
Reliable Reliable2
Reliable Pearson Correlation 1 .821**
Sig. (2-tailed) .000
N 273 273Reliable2 Pearson Correlation .821** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
20)Test re-test correlation value for the variable Trendy
Correlations
Trendy Trendy2
Trendy Pearson Correlation 1 .760**
Sig. (2-tailed) .000
N 273 273Trendy2 Pearson Correlation .760** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
21)Test re-test correlation value for the variable Youthful
Correlations
Youthful Youthful2
Youthful Pearson Correlation 1 .836**
Sig. (2-tailed) .000
N 270 270Youthful2 Pearson Correlation .836** 1
Sig. (2-tailed) .000
N 270 272**. Correlation is significant at the 0.01 level (2-tailed).
22)Test re-test correlation value for the variable Sexy
Correlations
Sexy Sexy2
Sexy Pearson Correlation 1 .865**
Sig. (2-tailed) .000
N 272 272Sexy2 Pearson Correlation .865** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
23)Test re-test correlation value for the variable Stylish
Correlations
Stylish Stylish2
Stylish Pearson Correlation 1 .929**
Sig. (2-tailed) .000
N 273 273Stylish2 Pearson Correlation .929** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
24)Test re-test correlation value for the variable Rational
Correlations
Rational2 Rational
Rational2 Pearson Correlation 1 .704**
Sig. (2-tailed) .000
N 273 273Rational Pearson Correlation .704** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
25)Test re-test correlation value for the variableLoyal
Correlations
Loyal Loyal2
Loyal Pearson Correlation 1 .732**
Sig. (2-tailed) .000
N 271 271Loyal2 Pearson Correlation .732** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
26)Test re-test correlation value for the variable Outgoing
Correlations
Outgoing Outgoing2
Outgoing Pearson Correlation 1 .743**
Sig. (2-tailed) .000
N 265 265Outgoing2 Pearson Correlation .743** 1
Sig. (2-tailed) .000
N 265 271**. Correlation is significant at the 0.01 level (2-tailed)
27)Test re-test correlation value for the variablePractical
Correlations
Practical Practical2
Practical Pearson Correlation 1 .759**
Sig. (2-tailed) .000
N 270 270Practical2 Pearson Correlation .759** 1
Sig. (2-tailed) .000
N 270 273**. Correlation is significant at the 0.01 level (2-tailed).
28)Test re-test correlation value for the variable Precise
Correlations
Precise Precise2
Precise Pearson Correlation 1 .701**
Sig. (2-tailed) .000
N 268 268Precise2 Pearson Correlation .701** 1
Sig. (2-tailed) .000
N 268 271**. Correlation is significant at the 0.01 level (2-tailed).
29)Test re-test correlation value for the variable Lively
Correlations
Lively Lively2Lively Pearson Correlation 1 .742**
Sig. (2-tailed) .000
N 270 270Lively2 Pearson Correlation .742** 1
Sig. (2-tailed) .000
N 270 272**. Correlation is significant at the 0.01 level (2-tailed).
30)Test re-test correlation value for the variable Engaging
Correlations
Engaging Engaging2
Engaging Pearson Correlation 1 .738**
Sig. (2-tailed) .000
N 271 271Engaging2 Pearson Correlation .738** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
31)Test re-test correlation value for the variable Wholesome
Correlations
Wholesome Wholesome2
Wholesome Pearson Correlation 1 .811**
Sig. (2-tailed) .000
N 272 272Wholesome2 Pearson Correlation .811** 1
Sig. (2-tailed) .000
N 272 273**. Correlation is significant at the 0.01 level (2-tailed).
32)Test re-test correlation value for the variable Clear
Correlations
Clear Clear2
Clear Pearson Correlation 1 .763**
Sig. (2-tailed) .000
N 271 271Clear2 Pearson Correlation .763** 1
Sig. (2-tailed) .000
N 271 272**. Correlation is significant at the 0.01 level (2-tailed).
33)Test re-test correlation value for the variable Friendly
Correlations
Friendly Friendly2
Friendly Pearson Correlation 1 .745**
Sig. (2-tailed) .000
N 271 271Friendly2 Pearson Correlation .745** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
34)Test re-test correlation value for the variable PassionateCorrelations
Passionate Passionate2
Passionate Pearson Correlation 1 .743**
Sig. (2-tailed) .000
N 268 268Passionate2 Pearson Correlation .743** 1
Sig. (2-tailed) .000
N 268 269**. Correlation is significant at the 0.01 level (2-tailed).
35)Test re-test correlation value for the variable Independent
Correlations
Independent Independent2
Independent Pearson Correlation 1 .809**
Sig. (2-tailed) .000
N 267 267Independent2 Pearson Correlation .809** 1
Sig. (2-tailed) .000
N 267 272**. Correlation is significant at the 0.01 level (2-tailed).
36)Test re-test correlation value for the variable Well-Mannered
Correlations
Well-Mannered Well-Mannered2
Well-Mannered Pearson Correlation 1 .708**
Sig. (2-tailed) .000
N 272 272Well-Mannered2 Pearson Correlation .708** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed).
37)Test re-test correlation value for the variable Optimistic
Correlations
Optimistic Optimistic2
Optimistic Pearson Correlation 1 .732**
Sig. (2-tailed) .000
N 269 269Optimistic2 Pearson Correlation .732** 1
Sig. (2-tailed) .000
N 269 273**. Correlation is significant at the 0.01 level (2-tailed).
38)Test re-test correlation value for the variable Enthusiastic
Correlations
Enthusiastic Enthusiastic2
Enthusiastic Pearson Correlation 1 .756**
Sig. (2-tailed) .000
N 271 271Enthusiastic2 Pearson Correlation .756** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
39)Test re-test correlation value for the variable Pleasant
Correlations
Pleasant Pleasant2
Pleasant Pearson Correlation 1 .810**
Sig. (2-tailed) .000
N 268 267Pleasant2 Pearson Correlation .810** 1
Sig. (2-tailed) .000
N 267 270**. Correlation is significant at the 0.01 level (2-tailed).
40)Test re-test correlation value for the variable Sincere
Correlations
Sincere Sincere2
Sincere Pearson Correlation 1 .799**
Sig. (2-tailed) .000
N 273 273Sincere2 Pearson Correlation .799** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
41)Test re-test correlation value for the variable Positive
Correlations
Positive Positive2
Positive Pearson Correlation 1 .724**
Sig. (2-tailed) .000
N 267 267Positive2 Pearson Correlation .724** 1
Sig. (2-tailed) .000
N 267 273**. Correlation is significant at the 0.01 level (2-tailed).
42)Test re-test correlation value for the variable Good-natured
Correlations
Good-Natured Good-Natured2
Good-Natured Pearson Correlation 1 .758**
Sig. (2-tailed) .000
N 271 271Good-Natured2 Pearson Correlation .758** 1
Sig. (2-tailed) .000
N 271 273**. Correlation is significant at the 0.01 level (2-tailed).
43)Test re-test correlation value for the variable Joyful
Correlations
Joyful Joyful
Joyful Pearson Correlation 1 .779**
Sig. (2-tailed) .000
N 272 272Joyful Pearson Correlation .779** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed).
44)Test re-test correlation value for the variable Honest
Correlations
Honest2 Honest
Honest2 Pearson Correlation 1 .821**
Sig. (2-tailed) .000
N 273 272Honest Pearson Correlation .821** 1
Sig. (2-tailed) .000
N 272 272**. Correlation is significant at the 0.01 level (2-tailed).
45)Test re-test correlation value for the variable Happiness
Correlations
Happiness Happiness
Happiness Pearson Correlation 1 .933**
Sig. (2-tailed) .000
N 273 273Happiness Pearson Correlation .933** 1
Sig. (2-tailed) .000
N 273 273**. Correlation is significant at the 0.01 level (2-tailed).
D) Confirmatory factor analysis outputs of Full version of PBPS
Raw Data from file 'D:\EDUCATION\Thesispreparation\cfa\original+50\ori50.psf'Latent Variables 1 2 3 4 5 6 7Relationships
Happines = 1Honest = 1
Goodnatu = 1Joyful = 1Positive = 1Sincere = 1Pleasant = 1
Enthusia = 1Wellmann = 1Optimist = 1
Independ = 1Passiona = 1Friendly = 1
Wholesom = 1Clear = 1Engaging = 1Lively = 1Precise = 1
Practica = 1Loyal = 1Outgoing = 1Rational = 1Stylish = 2Sexy = 2Youthful = 2Trendy = 2
Aggressi = 3Bold = 3
Adventur = 3Daring = 3
Courageo = 3Reliable = 4
trustwor = 4Respecta = 4Believab = 4Competen = 5Competit = 5Compatib = 5Comforta = 5Accessib = 6Accounta = 6Acceptin = 6Charming = 7Charisma = 7
Appealin = 7Measurement Equations
Happines = 0.87*1, Errorvar.= 0.70 , R² = 0.52(0.071) (0.069)12.16 10.05
Honest = 0.90*1, Errorvar.= 0.41 , R² = 0.67(0.062) (0.042)
14.58 9.76
Goodnatu = 0.79*1, Errorvar.= 0.49 , R² = 0.56(0.062) (0.049)12.80 9.99
Joyful = 0.93*1, Errorvar.= 0.53 , R² = 0.62(0.067) (0.054)13.82 9.87
Positive = 1.02*1, Errorvar.= 0.44 , R² = 0.70(0.067) (0.045)15.21 9.64
Sincere = 1.04*1, Errorvar.= 0.45 , R² = 0.71(0.068) (0.047)15.25 9.63
Pleasant = 0.90*1, Errorvar.= 0.52 , R² = 0.61(0.066) (0.053)13.66 9.90
Enthusia = 0.86*1, Errorvar.= 0.79 , R² = 0.49(0.074) (0.078)11.66 10.09
Wellmann = 0.99*1, Errorvar.= 0.65 , R² = 0.60(0.073) (0.066)13.48 9.92
Optimist = 1.01*1, Errorvar.= 0.57 , R² = 0.64(0.071) (0.058)14.14 9.83
Independ = 0.93*1, Errorvar.= 0.68 , R² = 0.56(0.073) (0.068)12.82 9.99
Passiona = 0.89*1, Errorvar.= 0.50 , R² = 0.61(0.065) (0.050)13.69 9.89
Friendly = 1.03*1, Errorvar.= 0.56 , R² = 0.65(0.072) (0.057)14.36 9.80
Wholesom = 0.93*1, Errorvar.= 0.72 , R² = 0.54(0.074) (0.072)12.57 10.02
Clear = 0.93*1, Errorvar.= 0.61 , R² = 0.58(0.070) (0.061)13.24 9.95
Engaging = 0.91*1, Errorvar.= 0.71 , R² = 0.54(0.073) (0.071)12.52 10.02
Lively = 0.87*1, Errorvar.= 0.75 , R² = 0.50(0.073) (0.074)11.93 10.07
Precise = 0.98*1, Errorvar.= 0.55 , R² = 0.64(0.070) (0.056)14.08 9.84
Practica = 0.97*1, Errorvar.= 0.68 , R² = 0.58(0.073) (0.068)13.14 9.96
Loyal = 0.85*1, Errorvar.= 0.64 , R² = 0.53(0.069) (0.064)12.33 10.04
Outgoing = 0.79*1, Errorvar.= 0.97 , R² = 0.39(0.078) (0.095)10.14 10.19
Rational = 0.88*1, Errorvar.= 0.86 , R² = 0.48(0.077) (0.085)11.49 10.11
Stylish = 1.20*2, Errorvar.= 0.55 , R² = 0.72(0.079) (0.074)15.08 7.41
Sexy = 1.21*2, Errorvar.= 0.76 , R² = 0.66(0.086) (0.092)14.06 8.20
Youthful = 1.25*2, Errorvar.= 0.54 , R² = 0.74(0.081) (0.076)15.37 7.12
Trendy = 1.07*2, Errorvar.= 0.77 , R² = 0.60(0.082) (0.088)13.02 8.75
Aggressi = 0.83*3, Errorvar.= 0.78 , R² = 0.47(0.074) (0.081)11.13 9.59
Bold = 0.92*3, Errorvar.= 0.67 , R² = 0.56(0.073) (0.073)12.53 9.24
Adventur = 1.08*3, Errorvar.= 0.52 , R² = 0.69(0.073) (0.062)14.71 8.29
Daring = 1.14*3, Errorvar.= 0.31 , R² = 0.81(0.068) (0.047)16.62 6.51
Courageo = 1.01*3, Errorvar.= 0.45 , R² = 0.69(0.069) (0.055)
14.72 8.29
Reliable = 0.83*4, Errorvar.= 0.75 , R² = 0.48(0.074) (0.079)11.21 9.48
trustwor = 0.94*4, Errorvar.= 0.68 , R² = 0.56(0.075) (0.075)12.59 9.08
Respecta = 1.14*4, Errorvar.= 0.38 , R² = 0.77(0.071) (0.056)15.91 6.84
Believab = 1.10*4, Errorvar.= 0.34 , R² = 0.78(0.069) (0.051)16.08 6.63
Competen = 1.04*5, Errorvar.= 0.37 , R² = 0.74(0.067) (0.050)15.48 7.35
Competit = 1.05*5, Errorvar.= 0.44 , R² = 0.72(0.070) (0.056)15.02 7.76
Compatib = 1.04*5, Errorvar.= 0.41 , R² = 0.72(0.068) (0.053)15.15 7.65
Comforta = 0.85*5, Errorvar.= 0.56 , R² = 0.57(0.068) (0.061)12.60 9.06
Accessib = 1.00*6, Errorvar.= 0.70 , R² = 0.59(0.080) (0.087)12.56 8.03
Accounta = 1.10*6, Errorvar.= 0.49 , R² = 0.71(0.076) (0.078)14.36 6.25
Acceptin = 1.01*6, Errorvar.= 0.55 , R² = 0.65(0.075) (0.076)13.47 7.25
Charming = 1.04*7, Errorvar.= 0.40 , R² = 0.73(0.069) (0.056)15.17 7.10
Charisma = 1.12*7, Errorvar.= 0.28 , R² = 0.82(0.068) (0.053)16.50 5.29
Appealin = 0.98*7, Errorvar.= 0.68 , R² = 0.59(0.076) (0.077)12.89 8.80
Correlation Matrix of Independent Variables
1 2 3 4 5 6-------- -------- -------- -------- -------- --------
1 1.00
2 0.45 1.00
3 0.37 0.57 1.00
4 0.62 0.32 0.40 1.00
5 0.58 0.39 0.36 0.65 1.00
6 0.53 0.30 0.25 0.63 0.51 1.007 0.56 0.52 0.58 0.40 0.53 0.51
Correlation Matrix of Independent Variables
7--------
7 1.0
Goodness of Fit StatisticsDegrees of Freedom = 924
Minimum Fit Function Chi-Square = 2238.09 (P = 0.0)Normal Theory Weighted Least Squares Chi-Square =2352.39 (P=0.0)
Estimated Non-centrality Parameter (NCP) = 1428.3990 Percent Confidence Interval for NCP = (1289.13; 1575.28)
Minimum Fit Function Value = 10.36Population Discrepancy Function Value (F0) = 6.61
90 Percent Confidence Interval for F0 = (5.97 ; 7.29)Root Mean Square Error of Approximation (RMSEA) = 0.085
90 Percent Confidence Interval for RMSEA = (0.080 ; 0.089)P-Value for Test of Close Fit (RMSEA < 0.05) = 0.00
Expected Cross-Validation Index (ECVI) = 11.9290 Percent Confidence Interval for ECVI = (11.27 ; 12.60)
ECVI for Saturated Model = 9.58ECVI for Independence Model = 171.59
Chi-Square for Independence Model with 990 Degrees of Freedom =36972.86
Independence AIC = 37062.86Model AIC = 2574.39
Saturated AIC = 2070.00Independence CAIC = 37259.95
Model CAIC = 3060.56Saturated CAIC = 6603.19
Normed Fit Index (NFI) = 0.94Non-Normed Fit Index (NNFI) = 0.96
Parsimony Normed Fit Index (PNFI) = 0.88Comparative Fit Index (CFI) = 0.96Incremental Fit Index (IFI) = 0.96Relative Fit Index (RFI) = 0.94
Critical N (CN) = 100.11Root Mean Square Residual (RMR) = 0.089
Standardized RMR = 0.056Goodness of Fit Index (GFI) = 0.67
Adjusted Goodness of Fit Index (AGFI) = 0.63Parsimony Goodness of Fit Index (PGFI) = 0.60
E) Results of Short version of the scale
E.1. Reliability Values
Cronbach’s alpha values
Factor 1Reliability Statistics
Cronbach'sAlpha N of Items
.795 4
Item-Total Statistics
Scale Mean ifItem Deleted
Scale Variance ifItem Deleted
Corrected Item-Total Correlation
Cronbach'sAlpha if Item
Deleted
Honest 9.7667 9.648 .639 .727Enthusiastic 9.8324 10.104 .592 .751Precise 9.9875 10.095 .614 .740Outgoing 9.9264 10.149 .578 .758
Factor 4Reliability Statistics
Cronbach'sAlpha N of Items
.771 4
Item-Total Statistics
Scale Mean ifItem Deleted
Scale Variance ifItem Deleted
Corrected Item-Total Correlation
Cronbach'sAlpha if Item
Deleted
Aggressive 9.5611 9.887 .558 .724Bold 9.0833 9.957 .589 .707Adventurous 9.1489 9.772 .573 .716Daring 9.3900 10.079 .569 .718
For other factors results are same as in the case of full versionevaluation
Test re-test results
Factor 1
Correlations
ShortFactor11 ShortFactor12ShortFactor11 Pearson Correlation 1 .805**
Sig. (2-tailed) .000
N 263 263ShortFactor12 Pearson Correlation .805** 1
Sig. (2-tailed) .000
N 263 269**. Correlation is significant at the 0.01 level (2-tailed).
Factor 4
Correlations
ShortFactor41 ShortFactor42
ShortFactor41 Pearson Correlation 1 .813**
Sig. (2-tailed) .000
N 269 269ShortFactor42 Pearson Correlation .813** 1
Sig. (2-tailed) .000
N 269 272**. Correlation is significant at the 0.01 level (2-tailed).
For other factors results are same as in the case of full versionevaluation
E.2. CFA Results
Latent Variables F1 F2 F3 F4 F5 F6 F7RelationshipsHonest = F1
Enthusia = F1Precise = F1Outgoing = F1Stylish = F2Sexy = F2Youthful = F2Trendy = F2
Aggressi = F3Bold = F3
Adventur = F3Daring = F3Reliable = F4
trustwor = F4Respecta = F4Believab = F4Competen = F5Competit = F5Compatib = F5Comforta = F5Accessib = F6Accounta = F6Acceptin = F6Charming = F7Charisma = F7
Appealin = F7
Measurement Equations
Honest = 0.85*F1, Errorvar.= 0.50 , R² = 0.59(0.068) (0.064)12.49 7.75
Enthusia = 0.84*F1, Errorvar.= 0.83 , R² = 0.46(0.079) (0.094)10.55 8.87
Precise = 0.99*F1, Errorvar.= 0.53 , R² = 0.65(0.074) (0.075)13.37 6.98
Outgoing = 0.82*F1, Errorvar.= 0.93 , R² = 0.42(0.082) (0.10)9.93 9.11
Stylish = 1.20*F2, Errorvar.= 0.54 , R² = 0.73(0.079) (0.073)15.14 7.35
Sexy = 1.21*F2, Errorvar.= 0.76 , R² = 0.66(0.086) (0.093)14.03 8.22
Youthful = 1.24*F2, Errorvar.= 0.56 , R² = 0.74
(0.081) (0.077)15.25 7.25
Trendy = 1.07*F2, Errorvar.= 0.76 , R² = 0.60(0.082) (0.087)13.12 8.71
Aggressi = 0.86*F3, Errorvar.= 0.73 , R² = 0.50(0.074) (0.079)11.52 9.19
Bold = 0.94*F3, Errorvar.= 0.64 , R² = 0.58(0.074) (0.073)12.72 8.73
Adventur = 1.13*F3, Errorvar.= 0.41 , R² = 0.75(0.073) (0.062)15.46 6.61
Daring = 1.08*F3, Errorvar.= 0.43 , R² = 0.73(0.072) (0.061)15.09 7.02
Reliable = 0.82*F4, Errorvar.= 0.75 , R² = 0.47(0.074) (0.079)11.17 9.49
trustwor = 0.94*F4, Errorvar.= 0.68 , R² = 0.56(0.075) (0.075)12.55 9.09
Respecta = 1.14*F4, Errorvar.= 0.38 , R² = 0.77(0.071) (0.056)15.92 6.80
Believab = 1.11*F4, Errorvar.= 0.34 , R² = 0.78(0.069) (0.051)16.12 6.55
Competen = 1.03*F5, Errorvar.= 0.37 , R² = 0.74(0.067) (0.050)15.46 7.35
Competit = 1.05*F5, Errorvar.= 0.44 , R² = 0.72(0.070) (0.056)15.03 7.74
Compatib = 1.04*F5, Errorvar.= 0.41 , R² = 0.73(0.068) (0.053)15.16 7.63
Comforta = 0.85*F5, Errorvar.= 0.56 , R² = 0.56(0.068) (0.062)12.58 9.06
Accessib = 1.00*F6, Errorvar.= 0.70 , R² = 0.59(0.080) (0.087)12.56 8.01
Accounta = 1.10*F6, Errorvar.= 0.49 , R² = 0.71(0.077) (0.078)14.30 6.30
Acceptin = 1.01*F6, Errorvar.= 0.55 , R² = 0.65(0.075) (0.076)13.50 7.19
Charming = 1.04*F7, Errorvar.= 0.41 , R² = 0.73(0.069) (0.056)15.06 7.20
Charisma = 1.12*F7, Errorvar.= 0.27 , R² = 0.82(0.068) (0.053)16.57 5.16
Appealin = 0.98*F7, Errorvar.= 0.67 , R² = 0.59(0.076) (0.077)12.93 8.79
Correlation Matrix of Independent Variables
F1 F2 F3 F4 F5 F6-------- -------- -------- -------- -------- --------
F1 1.00
F2 0.50 1.00(0.06)
7.96
F3 0.39 0.58 1.00(0.07) (0.05)
5.73 10.86
F4 0.60 0.32 0.37 1.00(0.06) (0.07) (0.07)10.81 4.70 5.58
F5 0.56 0.40 0.34 0.65 1.00(0.06) (0.06) (0.07) (0.05)
9.77 6.08 5.09 13.92
F6 0.47 0.30 0.23 0.63 0.51 1.00(0.07) (0.07) (0.07) (0.05) (0.06)
7.01 4.15 3.05 12.15 8.49
F7 0.60 0.52 0.57 0.40 0.53 0.51(0.06) (0.06) (0.05) (0.06) (0.06) (0.06)10.76 9.11 10.35 6.20 9.42 8.51
Correlation Matrix of Independent Variables
F7--------
F7 1.00
Goodness of Fit Statistics
Degrees of Freedom = 278Minimum Fit Function Chi-Square = 603.18 (P = 0.0)
Normal Theory Weighted Least Squares Chi-Square = 613.05 (P = 0.0)Estimated Non-centrality Parameter (NCP) = 335.05
90 Percent Confidence Interval for NCP = (267.26 ; 410.58)
Minimum Fit Function Value = 2.79Population Discrepancy Function Value (F0) = 1.55
90 Percent Confidence Interval for F0 = (1.24 ; 1.90)Root Mean Square Error of Approximation (RMSEA) = 0.075
90 Percent Confidence Interval for RMSEA = (0.067 ; 0.083)P-Value for Test of Close Fit (RMSEA < 0.05) = 0.00
Expected Cross-Validation Index (ECVI) = 3.5190 Percent Confidence Interval for ECVI = (3.20 ; 3.86)
ECVI for Saturated Model = 3.25ECVI for Independence Model = 45.56
Chi-Square for Independence Model with 325 Degrees of Freedom =9789.85Independence AIC = 9841.85
Model AIC = 759.05Saturated AIC = 702.00
Independence CAIC = 9955.73Model CAIC = 1078.78
Saturated CAIC = 2239.34
Normed Fit Index (NFI) = 0.94Non-Normed Fit Index (NNFI) = 0.96
Parsimony Normed Fit Index (PNFI) = 0.80Comparative Fit Index (CFI) = 0.97Incremental Fit Index (IFI) = 0.97Relative Fit Index (RFI) = 0.93
Critical N (CN) = 121.24
Root Mean Square Residual (RMR) = 0.087Standardized RMR = 0.054
Goodness of Fit Index (GFI) = 0.82Adjusted Goodness of Fit Index (AGFI) = 0.77Parsimony Goodness of Fit Index (PGFI) = 0.65