The impact of the use of neuromarketing on a brands’ CBBE

The impact of the use of neuromarketing on a brands’ CBBE

University of Amsterdam Faculty of Economics and Business

Master of Science in Business Administration Track: Marketing

Under supervision of: dr. Evşen Korkmaz

By:

Student: Dima Shipil’ko Student number: 10732284

Statement of originality

This document is written by Dima Shipil’ko who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for its contents

. 1 . Introduction

7

. 2 . Literature review

10

2.1 Neuromarketing 10

2.1.1 Functional Magnetic Resonance Imaging 13

2.1.2 Electroencephalography 14

2.1.3 Magnetoencephalogram 15

2.1.4 Heart Rate 15

2.1.5 Challenges for neuromarketing 16

2.2 Consumer-based brand equity

19

2.2.1 Defining customer-based brand equity 19

2.2.2 Ingredients for customer-based brand equity 20

2.3 Conceptual model and construction of hypotheses

22

2.3.1 Main Hypothesis 23 2.3.2 Moderating hypotheses 24 2.3.3 Conditional hypothesis 26

. 3 . Research method

27

3.2 Instrumentation

27

3.3 Procedure

31

. 4 . Results

32

. 5 . Discussion

40

5.1 Relevance and Implications

42

5.1.1 Academic relevance 42

5.1.2 Managerial implications 42

5.2 Limitations and future research

43

. 6 . Conclusion

45

. Reference list

47

Internet resources 58

APPENDIX 1 - CBBE & Uncertainty Avoidance questionnaire

60

APPENDIX 2 - Coca-Cola case

62

APPENDIX 3 - Pepsi-Cola case

63

APPENDIX 4 - Neuromarketing case

64

APPENDIX 5 - ANOVA for the experimental groups per CBBE variable

65

APPENDIX 6 - Effect size of the levels of Uncertainty Avoidance on the

CBBE variables

66

Abstract

Neuromarketing is a marketing method already being used for 13 years, but the method is still in its infancy. This method has much potential for marketers and firms, but also threads on thin ice for consumer acceptance, as marketing does not generate positive feelings within consumers in general (Heath & Heath, 2008). The current research was set out to find the impact that neuromarketing could have on the CBBE (Keller, 1993) of two known brands, namely Pepsi-Cola and Coca-Cola. Moreover, according to the Coca-Cola exper-iment (Du-Jian Gang et al., 2012), Coca-Cola is a brand with strong positive associations compared to Pepsi-Cola. This research tried to replicate the results of the former experi-ment, to test wether consumers have a more positive bias towards Coca-Cola than Pepsi-Cola. Furthermore, as Hofstede’s (1980) cultural dimensions seem to impact a consumers perception of a firms unethical behavior (Leonidiou et al., 2012), this research tried to find the moderating impact of Uncertainty Avoidance on the decline of CBBE when a firm uses neuromarketing as a marketing tool. 152 cases of survey data were collected from Dutch consumers, which showed a more positive bias towards Coca-Cola, but did not find a de-cline in CBBE when a brand used neuromarketing. Furthermore, no moderating relation-ship was found through Uncertainty Avoidance. This thesis is concluded by managerial implications and suggestions for future research.

Keywords: neuromarketing, CBBE, uncertainty avoidance, consumer response towards

Aknowledgments

I would like to thank the following people for their aid during the writing of this thesis. First of all, I would like to thank dr. Evşen Korkmaz for her helping hand and guidance during the thesis process. Furthermore, I would like to thank every individual that helped me with the study, for filling out my questionnaire and helping me find more respondents. I would also like to thank Anouar El Haji for clearing up my experimental design. Thank you all again for your support.

I hope you enjoy reading this thesis.

. 1 .

Introduction

In the 90’s, professor Gerald Zaltman of the Harvard Business School took an interest in investigating the consumers’ brain by using neuroscientific techniques to get a better un-derstanding of their decision making process (Kwiatkowska, 2008). This resulted in a new emerging field in marketing which, in 2002, was dubbed ‘neuromarketing’ (Smidts, 2002). This new field combined the research field of marketers and neuroscientists, in which the scientific methods like the functional magnetic resonance imaging (fMRI), electroen-cephalography (EEG), galvanic skin response (GSR), magnetoenelectroen-cephalography (MEG), steady state topography (SST), heart rate and respiratory rate (Giovanni et al., 2011; Vec-chiato et al., 2010) are used to measure consumer preferences, consumer behavior and their responses to marketing stimuli (Lewis & Bridger, 2005). This method also has opened up possibilities for marketers for more effective marketing programs and it also comple-ments modern marketing methods (Morin, 2011). Neuromarketing is said to be the best method so far to collect unbiased consumer information (Egrie & Bietsch, 2014) but it is also confined with ethical challenges. Marketing by itself walks a thin line between com-mercial persuasion and manipulation of the consumer (Wible, 2011), and with neuromar-keting the risk for manipulation has increased.

Since the method was first introduced, research on the topic has grown rapidly. Re-search subjects ranged from showing the important implications neuromarketing has for the marketers and consumers (i.e. Javor, Koller & Ransmayr, 2013; Morin, 2011; Hubert et al., 2009) to the ethical issues that arise from this new research method (i.e. Wilson, Gaines & Hill, 2008; Murphy, Illes & Reiner, 2008; Egrie & Bietsch, 2014). With the intro-duction of neuromarketing, the ethical aspects of this method have become a high priority for researchers. A big concern in this research is the question if marketers are able to by-pass the consumers’ free will (Fisher et al., 2010). With current technology, companies like

Shopperception and Anaxa Vida follow customers real-time by using in-store movement- and eye-tracking (for more information: www.shopperception.com and www.anaxa-vida.-com). This form of ‘stealth neuromarketing’ (Fisher, Chin & Klitzman, 2010) can be a pri-mary source of unconscious consumer preference and can help marketers construct more efficient, but also more intrusive, marketing strategies. This might also have an impact on the perception consumers have of companies that use this marketing method, but this has not been studied thoroughly. As the consumer is the centre of focus within marketing, he is the most important factor within marketing research (Heath & Heath, 2008). However as the latter authors propose, some but not much, research has been conducted on the con-sumers’ response and attitude towards marketing. Where marketers are searching for the best ways to sell their products, the response of the consumer has often been left out. The studies that have been conducted, showed mostly negative media coverage (Dalsace & Markovitch, 2009) and negative responses because of the lack of trust by consumers in marketing (Sheth & Sisodia, 2005; Gaski, 2008). With neuromarketing being a relatively new marketing method, no research has been conducted on the acceptance of this prac-tice by consumers. As this might have an impact on the way consumers judge a brand, it is therefore interesting to know how the use of neuromarketing affects a brands value. The value a brand has to a consumer can be described by Keller’s (1993) customer-based brand equity (CBBE). The author defines this as the added value a brand has for a

prod-uct, compared to the same product without the brand name. A famous neuromarketing ex-periment that shows this brand equity, was the ‘Coca Cola exex-periment’. For this, an fMRI scan was used to measure the preference for a brand, comparing Coca Cola with Pepsi-Cola. In the blind condition, the majority of participants preferred Pepsi-Cola over Coca Cola when tasting the drink. But when participants knew what drink they were served, the fMRI scans showed a strong preference toward Coca Cola because of the positive associ-ations they had with the brand (Du-Jian Gang et al., 2012, p. 285). The results of this study

show how much a brand can affect consumer decision. 


In this thesis we attempt to find out whether the knowledge of the use of neuromar-keting by a brand affects its CBBE and what moderates this relationship. In light of the Coca-Cola experiment, we use Coca-Cola and Pepsi-Cola to measure the effect of neu-romarketing on their CBBE. The reason for choosing Coca-Cola as the main brand is, first-ly, because it’s the world’s number four most valuable brand, secondly because Coca Cola has publicly claimed to use neuromarketing (respectively Forbes, 2014; 2013) and to repli-cate the results of the Coca-Cola experiment (Du-Jian Gang et al., 2012). With our re-search, we wish to fill the gap of consumer response toward neuromarketing. Our main research question is therefore “Does the use of neuromarketing generate a decline in the CBBE of a brand”. Testing this is important, because ethical marketing is a valuable aspect

of conducting this practice. We acknowledge that neuromarketing might be seen as an un-trustworthy method by the consumers as they do not seem to trust marketing in general. Therefore we want to state the challenges for consumer approval of neuromarketing. Fur-thermore, individuals may vary in their responses to neuromarketing. Some consumers tend to seize their brand purchase if they deem its practices to be untrustworthy. According to Leonidou et al. (2012) these individuals seem to score high in uncertainty avoidance (as described by Hofstede, 1980), which states that they respond negatively to ambiguous sit-uations and untrustworthy behavior. We therefore believe that individuals’ uncertainty avoidance will have a moderating effect on the relationship between neuromarketing and CBBE. In the following section we will discuss neuromarketing, its practices and its the ethical concerns regarding its practice. In this section we will also describe the term brand equity and we will propose hypothesis based on the assumptions made from the literature. In the later sections, we will describe our research method along with the results, our con-clusions, implications and future research.

. 2 . Literature review

There are many marketing research methods used today, but they seem to face a com-mon problem. These methods rely on the assumption that consumers are able to describe their own cognitive process, but this seems to be more hope than fact. The reason for this is that a consumers’ cognition has many subconscious elements, Morin (2011) concludes. He describes the fact that consumers could be prone to distorting the research, for rea-sons like incentives, time constraints or peer pressures. Kwiatkowska (2008) also argues that the modern marketing is based on theories from the 1950’s. She notes that the ways of our understanding of consumers has long been outdated and that new methods should be developed. Current research methods consist of qualitative as well as quantitative methods. Qualitative methods use an inductive approach on a small population, and they are a good way to provide a psychological analysis and form a theory. On the other hand, the amount of subjectivity leaves it open for bias of interpretation and its test procedure is harder to control (Amaratunga et al., 2002). Another way of studying consumer behavior, is the quantitative method. Herein, researchers use measurements to test hypotheses which is focused on a large population in a controlled situation. It is efficient and fast, but it is in-flexible and it can not generate theories (Amaratunga et al., 2002). Therefore, since 2002, neuromarketing has attracted the attention of many marketers.

Neuromarketing was described in 2002 by Smidts (2002) as the research that is able to measure neural activity in consumers brains while being exposed to marketing stimuli. This research combines neuroscientific and marketing methods to gather more ef-ficient information about consumers’ responses to marketing stimuli. Javor et al. (2013) have distinguished two definitions for the use of neuroscience for marketing purposes in their article, neuromarketing being the first and ‘consumer neuroscience’ as the second. They define consumer neuroscience as the scientific research of consumer behavior.

Neu-romarketing, on the other hand, was defined as the commercial application of neurology. Consumer neuroscience relies on studies done in other areas of neuroscience, for exam-ple social neuroscience and neuroeconomics (Smidts et al., 2014). These studies draw conclusions on framing (De Martino et al., 2006), self-control (Hare, Camerer & Ranger, 2009), heuristic choice (Venkatraman et al., 2009) and trust, fairness, and reciprocity (Hsu, Anen & Quartz, 2008). These research areas provide a solid basis for theorizing consumer behavior (Yoone et al., 2012). These theories can be used to describe consumer decision making and they can be used to build decision making models. These models and theo-ries, in their turn, can be applied in neuromarketing research.

When neuromarketing was introduced, it was seen as a very promising marketing tool of those times. Morin (2011) rendered that neuroimages could be used to measure consumers emotional responses and get an insight on their latent wants and needs. With this tool, marketers could get an inside-look into the mind of the consumer and therefore provide a more customized product-promotion program. The marketing tool was perceived to have much potential and its use by commercial companies has increased rapidly, from 13 in 2008, 60 in 2012 and to over 85 by 2013 (Levallois, Smidts & Wouters, 2013). Com-panies like Brighthouse, SalesBrain, BrainTrends, BrainSigns, etc. are specialized in con-ducting neuromarketing research and help brands effectively present their product. The previously named Coca Cola-experiment is an example of such research. The results of that experiment show that consumer decision making depends on more than just rational preferences. Another research by Hubert et al. (2009) also shows how much latent wants and needs affect the purchase decisions. They hypothesized that emotions and memories can influence the perception of product packaging, and that these perceptions can affect purchase decisions. The results state that perception of brands and products rely on con-sumers’ emotional experiences with that brand. This leads to the notion that effective mar-keting is lead by relevant emotional responses by consumers (Bagozzi, 1997). Although

the concept of emotions affecting consumer decisions is not new within marketing (e.g. Laros & Steenkamp, 2005; Holbrook & Hirschman, 1982), neuroscientific methods are more efficient in measuring emotional responses through neural activity than other market-ing methods. These responses are then measured in the limbic system, the part of the brain that is associated with emotions, seen in figure 1 (Joseph, 1992). (Neuro-)marketing is most effective when its able to tap into the relevant regions of the brain; i.e. the amyg-dala is associated with emotional memory (Phelps & Anderson, 1997), the hypothalamus with controlling motivational states (hunger, satisfaction, comfort etc.) (meta-health.com, 2012) and the hippocampus that controls long- and short-term memory (National Institute of Health, 2001). Various researchers believe that using neuroscientific methods for mar-keting can aid marketers to, more efficiently,

tap into consumer decision making process-es. Knutson et al. (2007) stated that the use of an fMRI scanner during a shopping task could aid in better predicting purchasing deci-sions. The use of EEG technology, wherein

brainwaves can be measured, prove to be effective for marketing efficacy (Ohme et al., 2009). Further-more, consumer neuroscience has the advantage of

refining current theories and models that can provide insights in consumer decision mak-ing (Greene et al., 2001). Like it has been noted above, current marketmak-ing methods seem to be outdated (Kwiatkowska, 2008) and according to Bercea (2013), neuromarketing is able to bridge the gap between current methods (i.e. qualitative and quantitative methods). “It implies defining a problem (qualitative approach), defining and test hypothesis (quanti-tative approach) and exploring the results in depth (quali(quanti-tative approach)” (Bercea, 2013, p. 9). Venkatraman et al. (2011) suggest that neuromarketing currently has contributed to figure 1. The limbic system of the brain.

Retrieved from http://webspace.ship.edu/

our understanding of consumer behavior by making it possible to measure preferences by measuring subjective value, that it is able to measure individual preferences which can aid in the understanding of how and when preferences can change (Chib et al. 2009; Mon-tague, King-Casas & Cohen, 2006; Plassmann, O’Doherty & Rangel, 2007). Neuromarket-ing seems to have much potential for marketers and companies, and it is seen as a

method that can aid them in many ways. It complements current marketing tools and it is able to help generate more efficient marketing strategies. But this method also faces its own challenges. Neuromarketing deals with criticism by researchers, is subjected to ethi-cal problems and might generate a negative response within consumers. These chal-lenges will be discussed later, in section 2.1.5. Firstly we will discuss the methods used by neuromarketing in the following paragraphs. As it was noted before, neuromarketing uses various neurological techniques (e.g. fMRI, EEG, MEG). We will describe these briefly be-low, but only describe the three most popular methods used within neuromarketing and one that contrasts the most with those methods which is also frequently used by neuro-marketers.

2.1.1 Functional Magnetic Resonance Imaging

Functional magnetic resonance imaging (fMRI) is a neurological non-invasive method (the body of the participant is not ‘invaded’ or cut open during research) used to measure and map the consequences of altered electrical brain activity (Matthews & Jezzard, 2004). These consequences can be measured through changes in blood perfusion, blood volume or blood oxygenation that are caused through neuronal activity within the brain. The last method is called blood oxygenation level dependent (BOLD) fMRI. Whenever neuronal ac-tivity is increased, blood flow to the brain increases along with the oxygen levels

(Matthews & Jezzard, 2004). This activity is then measured by the fMRI machine which produces images of a changing mental state (which is

displayed in Figure 2). This method is able to provide an image spatial resolution of a few millimetres and a tempo-ral resolution of a few seconds. In other words, this method can provide a more precise but a slower (than other methods) image of neuronal activity. This method can be used to provide images of activity within the previ-ously mentioned limbic system, and thus a more objec-tive consumer reaction to marketing stimuli. For a more detailed overview of workings of fMRI and brain

process-es during neuronal activity, we want to refer the reader to the book by Ulmer and Jansen (2013) called fMRI - Basics and Clinical Application.

2.1.2 Electroencephalography

Another method for measuring neuronal activity is called electroencephalography, or EEG (Matthews & Jezzard, 2004). This non-invasive method

localises the underlying electrical activity of the brain through electrical pulses sent out during brain activities by placing electrodes on the scalp. This method provides a fast temporal resolution (10 - 100 milliseconds), but a bigger spatial resolution (between 1 and several cen-timetres). Therefore this method is able to produce a quicker response of neuronal activity than the fMRI, but with a less precise position within the brain. To measure the neural activation within the brain, electrodes are

figure 2. fMRI scan of the brain. Retrieved

from

http://psychcentral.com/lib/img/fm-ri_scan.jpg

figure 3. Spike waves of EEG scan of the

brain. Retrieved from

http://upload.wikime-dia.org/wikipedia/commons/thumb/2/26/ Spike-waves.png/220px-Spike-waves.png

placed on the participants’ scalp that capture the change in voltage as neurons send out pulses when activated (Kennett, 2012). This technique allows marketers to measure fast changes in brain activity, but this method suffers from less precise localisation of brain ac-tivation than the fMRI. For a more detailed explanation of this method and its basics, the book Practical Approach to Electroencephalography by Liberson provides a great over-view.

2.1.3 Magnetoencephalogram

A more recent method than the EEG is the magne-toencephalogram, MEG for short (Lancet, 1990). Neuronal activity produce electrical charges that are measured by EEG, but these electrical charges also produce magnetic fields. These magnetic fields therefore are then measured by the MEG, also a non-invasive method. This method also has the same temporal resolution as the EEG, but it

bene-fits from a better spatial resolution (Lancet, 1990). This latter benefit stems from the fact that the human scalp is relatively bad with conduction, and thus may distort the electrical charges sent out by neural activation, whereas magnetic fields are not distorted. Therefore MEG’s implication for marketers are similar to EEG, but with a more precise brain activa-tion image. For a more detailed discussion on MEG, Clinical Magnetoencephalography and Magnetic Source Imaging by Papanicolaou provides a good overview of this method.

2.1.4 Heart Rate

An increased heart rate (HR) is a physical reaction to sympathetic nervous activation (Vecchiato et al., 2010). The sympathetic nervous system is activated whenever a person

figure 4. MEG scan. Retrieved from: http:// mommasmoneymatters.com/wp-content/up-loads/2012/04/meg.jpg

is stimulated or excited, otherwise known as the ‘fight-or-flight’ reaction. This method mea-sures the HR when participants are exposed to (marketing) stimuli, and thus meamea-sures their excitement. If excitement occurs, marketers are able to relate this to emotional re-sponse toward the marketing stimuli. We shall not go into greater detail for this method, as it mostly speaks for itself. In the following section we shall discuss the challenges faced by neuromarketing.

2.1.5 Challenges for neuromarketing

Research on the benefits of neuromarketing has also been accompanied by research on the challenges this method faces. Ariely and Berns (2011) distinguish the hypes from hopes for neuromarketing in their article. They note that this research area has much po-tential, but do state that this method is also highly expensive. Furthermore, neuromarket-ing companies suggest that this method has the potential of findneuromarket-ing the ‘buy-button’ within the consumers (Javor et al., 2013). Although these companies try to convince that this buy-button exists, the latter authors suggest otherwise. This means that neuromarketing by itself is not a holistic marketing method, but is seen as more complementary to other current methods (Javor et al., 2013). Furthermore, and more importantly, neuromarketing is also subjective to ethical issues (Dinu, 2013). This method being a primary subject in ethical marketing discussions, has attracted much research attention (Wilson, Gaines & Hill, 2008; Fisher, Chin & Klitzman, 2010; Lindell & Kidd, 2013). An issue that has been emphasized by Wilson et al. (2008), is that consumers lose their free will in buying deci-sions if marketers are able to surpass the consumers unwillingness to buy a product. If marketers are able to convince consumers without their consent, then this can be seen as an intrusion into the consumers mind (Canli & Amin, 2002; Illes, 2003; Hyman, 2004; Kennedy, 2004; New York City Bar Association, 2005). This could allow marketers to ma-nipulate consumers into making purchase decisions (Neuroscience, 2004). Wilson et al.

(2008) conclude that consumers should be informed about the use of neuroimaging data for commercial use, and they should be able to give their consent in the use of this method for influencing consumer decision making while purchasing products. The authors state that future research must be conducted on the ownership of neuroimaging information, how can it be combined with existing databases and under what conditions can it be used by companies and/or sold to third parties.

Another study conducted by Fisher et al. (2010) states a similar research question and results, namely what ethical issues can arise from the use of neuromarketing in com-mercial applications. They conclude that neuromarketing can have significant professional, ethical and scientific concerns. Professional ethics can be a difficult subject to discuss and the new practice of neuromarketing can host contradicting opinions on the use of this re-search method. Fisher et al. (2010) also point out the possibility of stealth neuromarketing, but concluded that this wasn't possible at that time with the technology available. But as Foscht and Swoboda (2012) predicted, it would be possible to monitor consumers in the store much more effectively. Like it has been described in the introduction, today’s compa-nies like Shopperception and Anaxa Vida use real-time in-store monitoring and eye-track-ing of consumers. In the eye of the consumer, neuromarketeye-track-ing methods can hold an immi-nent threat to their free will. Smidts et al. (2014) also bring up the advancement in con-sumer neuroscience, in which they propose marketing possibilities within the genetics and molecular neuroscience. This poses the question if marketers, with the current technology and upcoming technologies, will still be able to conduct ethical marketing practices. Be-cause of the likelihood of this threat, the ‘Neuromarketing Science and Business Associa-tion’ (NMSBA) has drawn up a Code of Ethics that prevents the use of neuromarketing re-search with the intention to manipulating the consumer (NMSBA, 2013). All its members are bound to comply to this code and are not allowed to take advantage of the partici-pants. Still a lot of research must be done on this fairly new research method. Fisher et al.

(2010) provide future research questions comparable to Wilson et al. (2008), stating that research is needed to see by whom and for which goals neuromarketing is being used, and ask if there is a perceived return on the use of this method. Effective use of the

method may benefit a companies revenue, but this aspect is also susceptible to consumer response to neuromarketing. 


If neuromarketing is to be used unethically, this might damage the consumer trust and thus result in a negative impact on the company using the method. Therefore mar-keters must be aware of the consumer response toward neuromarketing. As there is some research on consumer acceptance of marketing (Sheth and Sisodia, 2005; Dalsace & Markovitch, 2009; Gaski, 2008; Kachersky & Lerman, 2013), no research could be found on consumer attitude toward neuromarketing. The research that has been conducted on consumers’ view on marketing, showed mostly negative responses because of the belief that marketing conducted untrustworthy and unethical behavior. Consumers believe that marketing efforts are mostly used to benefit the firms instead of satisfying consumers’ wants and needs (Kachersky & Lerman, 2013). Marketers must therefore be careful in us-ing neuromarketus-ing in the wrong ways, because consumers might respond negatively to this practice if they perceive it as unethical (Yoo & Donthu, 2002). These responses may be influenced by individual characteristics, such as Hofstede’s (1980) cultural dimensions (i.e. uncertainty avoidance, individualism). Yoo and Donthu (2002) concluded that uncer-tainty avoidance and collectivism is positively related to firms’ marketing ethics. Other search also shows that individual perception of ethical attitudes and judgement is also re-lated to individualism and uncertainty avoidance (Jackson, 2001). These results show that ethical behavior by brands will be judged differently by an individuals level of the cultural dimensions. The way a brand is judged will have an effect on its consumer value. With our research, we wish to measure if the use of neuromarketing will affect the value of a brand

that uses this marketing method. This value of a brand is described by customer-based brand equity and is described in the following section.

2.2 Consumer-based brand equity

Brand equity has been studied on multiple accounts (i.e. Aaker & Biel, 1992; Farquhar, 1989) and is described as the added value a brand name has to a product opposing the same product without the brand name. In this section we firstly describe customer-based brand equity (CBBE) as defined by Keller (1993) in his article. This is a logical step to take, as our research is also focused on consumer response to neuromarketing. Furthermore we will also briefly discuss what factors build up customer-based brand equity, after which we will present our hypotheses based on this description of CBBE. Herein we will describe how CBBE might be affected by the use of neuromarketing by brands.

2.2.1 Defining customer-based brand equity

Branding of a product is a practice that has been done for centuries. The reasons for branding consist of assuring consumers of a certain quality, legal protection and offering value to a product (Farquhar, 1989). The latter is also known as brand equity (Leuthesser, 1998). Much research, especially within Marketing Science Institute (MSI), has been con-ducted on the concept of brand equity and how this helps to build, manage and extend a brand (i.e. Broniarczyk & Alba, 1994; Farquhar, 1989; Feldwick, 1996). Keller (1993) de-fines CBBE “as the differential effect of brand knowledge on consumer response to the marketing of the brand” (p. 8). Differential effect, in Keller’s (1993) paper, is defined as the effect a known brand has over a fictitious brand when the same marketing mix occurs. Consumer response can be seen as the consumers’ preference, perception and

behav-iours that occur from the marketing mix. And the latter, brand knowledge, he describes by brand awareness and brand image. A clear overview of the latter concepts can be seen in

Figure 5. A more detailed description can be found in Keller’s (1993) article, as it goes be-yond the scope of this thesis. Further in this section, we will focus on building customer-based brand equity.

2.2.2 Ingredients for customer-based brand equity

To achieve a high level of customer-based brand equity, a brand must develop strong brand associations through well chosen brand identities (i.e. name, logo) and these must be supported by a strong marketing program (Keller, 1993). The choice of brand identities can help enhance brand awareness or link brand associations. A well chosen brand name, for example, can affect the recall and recognition of the brand (Alba & Hutchinson, 1987). The latter authors conclude that a brand name/logo must be simple, familiar and distinctive

for a strong and positive affect. For this to be effective, a supporting marketing program must be developed that can increase brand awareness and the intention to purchase the brand’s products (Keller, 1993). The marketing program can help increase the brand famil-iarity, which is defined as the accumulated product experience by the consumer (Alba &

Hutchinson, 1987). A well chosen marketing program can affect favourable, strong and unique brand associations and experiences, which can help increase the CBBE (Keller, 1993).

Besides these primary and more direct sources of brand associations (i.e. brand name or communications by the brand), consumers also develop their own associations that can be indirectly linked to the brand. This is what Keller (1993) calls secondary asso-ciations. These can be evoked through the company, the country of origin, the distribution channels, celebrity spokesperson or an event. The first three are associations from factual sources (who makes it, where is it made, where is it purchased) and this information is mostly available to everyone. For example, consumers may deem a company to be reli-able with their products (i.e. Apple). Consumers may also think the country of origin plays a big role in a product’s reliability (i.e. German cars are seen as efficient and reliable). Also, the distribution channels may play a role in forming brand associations (i.e. certain retailers selling certain brands). The latter sources of secondary associations, the celebrity spokesperson or an event, “occur when primary brand associations are for user and usage situation attributes” (Keller, 1993). When a celebrity promotes a brand, associations to-wards this celebrity can spill over to the brand (Rossiter & Percy, 1987). This is also possi-ble when the brand is linked to a certain event by the consumer (Keller, 1993). These sec-ondary associations may become important when a consumer does have existing brand associations, but when these lack power. Secondary associations can then help to achieve strong and favourable associations toward the brand, however relying on these can be risky as there is less control over the consumers’ secondary attributes toward the brand. 


In the following section we construct our hypotheses and conceptual model based on the literature noted above. Afterwards we describe our research method in which we give a detailed explanation of the instrumentation used for our research. We conclude our research by presenting the results, discussion and room for future trajectories.

2.3 Conceptual model and construction of hypotheses

In this thesis we want to describe the relationship between Neuromarketing and CBBE. This relationship is shown in Figure 6, our conceptual model, with ‘Brand’ and ‘CBBE’ as the two major factors. Within this model we describe the positive relationship between the brand and CBBE. As described in the previous section, a brand can have a positive effect on the consumer value. We believe that the brand we use, namely Coca-Cola, will have a positive effect on its value because of its positive associations (Du-Jian Gang et al., 2012). However, we do believe that this relationship may be affected by factors, which we de-scribe as ‘neuromarketing’ and ‘uncertainty avoidance’. We expect that the use of neuro-marketing will have a negative impact on the CBBE of a brand because of the lack of trust in marketing in general as we described earlier. Hence the negative effect of this moderat-ing relationship between neuromarketmoderat-ing and CBBE. However, we do believe this effect to be less strong when a highly valued brand (Forbes, 2014) uses this method, than when a lesser valued brand uses this marketing tool. Furthermore, this model also investigates the moderating role of Hofstede’s (1980) uncertainty avoidance on the moderating relationship of neuromarketing. We believe that consumers that have a high level of uncertainty avoid-ance will also be more reluctant to purchase brands’ products when they deem the brands marketing practices to be unethical. We choose this moderator because of its highest im-pact on consumer trust in case of unethical behavior by a firm (Leonidou et al., 2012). A total of 5 hypotheses have been constructed which we describe in the following section. During the study, we provide our participants with a case wherein neuromarketing will be

explained and also will be placed in a situation where it is used by a brand for the design of their new advertising campaigns.

2.3.1 Main Hypothesis

As we described in the previous section, the amount of value a product of a known brand has above the same product without the brand name, is described by brand equity (Aaker & Biel, 1992). Keller (1993) looked at this concept from the consumers’ point of view and described this as CBBE. He notes that a product from a well known brand, with positive associations, will have a higher CBBE than the same product without the brand name. As we noted in the introduction, and in light of the Coca-Cola experiment (Du Jian Gang et al., 2012), Coca-Cola is a brand with positive associations. These associations, we believe, will have a more positive effect on a drink of Coca-Cola than on the same drink with Pepsi-Cola as a brand. As it has been concluded by the experiment by Du Jian Gang et al.

(2012), Pepsi-Cola has a lower CBBE than Coca-Cola. Therefore, we present the following hypotheses:

figure 6. Conceptual model

CBBE

+

-

+

H1a: Pepsi-Cola will have a lower CBBE than Coca-Cola

H1b: Coca-Cola will have a higher CBBE than Pepsi-Cola

2.3.2 Moderating hypotheses

Leonidou et al. (2012) noted that there are several factors that result in a positive relation-ship between consumers and a brand. These are ‘trust’, ‘satisfaction’ and ‘loyalty’. They conclude that when consumers trust a brand, they will be more likely to be satisfied with the brand and thus loyal to it. Consumers that are loyal tend to reinforce the relationship they have built with a brand, rather than risking a new relationship which might bring un-certainties (Macintosh et al., 1992). Trust plays a big role in a brands’ performance, com-pany loyalty, purchase intention and market performance (Chaudhuri & Holbrook, 2001; Metzler et al., 2008), although Dick and Basu (1994) do conclude that satisfaction does not lead to repurchase all the time. Anyhow, there is evidence that in case of good experi-ences with a firm, consumers tend to increase repurchase behavior and commitment (An-tón et al., 2007). Thus, it could be stated, that when consumers are satisfied with a brand, they most likely trust the brand and its actions. A known brand like Coca-Cola seems to have consumer trust, as it is one of the worlds most valuable brands (Forbes, 2014; Sta-tista, 2014), and has almost fifty percent of the global market share of carbonated bever-ages in 2015 (Statista, 2015). Following the findings by Leonidou et al. (2012), it can be believed that consumers will have positive associations with Coca-Cola, and thus be satis-fied and trust the brand and its actions. Also, Hustvedt and Kang (2013) noted that building consumer trust requires transparency by the firm. They state that when a firm is transpar-ent about their actions, this can result in an inttranspar-ention to purchase their products and spread positive word of mouth. We believe that the knowledge of Coca-Cola using neuro-marketing will not cause a significant decline in its CBBE because of the satisfaction, trust

and loyalty for the brand that consumers already seem to possess. Our hypotheses are therefore:

H2a: The use of neuromarketing by the brand Coca-Cola will not affect its CBBE

H2b: The CBBE of Coca-Cola will be higher than Pepsi-Cola if neuromarketing is

used

According to the notions made above, consumers that have less positive associations with a brand, will not have the same trust in that brand as they do with a brand like Coca-Cola. Therefore it is possible that Pepsi-Cola, with less positive associations (Du Jian Gang et al., 2012), will generate more feelings of distrust in their marketing actions, as consumers do not seem to trust marketing in general (Heath & Heath, 2008; Pollay & Mittal, 1993). This is mostly caused by consumers’ suspicion of marketing deception. As Fisher et al. (2010) note, neuromarketing has manipulative potential, and can thus lead to a higher sense of distrust within consumers. The lack of trust in marketing and no experience with an unknown brand may cause greater distrust in that brand. We therefore state the hy-potheses:

H3a: The use of neuromarketing by the brand Pepsi-Cola will have a negative im

pact on its CBBE, than when this tool is not used

H3b: The CBBE of Pepsi-Cola will be lower than Coca-Cola if neuromarketing is

2.3.3 Conditional hypothesis

Neuromarketing being a relatively new marketing method, its use and implications are not always straightforward and uncertain at times according to research, as mentioned in the literature review. Also no research on consumer response to this method has been con-ducted. Therefore we believe consumer knowledge on neuromarketing will unlikely be de-tailed, and for them will mostly be an unknown area. According to Hofstede (1997), people differ when it comes down to ambiguous or unknown situations. Some people are more comfortable than others with situations that require coping with uncertain moments in their lives. This is one of Hofstede’s (1980) cultural dimensions, namely uncertainty avoidance. Individuals that have a high degree of uncertainty avoidance tend to feel more uncomfort-able in situations they cannot control. These individuals feel more at peace in environ-ments that are controlled and are structured in the way they know (Vitell et al., 2003). On the other hand, individuals that tend to score low on uncertainty avoidance are more com-fortable taking risks, feel more comfort in situations that are not part of a routine and tend to be more flexible than their counterparts (Yoo & Donthu, 2002). Both the latter authors and Vitell et al. (2003) confirmed that ethical behaviors by a firm and uncertainty avoid-ance correlate positively. When people with high uncertainty avoidavoid-ance perceive a firms practices to be unethical, this lowers their trust in the firm (Leonidou et al., 2012). Thus it can be stated that individuals in high uncertainty avoidance are more reluctant to purchase products from companies they deem to behave untrustworthy. As we mentioned earlier, research showed that consumers seem perceive marketing as untrustworthy. If this is the case, and consumers believe neuromarketing to be unethical, we believe that individuals with high uncertainty avoidance will respond more negatively to a brand when this brand uses neuromarketing as a method. Therefore, our last hypothesis is:

H4: The negative moderating effect of neuromarketing on the relationship between

brand and CBBE will be higher in individuals with a high level of uncertainty avoid ance than individuals with lower levels of uncertainty avoidance.

. 3 . Research method

In this chapter we describe our research methodology. In the first section our research sample is described after which, in the second section, we present our instrumentation, research design and measurement. The third section provides the procedure and how the data has been collected.

3.1 Research sample

For this thesis, the population was taken from the Cola consumer group. Coca-Cola’s main target group is mostly the younger generation (12 to 18 years), but they also focus on young adults with their Coca-Cola Zero and Light campaigns (cocacolaneder-land.nl, 2015). For our research we seek a group that has more knowledge on marketing, and thus for this reason, we believe that the young adult group is most representative. In the Netherlands there are 4.1 million people between the ages 20 and 40, which covers the largest part of the young adult group (CBS, 2014). For this study a non-probability sampling method was used, because no sampling frame could be achieved for such a large population. The study took place mostly among students from the University of Ams-terdam, most of which came from the master Business Administration. There are around 2000 students participating in the master, and according to the power analysis from www.-surveysystem.com/sscalc.htm, around 330 responses were needed.

In this study, we assigned the participants randomly into four different groups. These were named the control, brand, neuromarketing and interaction. To measure the effect of neu-romarketing and its practices (conditions) on the CBBE of Coca-Cola, we compared the conditions’ effect on CBBE with each other. The groups used in this experiment are dis-played in Table 1. Because we wanted to measure if neuromarketing and its practices caused a decline in CBBE, we believe that the experimental design would be the best method for this study. This is because of random assignment to the conditions, the control group and the manipulation of variables. Statistical data was acquired through surveys filled out by the participants, because surveys form a formal method to test hypotheses statistically in quantitative research. The surveys were filled out at one point in time, and therefore this makes this study cross-sectional.

3.2.1 Research design

For the study, all participants filled out the questionnaire measuring CBBE, which also in-cluded demographic questions and which can be found in Appendix 1. In addition to the brand equity questionnaire, the participants in the brand and interaction condition also filled out items that measured attitude, ownership, purchase intention and experience with the brand (Yoo & Donthu, 2001). This was done in order to measure if participants are fa-miliar with our chosen brands and if they have experience with the products. Also, we wanted to account for any positive bias towards Coca-Cola, as we expected the brand to

Table 1: Experimental Groups

Brand used

Neuromarketing used

No Yes

No Control Brand

be positively rated by most of the participants (according to the Coca-Cola experiment mentioned earlier and Coca-Cola’s status as a brand). The items on purchase intention, attitude and involvement of the brand were placed after CBBE. The reason for placing atti-tude and intention after brand equity, was to reduce the halo effect that is common in mul-tiattribute attitude models, “in which subjects distort their perceptions when expressing their overall attitudes before they evaluate details that contribute to the attitudes” (Yoo & Donthu, 2001, p. 5).

Firstly, participants in the control group read a case on a product of

‘Pepsi-Cola’ (Wikipedia, 2015; Appendix 3) which sells a similar carbonated soft drink as Coca-Cola, which was followed by the questionnaire on CBBE. Participants in the brand condi-tion had a similar procedure, except in their essay on Pepsi-Cola was replaced by an es-say on Coca-Cola (Appendix 2). Secondly, participants in the neuromarketing and interac-tion condiinterac-tions also read an essay on a product of a brand (respectively Pepsi-Cola and Coca-Cola), and this was also followed by an essay on neuromarketing and the fact that their brand uses this tool for the design of their latest marketing campaigns (Appendix 4). Moreover, these latter conditions also filled out the questionnaire on CBBE, and lastly a questionnaire on Uncertainty Avoidance as well (Appendix 1).

3.2.2 Measurement

In this study, two sets of questionnaires were used to measure CBBE and uncertainty avoidance. The items of these questionnaires were measured through, respectively, a five-point and seven-five-point Likert scale. Apart from these items, the concepts of brand experi-ence and brand usage/ownership were measured by yes or no items. Brand experiexperi-ence was measured by “Have you ever bought Coca-Cola?” and brand usage/ownership by “Do you currently use/own Coca-Cola?”. Purchase intention was measured by the items “I would like to buy Coca-Cola” and “I intend to purchase Coca-Cola” on a five-point scale.

Attitude was measured by five-item scales of “very bad/very good”, “very nice/very awful”, “very attractive/very unattractive”, “very desirable/very undesirable” and “very likeable/very unlikable”. Lastly, brand involvement was measured by five-point items: “I am very in-volved with Coca-Cola”, “I drink Coca-Cola very often”, “I am a Coca-Cola expert” and “I am not interested in Coca-Cola” (Yoo & Donthu, 2001).

As Keller (1993) only provided the ingredients for CBBE, but has not defined a clear scale on how the concept should be measured. Therefore we borrowed the scale from Yoo and Donthu (2001) for our first questionnaire on CBBE. The authors comprised a multidi-mensional scale for the measurement of this concept based on extensive research on the subject. The questionnaire consists of four variables, namely ‘brand loyalty’, ‘brand aware-ness and associations’, ‘perceived quality’ and ‘overall brand equity’. Brand loyalty was described as the intention to buy the brand as a primary choice (Yoo and Donthu, 2001) and consists of three items. Perceived quality is the subjective evaluation by consumers on product quality (Zeithaml, 1988, p. 3), and consists of two items. Brand awareness is “the ability for a buyer to recognise or recall that a brand is a member of a certain product category” (Aaker, 1991, p. 61), and is measured by two items. Brand associations is de-fined as everything that the memory links to a brand (Aaker, 1991), which can be strength-ened by experiences, which has three items. Overall brand equity measures the brand eq-uity in general, which is assessed by four items. The items are measured by a five-point Likert scale, ranging from 1 = “strongly disagree” to 5 = “strongly agree”. The items cho-sen all showed a Cronbach’s alpha higher than 0.70 (Yoo & Donthu, 2001). A higher score on all four variables for a brand, would imply a higher CBBE for that brand.

The second questionnaire on uncertainty avoidance, has been taken from Hofst-ede's (1980) cultural dimensions. This questionnaire measures the amount of risk avoid-ance or risk seeking behavior within participants. The measurement of this variable is measured through four items, which also had a Cronbach’s alpha higher than 0.70. These

items were measured by a seven-point Likert scale, ranging from 1 = strongly disagree to 7 = strongly agree.

3.3 Procedure

The data for this research was collected through an online survey. A pilot study was con-ducted to account for any inconsistencies in the research design, after which the main study has been done.

3.3.1 Pilot study

A pilot study was conducted among five respondents, four male and one female, prior to the research on the 7th of May 2015. This was done in order to account for any inconsis-tencies and inconveniences that may occur during the research, thus their results were not included in the main research. The respondents were mostly personal acquaintances of the researcher, which differ from education and lifestyle. Following the pilot study, we re-moved the items on ‘product involvement’, as this seemed not to be of value to the results. Furthermore, we initially used ‘Brand X’ to compare it to Coca-Cola on their CBBE, but we noticed that the participants associated a ‘cola’ product directly to Coca-Cola. To account for this occurrence, we changed Brand X to ‘Pepsi-Cola’, as to create more contrast be-tween the brands, which is also a replication of the Coca-Cola experiment (Du Jian Gang et al., 2012) mentioned before. Afterwards, we also adjusted our hypotheses according to this change. SPSS 22.0 for Mac was used to measure reliabilities.

3.3.2 Main Study

Qualtrics.com was used to distribute the online survey, and the data was collected through this website. On the 8th of May 2015 the survey was distributed to the researcher’s per-sonal network through social media (Facebook) and through e-mailing the Business

Stud-ies student group. Also, the respondents were asked to spread the survey in their personal network. Therefore, the snowball technique helped to gather more respondents than the researcher could reach himself. The survey was online until the 21st of May 2015. No ex-act number of surveys spread could be noted, as it was unclear how many surveys were spread among the personal networks of the researchers acquaintances.

. 4 . Results

A total of 205 responses were returned after thirteen days. Of these responses, 53 partici-pants did not fill out more than 50% of the questionnaire. Therefore, these 53 were dis-carded from the data. The total number of participants consisted was 152, male (n = 93) and female (n = 59) with a mean age of 29. Three items from the questionnaire were re-coded, namely “I have difficulty in imagining Coca-Cola/Pepsi-Cola in my mind”, “Uncer-tainty is a normal feature of life and each day must be accepted as it comes” and “Fear of

ambiguous situations and of unfamiliar risks is normal” because these were

counter-in-dicative.

4.1 Reliability

Reliable items are items that measure the variable that they should measure. Therefore, for the items in used in our research, a reliability analysis was conducted. A way to mea-sure the reliability of items is through the Cronbach’s alpha. When the alpha is >.70, the item can then be presumed as reliable. Table 2 shows the reliability analysis, where the reliability of the variables is displayed per brand category. As can be seen below, most variables were reliable, as the alpha is above .70. However, for the variables Perceived Quality of Coca-Cola, Brand Awareness/Associations of Coca-Cola, Purchase Intention of Coca-Cola and Uncertainty Avoidance, the scale did not reach above the threshold.

Firstly, the variable Perceived Quality of Coca-Cola reached an alpha of .60. This indicates that the items within the scale are not internally consistent enough for this con-struct to be reliable. Deleting items within the variable would not affect the measurement in any way (probably because this variable consisted of only two items). However, this vari-able was needed for testing the hypotheses. Therefore it was kept for further analysis. The same also goes for Brand Awareness/Associations of Coca-Cola. The alpha of this vari-able does not reach a positive score, namely a -.31 alpha. However, deleting the item “Some characteristics of Coca-Cola come to my mind quickly” raises the alpha to .69. As the score increased significantly by deleting the item, this item was discarded from the variable. Furthermore, Purchase Intention of Coca-Cola scored .04 on reliability. As this variable was merely to check for positive bias of Coca-Cola, it was not needed for the de-fined hypotheses. Therefore it was removed from the research, and is not mentioned in Table 2 below. Because we will not be able to compare Coca-Cola with Pepsi-Cola on Purchase Intention, we have also removed this variable for Pepsi-Cola.

The alpha for Uncertainty Avoidance did not reach a positive score, namely -.56. This means that the internal consistency was negative between the items. The results showed that these items were negatively correlated with each other. This suggests that this variable does not measure what it should measure. This could mean that the partici-pants, for some reason, did not answer consistently on the the scale items. However, while this variable and its items were unreliable, it was needed for one of the hypotheses. Never-theless, as we we used this variable in our analysis, we cannot say anything conclusive about the measures.

4.2 Correlation Check

Pearson’s correlation coefficients were calculated to measure if variables are related. Ta-bles 3 and 4 below show these correlations. It can be seen that mostly the correlations

from the Pepsi-Cola measures were significant (p<0.05), but only six from the Coca-Cola measurements had reached significance. For both groups (Pepsi-Cola and Coca-Cola), the measurement of Uncertainty Avoidance was not related to any of the other variables. This is also in line with the negative Cronbach’s alpha that was described for Uncertainty Avoidance in the previous section.

Table 2: Reliability of scales*

Variable N of Items Cronbach’s Alpha*

Brand Loyalty Pepsi-Cola 3 .79

Perceived Quality Pepsi-Cola 2 .71

Brand Awareness/Associations Pepsi-Cola

5 .75

Overall Brand Equity Pepsi-Cola 4 .89

Brand Loyalty Coca-Cola 3 .82

Perceived Quality Coca-Cola 2 .60

Brand Awareness/Associations Coca-Cola

5 .69

Overall Brand Equity Coca-Cola 4 .87

Attitude Pepsi-Cola 5 .95

Attitude Coca-Cola 5 .88

Uncertainty Avoidance 4 -.56

4.3 Model Testing

This research was based upon testing wether Coca-Cola would have a higher CBBE than Pepsi-Cola (hypothesis 1a and 1b), if the knowledge of the use of neuromarketing by a brand would negatively affect its CBBE (hypothesis 2a and 3a) and if Coca-Cola would have a higher CBBE than Pepsi-Cola if those brands used neuromarketing as a tool (hy-pothesis 2b and 3b). Furthermore, we tested if the different levels of Uncertainty

Avoid-Table 3: Means, Standard Deviations, Correlations and Cronbach’s Alpha of Pepsi-Cola related questions

Nr Variable Mean SD 1 2 3 4 5 6 7 1 Loyalty 1,71 ,69 (,79) 2 Quality 3,09 ,87 ,448** (,71) 3 Awareness/ Associations 3,96 ,68 ,09 ,240* (,75) 4 Overall Equity 2,32 ,86 ,530** ,465** ,244* (,89) 5 Attitude 1,59 ,45 ,517** ,681** ,279* ,492** (,95) 6 Uncertainty Avoidance 3,95 ,48 ,04 -,12 ,22 ,09 -,12 (-,56) *p<0.05, **p<0.01 (one-tailed)

Table 4: Means, Standard Deviations, Correlations and Cronbach’s Alpha of Coca-Cola related questions

Nr Variable Mean SD 1 2 3 4 5 6 1 Loyalty 3,15 1,12 (,82) 2 Quality 3,71 ,71 ,630** (,60) 3 Awareness/ Associations 4,06 2,31 -,15 -,14 (,69) 4 Overall Equity 3,10 ,95 ,486** ,419** -,06 (,87) 5 Attitude 1,85 ,29 ,376** ,567** -,03 ,222* (,88) 6 Uncertainty Avoidance 3,95 ,48 -,00 ,08 ,14 -,05 ,03 (-,56) *p<0.05, **p<0.01 (one-tailed)

ance would affect the impact of neuromarketing on CBBE (hypothesis 4). All tests were tested at a significance level of p<.05, and thus every result above that level was rejected.

For the the study, the participants were first divided into four groups, namely the ex-perimental groups mentioned before. Furthermore, the mean scores of the CBBE con-structs were calculated and can be found in Table 5 below. To compare the means of the CBBE variables between the experimental groups, a One-Way ANOVA was used to test the first four hypotheses (1a, 1b, 2a and 2b). This method requires a few assumptions to be tested, namely normality, homoscedasticity and independence of the variables.

The normality check showed that most variables were normally distributed, besides the variable Brand Attitude. The distribution of the data for the variable Brand Awareness/ Associations was slightly platykurtic and Brand Loyalty was slightly leptocurtic. As normali-ty does not have a strong effect on the Type 1 error, which states that the null hypothesis will unjustly get rejected, normality of the variables was accepted.

Also homoscedasticity was checked for the variables. Most variables were ho-moscedastic, besides the variables Brand Loyalty and Brand Attitude. The Levene’s statis-tic for these latter variables was significant, which implies heteroscedasstatis-ticity. However, the F-test is robust enough for violation of this assumption, thus this does not impact the re-sults greatly.

The independence of variables was also tested. This resulted in a intraclass correla-tion coefficient of .24, and thus a low independence between the variables. This low inde-pendence has a major effect on the Type 1 error. This means that the chance, that the null hypothesis will get rejected while it should be accepted, increases dramatically. Therefore this can have a significant impact on the results. However, the analysis was still done.

Hypothesis 1a, 1b, 2a, 2b, 3a, and 3b were tested using a one-way ANOVA, paring the means of the different experimental groups. This is the most viable test to com-pare the different scores of CBBE variables between the experimental groups. As noted

above, table 5 shows these mean scores on the variables by the different groups. The dif-ferent mean scores by the groups is also shown in Figure 7 below.

Table 5: Mean scores on the Brand Equity variables

Experimental Groups

Score on the different variables of Brand Equity: Mean (SE)

Loyalty Quality Awareness/ Associations Overall Equity Control (n=35) 1,88 (0,80) 3,17 (0,98) 3,89 (0,71) 2,37 (0,84) Brand (n=41) 3,08 (0,96) 3,63 (0,72) 4,30 (0,48) 3,18 (0,93) Neuromarketing (n=38) 1,56 (0,53) 3,01 (0,76) 4,03 (0,65) 2,26 (0,88) Interaction (n=38) 3,23 (1,27) 3,80 (0,70) 3,30 (0,53) 3,01 (0,98) Notes:

1. Participants in the Control and Neuromarketing group answered Pepsi-Cola related questions,

Brand and Interaction group answered Coca-Cola related questions

2. Responses were on a scale from 1: Strongly Disagree, 5: Strongly Agree

figure 7. Mean scores on Brand Equity variables by experimental groups (error bars represent +/- 2SD)

The results showed significant differences between the groups for the variables Brand Loyalty, F (3, 148) = 30.99, p < .05, Perceived Quality, F (3, 148) = 8.48, p < .05, and Overall Brand Equity, F (3, 148) = 9.72, p < .05. However, for the variable Brand Aware-ness/Associations, no statistically significant differences were found between the experi-mental conditions, F (3, 148) = .60, p = .62.

To compare the mean scores on the variables between the single groups (Brand - Control; Interaction - Neuromarketing; Neuromarketing - Control; Interaction - Brand), a Tukey post-hoc test was analyzed. This test revealed a statistically significant higher score for Brand Loyalty for the Brand condition compared to the Control condition (p = .00) and for the Interaction condition condition compared to the Neuromarketing condition (p = .00). However it did not reveal a statistically significant difference between the Neuromarketing condition and Control condition (p = .48), and also not between Interaction condition and the Brand condition (p = .90). For Brand Quality, the test showed statistically significant higher scores for the condition Interaction compared to Neuromarketing (p = .00). Between the conditions Brand and Control, Neuromarketing and Control, and Interaction and Brand no statistically significant differences were found (respectively p = .06; p = .83; p = .78). As expected, the variable Brand Awareness/Associations did not show any statistically signifi-cant differences between the experimental groups (Brand - Control, p = .74; Interaction - Neuromarketing, p = .95; Neuromarketing - Control, p = .99; Interaction - Brand, p = .60). For the variable Overall Brand Equity, statistically significant difference were found in higher scores for Brand condition compared to Control condition (p = .00) and Interaction compared to Neuromarketing (p = .00). No statistical significant differences were found be-tween Neuromarketing and Control groups (p = .96) and Interaction and Brand groups (p = .85). The results show that the Brand condition scores higher than the Control condition on two of the four variables. Therefore hypothesis 1a and 1b are partly supported. The Tukey post-hoc test did not reveal a lower score on CBBE in the Interaction condition compared

to the Brand condition. This supports hypothesis 2a. However, also no statistically signifi-cant difference was found between the Neuromarketing and Control condition. Therefore hypothesis 3a is rejected. Moreover, the above statistics do show some evidence of a higher score of the Interaction condition compared to the Neuromarketing condition in three of the four variables. Therefore hypothesis 2b and 3b are partially supported.

An interesting result was found when comparing the Interaction with the Control condition. In three of the four variables, namely Loyalty, Quality and Overall Brand Equity, the Interaction condition scored statistically significantly higher than the Control condition (respectively p = .00; p = .01; p = .02). These results would also partially support our hy-potheses by the fact that Coca-Cola has a higher CBBE than Pepsi-Cola, even when the former brand uses neuromarketing. Also, the Brand condition scores higher on the previ-ously mentioned variables compared to the Neuromarketing condition (respectively p = . 00; p = .00; p = .00).

To test for the moderating effect of Uncertainty Avoidance, the PROCESS test by Hayes (2013) was used to test the amount of variance explained by the variable (signifi-cance of ∆R2_{). Table 10 below shows these results. As can be seen, Uncertainty}

Avoid-ance does not seem to explain any of the variAvoid-ance in the CBBE variables. Therefore, hy-pothesis 4 was rejected.

Accounting for the expected positive bias for Coca-Cola, we measured the amount of experience and usage/ownership of Coca-Cola and Pepsi-Cola. All the participants that

Table 10: Effect size of the moderator Uncertainty Avoidance on the Brand Equity constructs

Construct R2-change F DF1 DF2 p Brand Loyalty ,000 ,014 1 72 ,906 Brand Quality ,008 ,741 1 72 ,392 Brand Awareness/ Associations ,004 ,279 1 72 ,599

filled in the Coca-Cola related questionnaire have previously bought the drink (n = 79). Of the participants that filled in the Pepsi-Cola questions, the majority claimed to have bought the drink before (n = 66) and a small part claimed that they haven’t (n = 7). Furthermore, the majority of Coca-Cola participants claimed to currently use Coca-Cola (n = 42) and the minority claimed that they did not (n = 37). For Pepsi-Cola it was the other way around (respectively n = 11; n = 62). Moreover, comparing the conditioned groups by the variable Attitude, the difference between the Brand and Control condition showed a statistically

significant difference (p = .00). Therefore, this result shows a more positive Brand Attitude for Coca-Cola than for Pepsi-Cola.

. 5 . Discussion

The previously mentioned Coca-Cola experiment by Du-Jian et al. (2012) showed that consumers that compare Coca-Cola with Pepsi-Cola are more positively biased towards Coca-Cola because of its positive associations with the brand. The results in the current study showed a similar bias towards Cola. For this reason, we also expected Coca-Cola to have a reasonably higher CBBE (Keller, 1993) than Pepsi-Coca-Cola. This however was only partly supported by the results, as only the half of the variables on the CBBE survey were answered more positively for Coca-Cola than for Pepsi-Cola. The other half of the results were not statistically significant. Therefore some evidence has been found for a stronger positive bias towards Coca-Cola, in line with previous research.

Research by Heath & Heath (2008) showed that marketing lacks consumer ap-proval and trust (Sheth, Sisodia & Barbulescu, 2006). Because neuromarketing is a rela-tively new marketing method, no research has been conducted on the approval of this method among consumers. Therefore, our current research has tried to measure the effect neuromarketing would have on the CBBE of a brand. The results showed no significant

results containing a different CBBE for the brand that used neuromarketing compared to the same brand that did not use this method. However, the results of this research did show a slight negative difference between Pepsi-Cola that used neuromarketing and Pep-si-Cola that did not use this marketing method. This does suggest a slight negative re-sponse to neuromarketing if a brand does not have strong positive associations. Although this notion is suggested, no causal effect could be concluded, as the results were not sta-tistically significant. Moreover, the results showed no difference in the CBBE of Coca-Cola when the brand used neuromarketing, compared to when it did not use this tool. This con-firmed our hypothesis that a brand with strong positive associations is not strongly affected by its practices. Nevertheless, comparing the scores on CBBE while a brand used neuro-marketing, these results showed a more positive CBBE for Coca-Cola than for Pepsi-Cola. But these results were only partly supported by the analysis, as differences on three out of four CBBE variables were significant. This, again, partly supports the notion of a more pos-itive bias towards Coca-Cola than for Pepsi-Cola.

In their research, Leonidou et al. (2012) have concluded that consumers can re-spond differently to unethical behaviors by companies compared to other consumers. This depends on their score on Hofstede’s (1980) cultural dimensions. Uncertainty avoidance was one of these dimensions that had the strongest moderating effect on unethical behav-iors by firms. Consumers that scored highly on this dimension, also perceived a firms un-ethical behavior as a valid reason to stop the repurchase of the firms products. As market-ing mostly triggered a negative response within consumers (Gaski, 2008), because of its lack of trust, we believed that this could be seen as unethical behavior by consumers that scored highly on uncertainty avoidance. Because of the high potential of deception and manipulation by neuromarketing (Morin, 2011), we expected neuromarketing to generate a stronger negative response for individuals with a high uncertainty avoidance score. How-ever, the results showed no moderating relationship of uncertainty avoidance on the CBBE