The Effect of Green Marketing on Consumer Brand Perception, and the Moderating Role of Culture

(1)

The Effect of Green Marketing on Consumer Brand Perception, and

the Moderating Role of Culture

Including an empirical study on skincare brand advertisements

Dissertation for the DDM Advanced Marketing and International Business Management

Lise Smit C200348140

S2666226

Newcastle University Business School Dr Qionglei Yu

(2)

2

Abstract

The purpose of this study is to investigate the effect Green Marketing has on Consumer Brand Perception (CBP). Moreover, the moderating effect Culture may have on this relationship will be assessed. The literature background in the first half forms the basis of the relationships that will be investigated. Next, the CVSCALE (Yoo, Donthu and

Lenartowicz, 2011) is used to measure Hofstede’s (1980, 2001) Cultural Values of individual respondents during the online survey. Resulting, the finding confirm Green Marketing

(3)

3

1. Introduction

Buying and consuming products are a key component of modern-day life. This is due to the plethora of options offered, providing people the chance of surrounding oneself with the things they desire. However, these increasing consumption rates lead to dwindling resources, and numerous studies suggest this accelerates climate change and will ultimately destroy (natural) resources worldwide. (Bendell, 2018) Rising awareness of these issues lead to increasing environmental responsibility among companies (Chen et al., 2016) and consumers (Chin et al., 2018) in the form of Sustainable Marketing and Sustainable Consumption, respectively. However, this research will primarily focus on consumers by investigating the effect Green Marketing has on Consumer Brand Perception, i.e. a primary precursor of deliberate consumption. (Brunk, 2010a)

Besides the environmental motivations to invest in Green Marketing are economic incentives, due to potential improvement of firm performance. (e.g., Benabou and Tirole, 2010; Choi and Wang, 2009; Fraj-Andrés, Martinez-Salinas and Matute-Vallej, 2009; Waddock and Graves, 1997) Moreover, Hur et al. (2014) suggest there is a relation between Green Marketing and Consumer (Brand) Perceptions (Bhattacharya and Sen, 2004; Du et al., 2010; Heikkurinen, 2010; Maignan and Ferrell, 2001; Melo and Garrido-Morgado, 2012; Roberts and Dowling, 2002), which will be further examined in this research.

Both marketing and consumer preferences are very context-specific, meaning inferences are dependent on the chosen industry as well as the country. Therefore, assessing Culture’s influence on Consumer Brand Perception is valuable for both national and international companies. More specifically, it may contribute to a better understanding of how to market a brand within one specific country. Additionally, it also highlights the importance of adapting this (national) strategy when crossing borders.

Prior research indicates Culture affects Consumer Behavior (Kire and Rajkumar, 2017; Lowe and Corkindale, 1998; Lynn, Zinkhan and Harris, 1993) due to the norms and values that are rooted in distinct societal groups, despite globalization. (Zhu, Quan and Xuan, 2006)

(7)

7

Nowadays there is a wide variety of national culture theories (e.g., House et al., 2004; Inglehart, 1997, 2000; Ghemawat, 2001; Schwartz, 1994b, 2004) available that may help researchers investigate this concept. Most of them are (at least partially) based on the precursor, Geert Hofstede and his Cultural Dimensions. (1980, 2001) Therefore, Hofstede’s Cultural Dimensions will be used to study Culture.

After the literature background, an empirical research will follow in order to collect primary data about Green Marketing, Consumer Brand Perception and Culture. The scope of this empirical study will be limited to the skincare segment of the beauty – or cosmetics – industry. Since the three main concepts are rather context-specific, focusing on a particular industry is useful in order to gather relevant insights. The global cosmetics market is considerable in size and has shown annual growth since 2004. (Ridder, 2020) Skincare makes up 40% of this market (Statista, 2020), which makes it a relevant segment to focus on. Moreover, skincare products are often particularly subjected to pertinent issues within Green Marketing. Greenwashing, for instance, concerns falsely overstating a product’s ‘greenness’ or sustainability and thereby misleading consumers in an attempt to manipulate their perceptions. (Hayder, 2017; Choudhary and Gokarn, 2013)Recent studies investigating current challenges in Green Marketing and Green Consumption (Chin et al., 2018), emphasize the relevance of these phenomena in the skincare market. Therefore, specifically focusing on skincare brands in the empirical research will presumably yield valuable insights on these matters.

The purpose of this study is to further investigate whether Green Marketing, in fact, positively influence Consumer Brand Perception. Additionally, how Hofstede’s Cultural

Dimensions (1980, 2001) moderate this relationship will be assessed by applying the CVSCALE. (Yoo, Donthu and Lenartowicz, 2011) More specifically, Power Distance, Collectivism and Long-Term Orientation predicted to strengthen the influence Green Marketing has on Consumer Brand Perception. Contrarily, Masculinity and Uncertainty Avoidance are expected to have a weaking, moderating effect.

(8)

8

2. Literature Background

This literature background will provide an overview of previous research concerning the main subjects of Sustainable Green Marketing, Consumer Brand Perception, Culture and its related areas.

2.0 Sustainability

Sustainability can be defined as ‘the consumption of goods and services that meet

basic needs and quality of life without jeopardizing the needs of future generations’ (OECD, 2002). Due to the growing urgency of addressing climate change, sustainable practices become increasingly embedded in our lives. (Choudhary and Gokarn, 2013) Consequently, questions regarding accountability about these environmental concerns arise. However, this research will primarily address responsibility from a business perspective.

According to the Business Development Matrix (Edward and Tallontire, 2009) undertaking sustainable practices as part of ‘societal development engagement’ may stem from two opposing incentives and resulting outcomes. Supporters of the interdependent

approach (e.g., Blowfield, 2005a; Frederick, 2006; Freeman, 2017; Scherer and Palazzo,

2007) claim business operations and societal problems are not independent and can therefore not be treated as such. (Edward and Tallontire, 2009)

Contrarily, instrumental relationship proponents (Friedman, 1988; Porter and Kramer, 2006; Whelan, 2012) claim business responsibilities are limited to optimizing their own efficiency and effectiveness, (Edward and Tallontire, 2009), i.e., ‘economic responsibility’. (Carroll, 1979, 1991) However, consumer attention towards to the ‘ethical reputation of firms’ (Garcia de los Salmones et al., 2009) as well as their preference for ‘socially

responsible companies’ (Maignan and Ferrell, 2001) increases along with their expectations of both companies and their products. (Mulki and Jaramillo, 2011) Therefore, engaging in Sustainable Marketing may, in fact, enhance firm performance. (e.g.,Benabou and Tirole, 2010; Choi and Wang, 2009; Fraj-Andrés, Martinez- Salinas and Matute-Vallej, 2009; Waddock and Graves, 1997) Moreover, the rising sales of Green Products imply consumers’ willingness to buy such products is growing. (Chen, 2008b) Therefore, Sustainable Marketing could be embedded in economic responsibility after all.

(9)

9

Figure 2.0: ‘Societal development engagement’ according to the Business Development

Matrix (Edward and Tallontire, 2009)

Instrumental Integrative Interdependent

Low | High

In addition to the varying rationales of businesses itself, the growing interest in sustainable development (Bandura, 2007; Gordon, 2011) in relation to environmental impact and climate change is visible in politics as well, i.e., ‘legal responsibility’. (Carroll 1979, 1991) For instance, the Sustainable Development Goals (SDGs) established by The United Nations in 2015. These goals address current global challenges by proposing seventeen specific incentives to be attained in 2030 in order to ‘achieve a better and more sustainable future for all.’ (United Nations, 2020) Appendix B displays all SDGs and highlights the goals that indirectly (6, 7, 11) and directly (12-15) address the environment and climate change. Since the UN chose to dedicate seven out of the seventeen SDGs concerns to environmental issues, the global relevance of these problems is highlighted. Moreover, since the SDGs are developed by the UN, legal responsibility is linked to political policies that may ensure businesses will, in fact, operate in more sustainable ways. Consequently, engaging in sustainable practices is increasingly important for businesses, as seen from several levels of the CSR Pyramid. (Carroll 1979, 1991)

2.1 Sustainable Marketing

Marketing plays an increasingly important role in modern life and can be defined as

‘(…) the activity, set of institutions, and processes for creating, communicating, delivering, and exchanging offerings that have value for customers, clients, partners, and society at large.’ (American Marketing Association, 2017)

Subsequently, Sustainable Marketing is ‘about understanding and managing

marketing’s pivotal role in the future of business and society (…) is the process of creating, communicating and delivering value to customers in such a way that both natural and human capital are preserved or enhanced throughout.’ (Martin and Schouten, 2012, p. 10) Moreover, it ‘(…) seeks a solution in which commercial goods can be marketed in a responsible way that does not adversely impact upon sustainability’. (Gordon et al. 2011, p. 147)

(10)

10

Early definitions of Sustainable Marketing resonate with the Triple Bottom Line (Elkington, 1994), which entails Profit, Planet and People – also known as the three Es: Equity, Economic and Environment. The Triple Bottom Line – TBL from now on – requires the satisfaction of ‘organizational goals’ (profit) and ‘customer needs’ (people), while making sure ‘(…) the process is compatible with ecosystems’ (planet). (Fuller, 1999; Gordon, 2011) Thereafter, Emery (2012) added ‘economically fair’, ‘socially equitable’ and

‘environmentally friendly’.

Regardless of semantics, the three components remain unchanged for all theories. Therefore, this research will refer to Elkington’s (1994) original concepts of Profit, Planet and People. However, it was Elkington (2018) himself who revised the TBL 25 years after introducing it. He believes sustainability is overwhelmingly measured in terms of profits or losses while no alternatives to measure human and environmental wellbeing are offered. Consequently, an overemphasis is placed on Profit, continuously prioritizing it over Planet and People, which is not what the TBL was intended for. This could translate to the practice of Greenwashing, which will be discussed in section 2.1.1. Elkington (2018) thinks numerous businesses are missing the point of the TBL when merely using it as an accounting tool rather than a concept that provokes thoughts on how to add social and environmental value as well.

Seemingly, The United Nations’ (1987) Sustainable Development precedes Sustainable Marketing, as it strives towards Sustainable Consumption and sustainable economic growth while protecting the environment. Choudhary and Gokarn (2013) link the phenomenon of Sustainable Development to Sustainable Consumption. The first concerns ‘maintaining long-term economic, social and environmental capital’ (p. 30) while the latter refers to utilizing resources in ways that leads to minimal environmental costs and supporting human wellbeing simultaneously. (Choudhary and Gokarn, 2013) This idea indicates they are codependent as development and consumption cannot exist without one another.

(11)

11

In addition to its embeddedness in Sustainable Marketing, Green Marketing is (also) used interchangeably with ‘Environmental Marketing’. (Fraj-Andrés, Martinez-Salinas and Matute-Vallej, 2009) This type of marketing, in turn, is directly linked to the Planet

(Elkington, 1994) component of the TBL discussed earlier. Figure 2.1 below depicts how Green Marketing overlaps with both Sustainable Marketing and Environmental Marketing, as suggested by Pride and Ferrell. (1993) Due to its holistic focus as well as the environmental incentives of this research, Green Marketing will be further discussed below.

Figure 2.1: Venn Diagram depicting Sustainable, Green and Environmental Marketing

2.1.1 Green Marketing

In 1992 Charter defined Green Marketing as ‘a holistic and responsible strategic management process that identifies, anticipates, satisfies and fulfils stakeholder needs, for a reasonable reward, that does not adversely affect human or natural environmental well-being.’ (Choudhary and Gokarn, 2013, p. 27) Notably, this definition is quite similar to that of Sustainable Marketing, which might explain why the terms are often used interchangeably. In the following years Pride and Ferrell (1993) referred to Green Marketing as ‘an

organization's efforts at designing, promoting, pricing and distributing products that will not harm the environment.’ (Choudhary and Gokarn, 2013, p. 27) Lastly, Polanski (1994) described it as ‘all activities designed to generate and facilitate any exchanges intended to satisfy human needs or wants (...) with minimal detrimental impact on the natural

environment.’ (Choudhary and Gokarn, 2013, p. 27) Though the various Green Marketing

(12)

12

definitions highlight different aspects, they emphasize its essence, environmental and human well-being.

In his 2001 article Peattie states the evolution of Green Marketing consists of three sequential phases, namely the Ecological phase, the Environmental phase and the Sustainable phase. Whereas the first phase concerned solving immediate environmental problems and the second surrounded implementation of cleaner technologies (Choudhary and Gokarn, 2013), the third and currently prevailing phase is seemingly a more holistic and proactive approach. This includes better regulation in terms of pollution and energy-efficient production for new

and current products. (Dono, 2010) In other words, the current Green Marketing phase

emphasizes one’s active stance towards limiting environmental impact rather than merely solving issues resulting from earlier problems. Consequently, in the remainder of this study, Green Marketing is defined as ´the marketing of green products/services that applies and optimizes sustainable practices throughout the entire process in order to limit any harm to the environment and human resources.’

(13)

13 2.2 Brand Perceptions and Consumer Behavior

Brand Perception is an extensively researched subject as it precedes Consumer Behavior (Romaniuk and Sharp, 2003) – which may ensue sales and profits. This results in the billions of dollars researchers and companies spend on research that aims to ‘identify important factors that influence on consumer decisions.’ (Mirabi, Akbariyeh and

Tahmasebifard, 2015, p. 267) Hence, Brand Perceptions presumably provide useful insights for marketeers when deciding upon taking risks that might improve their marketing and financial performance. (Choi et al., 2007; Chow and Holden, 1997) Additionally, studies identified (Brand) Perception as a key indicator of Consumer Behavior. (Faisal-E-Alam, 2020; Romaniuk and Sharp, 2003) This is partially based on the Theory of Planned Behavior (Ajzen, 1985, 1987), which elaborates upon the link between perceptions and behavior. Due to the lack of existing literature linking Green Marketing, Brand Perceptions and Culture, Consumer Behavior and its related concepts will be discussed to bridge the gap.

Consumer Behavior can be defined as ‘the decision process and physical activity

which the individuals engage in evaluating, acquiring, using or disposing of goods and services.’ (Loudon and Della Bitta, 1984) Furthermore, it refers to analyzing habits in terms of choices, purchases and consumption – or any motives refraining from this. (Solomon, 2009) This includes studying the links between feelings, thoughts and circumstances, that ultimately result in actual behavior. (Faisal-E-Alam, 2020)

During all product phases – research, design, production, marketing, etc. – consumers remain central as they are targeted to buy and consume the product. Hence, determining why consumers would or would not choose one product brand over the other enables marketeers to predict and influence buying behavior. This is where Brand Perception provide value about the why, which cannot be derived from Consumer (buying) Behavior alone. Since these personal wishes are often implicit and mostly inscrutable, brand marketeers continuously search for ways to fathom these.

2.2.1 The Behavioral Sequence

The Theory of Reasoned Action (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975) and its successive Theory of Planned Behavior (Ajzen, 1985, 1987) are useful models in explaining the relationship between human attitudes and actual behavior from a

(14)

14 Figure 2.2.1: The Behavioral Sequence

First, Beliefs consist of behavioral, normative and control beliefs (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975), but will not be further discussed in this study as they require a more psychological research context. Notably, however, the Beliefs in this model resonate with the beliefs (Brunk, 2010a) and consumer values (Kim and Chung, 2011) that are identified as indicators of attitude, Purchase Intention and buying behavior.

Second, Attitudes are the result of (un)favorable evaluations of the subject in question. (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975) Subsequently, Brand Attitudes can be defined as consumer’s overall brand evaluation (Wilkie, 1986), which may consist of Salient Believes and Evaluative Judgements. Salient Beliefs refer to ‘the extent to which consumers think the brand has certain attributes or benefits. These beliefs are assessed as positive or negative in ‘evaluative judgements. (Keller, 1993) Several attributes and benefits will be applied in the primary research of this study. (section 4.3.1)

Third, Behavioral Intention (intention in short) concerns the intention to perform certain behavior, in which the intention itself reflects how strong the desire is and/or how much effort the individual is willing to put in it. (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975) Generally, ‘the stronger the intention to engage in a behavior, the more likely should be its performance.’ (Ajzen, 1991) Following, ‘Purchase Intentions are an individual’s conscious plan to make an effort to purchase a brand.’ (Spears and Singh, 2004, p. 56)

Fourth and last, Behavior refers to the actual performance of certain behavior, which is directly related to Consumer Behavior in this context. One may assume the marketeers’ incentive is to cultivate favorable evaluations and attitudes, and steer consumers towards purchase intention (Kim and Chung, 2011; Spears and Singh, 2004) and ultimately Consumer (buying) Behavior. (Brunk, 2010a)

Belief

Attitude

Behavioral

(15)

15 2.2.2 Consumer Brand Perception

Keller (1998) describes a Brand as ‘(…) a product, but one that adds other dimensions that differentiates it in some way from other products designed to satisfy the same needs. (p. 3) Subsequently, Brand Perception(s) can be defined as ‘the attitudes, perspectives, and views consumers hold toward a Brand.’ (Guthrie, Kim and Jung, 2008, p. 166)

More specifically, CBP will be used to determine what the effect of Green Marketing will be, as seen from the consumer. Prior studies indicate Green Marketing positively

influences Consumer Behavior due growing environmental concern. (Choudhary and Gokarn, 2013) This includes consumers’ ‘less hostile’ behavior patterns of buying more eco-friendly products and services that have detrimental effects on the environment. (Choudhary and Gokarn, 2013; Jain and Kaur, 2004) Moreover, since Chen (2010) found Green Brand Image positively influences Green Brand Equity, incentives are in place to investigate whether the same applies to Green Marketing and CBP. Due to the priorly established link between perceptions and behavior, one may predict Green Marketing may contribute to enhanced CBP as well.

2.3 Culture as Moderator

Much research (e.g., Foscht et al., 2008; Kim and Bae, 2016), has been done to investigate the effect culture has on Consumer Brand Perceptions (CBP) and it can be said that culture’s role cannot be ignored when looking for in-depth inferences regarding Consumer Behavior and its related concepts. (Fan and Xiao, 1998; Lowe and Corkindale, 1998; Lynn, Zinkhan and Harris, 1993; Sproles and Kendall, 1986) For instance, the role culture plays in shaping how consumers perceive Brand Identity (Aaker, 1996), is

considerable. (Kushwah et al., 2019) Similarly, cultural differences influence the way in which consumers perceive Aaker’s (1997) Brand Personality. (Aaker and Maheswaran, 1997) Consequently, how culture is expected to influence Consumer Brand Perception will be discussed in section 3. But first, more attention will be paid to what culture is,

Culture can be defined as a complex system consisting of shared beliefs, values,

symbols and customs (Van Gelder, 2003), which are passed down from each generation to the next. (Trompenaars, 1994) These components form the underlying framework that define and guide individuals’ behavior and their subsequent interactions. (Johansson, 2000)

(16)

16

Another widely applied definition of culture is ‘the collective programming of the mind that distinguishes the members of one group or category of people from others.’ (Hofstede, 1991, p. 5) These distinguishing groups or collectives may include – but are not limited to – organizations, ethnic groups, gender, class, religious groups, teams, etc. Notably, the social and subjective norms embedded in the Belief component of the Theory of Planned

Behavior (Fishbein and Ajzen, 1975) are based on culture-specific settings, which indicates

Culture and Consumer Behavior are intrinsicallyrelated. Within a business context, however, nations form the most frequent level of analysis as this enables cross-country comparisons and related inferences. The fact that laws, politics, payment units, language(s), but also holidays, and certain uses are often – but not always – homogenous within a country, makes countries relevant units of analysis of comparison when companies investigate their

opportunities. Yet, this research will be conducted in a non-country specific cross-cultural context since the focus is on culture and its dimensions rather than on certain nationalities and the differences among them. Nationality, however, will be used as a control variable to check whether strong correlation between two or three of the main concepts is present due to respondents’ nationalities, to avoid oversimplified or incorrect conclusions. (section 4.3.4)

2.3.1 National Culture Measures

There are various theories that enable one to measure and compare national culture. Due to culture’s complex and ambiguous nature, no ideal tool exists as it is impossible to capture everything a specific culture entail. Consequently, researchers introduced theories that enabled them to measure specific elements – also known as (Cultural) Dimensions.

In 1980 Geert Hofstede conducted his study of National Culture and introduced his four Cultural Dimensions - Power Distance, Individualism, Masculinity and Uncertainty Avoidance – after which he added Long-Term Orientation as the fifth one in 1991. Based on Minkov’s (2010) World Value Survey data, Hofstede introduced Indulgence (versus

Restraint) as the sixth and final Cultural Dimension. However, due to the current scarcity of data on Indulgence, it is not included in this research.

Inspired by Hofstede’s (1980, 2001) framework, the similar theories of the Cultural

Distance Index (Kogut and Singh, 1988), Schwartz’s (1994b, 2004) Cultural Value

Dimensions, Inglehart’s (1990, 1997, 2000) two dimensions of cross-cultural world variation,

(17)

17

seemingly less than Hofstede’s. Thus, its applicability to a wide range of studies around the world indicates Hofstede’s theory is the most suitable for this research, too.

2.3.2 Hofstede’s Cultural Dimensions

Despite receiving ample criticism over the years (Kirkman, Lowe and Gibson, 2006, 2017) Hofstede’s Cultural Dimensions (1980, 2001) continues to be the most important theory in a cross-cultural context. (Chandy and Williams, 1994; Søndergaard, 1994) Essentially, frequent criticism concern its face validity (labelling issues), its lack of

psychometric properties (Spector, Cooper and Sparks, 2001) and lack of dynamics

(Baskerville, 2003; McSweeney, 2002; Søndergaard, 1994) as well as the fact that the original theory is merely based on IBM employees. However, the Cultural Dimensions (CDs from now on) owe its continuous relevance, among other things, to its full cover of major cultural conceptualization (Clark, 1990) and its empirical development – as opposed to the conceptualization stages of others. (Yoo, Donthu and Lenartowicz, 2011) Moreover, since the CD scores depict national culture in relation to other countries, studies indicate that country culture changes, but their positions do not. (Inglehart and Baker, 2000; Beugelsdijk and Welzel, 2018) While countries may show increased or decreased (‘absolute’) scores when compared to previous years, their relative cultural positions remain stable over time. In other words, national cultures do change, but all in the same direction (Beugelsdijk, 2019;

Beugelsdijk and Welzel, 2018), i.e., parallel. (Inglehart, 1997) Thus, Hofstede’s (1980, 2001) CDs remain relevant and will be applied to this study.

Moreover, Culture influences CBP of Brand Identity (Kushwah et al., 2019), which refers to the specific image marketeers aimed to convey to consumers. There is, however, no research that investigates Green Marketing, Consumer Brand Perception and Culture

altogether. More specifically, Culture has not been examined as a moderator rather than an independent variable in this context. Therefore, the moderating effect culture may have on the relationship between Green Marketing and Consumer Brand Perception remains unknown and requires investigation.

Previously, Leonidou et al. (2013) found that Hofstede’s CDs can predict how

consumers perceive ‘unethical marketing’ through the traits of idealism and egoism. Whereas Power Distance and Uncertainty Avoidance leads to idealism, which is positively associated with unethical marketing perception, Individualism and Masculinity contribute towards

(18)

18

2013) Though (perceived) unethical marketing itself is not being examined in this study, understanding how specific CDs may lead to increased awareness and negative perceptions towards this behavior, is useful. Moreover, Malik and Singhal (2017) found Collectivism and Long-Term orientation positively affect ‘Consumer Environmental Attitude’ and the resulting ‘Willingness to Purchase Environmentally Friendly Products’. As opposed to their research, specific claims about the environment itself will not be made here, in order to assess

subconscious perceptions.

2.4 The Cosmetics Industry

Since the empirical phase of this study will use cosmetic products to find answers on the research questions, this section provides a background of the cosmetics industry. In addition to its current multimillion sales numbers, the cosmetics industry has been expanding for several years Statista Research Department, 2020) and is predicted to continue doing so in the foreseeable future. (Data Bridge Market Research, 2019) This continuous growth is partially caused by the increasing number of both – offline and online – options for buying cosmetics. (Dai and Pelton, 2018) Its ‘high competitive density’ (Dai and Pelton, 2018, p. 269) makes it a relevant industry as competition often leads to a continuous demand for new marketing strategies. Moreover, beauty companies are subject to recessions and other crises due to the non-essential nature of their products. However, 2003 NPD Reports show that 38% of the survey respondents would spend less money on perfume and fragrances during crises while this number is 23% when it concerns skincare/make-up and toiletries (Kumar, Massie and Dumonceaux, 2006), inferring skincare is less dependent on the economy than other segments within the beauty industry.

(19)

19 2.4.1 The Skincare Segment

Cosmetics have been around for thousands of years and though people might automatically think about makeup, the industry is considerably more diverse than this. (Kumar, Massie and Dumonceaux, 2006) There are various segments within the industry – that may vary per country – the focus will be on skincare products because they make up 40% of the market. (Ridder, 2020)

(20)

20

3. Hypotheses

Following the previous sections, prior literature indicates there is a positive relationship between Green Marketing and Consumer Brand Perception. Moreover, since Consumer Brand Perception (CBP) is affected by Culture (Foscht, et al., 2008), its

moderating effect will be assessed. The Cultural Value Scale (CVSCALE) from Yoo, Donthu and Lenartowicz (2011) will be used to measure Hofstede’s (1980, 2001) dimensions of national culture on the individual level. Consequently, Uncertainty Avoidance and Long-Term Orientation are expected to weaken the relationship between Green Marketing and CBP while Power Distance and Collectivism are predicted to strengthen it.

Hence, the research questions are, what effect does Green Marketing have on Consumer Brand Perception – and how does Culture influence this relationship? Consequently, the hypotheses are developed as follows.

3.1 Hypotheses Development

Prior research showed Green Marketing – and its overlapping concept of Sustainable Marketing – positively influences Consumer Behavior. (e.g., Du et al., 2010;

Garrido-Morgado, 2012) Moreover, the key predictor of Consumer Behavior is said to be Consumer Brand Perception. (Faisal-E-Alam, 2020; Romaniuk and Sharp, 2003) This is underlined by the Theory of Planned Behavior (Ajzen, 1985, 1987), which explains how perceptions may lead to behavior. Thus, the first hypothesis is as follows.

H1: Green Marketing positively influences Consumer Brand Perception

Thereafter, the CVSCALE (Yoo, Donthu and Lenartowicz, 2011) will be used to assess the moderating effect Hofstede’s Cultural Values (1980, 2001) have on the relationship between Green Marketing and Consumer Brand Perception. The previously discussed effect the Cultural Values may have on idealism and egoism (Leonidou et al, 2013) will be used to predict its influence on each of the Values.

(21)

21

idealistic people tend to ‘do right’ (Singhapakdi et al., 1999) and are inclined to follow higher moral duties and standards (Forsyth, 1980), positive perceptions of Green Marketing can be expected. Granted, however, no dishonest (marketing) practices are in place (Nebenzahl et al., 2001) – i.e., Greenwashing (section 2.1.1). Moreover, studies (Lu, Rose and Blodgett, 1999; Singhapakdi et al., 1999; Vitell, Nwachukwu, and Barnes, 1993) researching how Culture influences ‘ethical marketing decision-making’ indicate people with high Power Distance tendencies ‘usually adhere to strict deontological norms’. (Leonidou et al, 2013, p. 542) These norms might translate to enhanced CBP when people are aware the (potential) benefits it may have. Therefore, an intensifying moderating effect of Power Distance on CBP resulting from Green Marketing is predicted.

Second, Individualism (vs. Collectivism) concerns ‘the degree of interdependence a society maintains among its members’ (Hofstede Insights, 2021), which is a continuous tradeoff between the ‘self’ and the group. Individualism is expected to have a neutral or negative effect on Green Marketing due to its link with egoism. (Leonidou et al., 2013) This is because individualistic people question higher ethical standards and are less likely to take society’s needs into consideration (Singhapakdi et al., 1999) Moving to the other side of the Individualism spectrum, Yoo and Donthu (2002) revealed Collectivism is positively

associated with ‘perceived marketing ethicality’ because this tendency inclines to build harmony among related groups. (Vitell, Nwachukwu, and Barnes, 1993) The same applies to Malik and Singhal’s (2017) ‘Consumer’s Environmental Attitude’ and its associated

‘Willingness to Purchase Environmentally Friendly Products’ a result of Collectivism. Since the CVSCALE uses ‘Collectivism’ rather than Individualism, this term will be used in the remainder of the study. Consequently, Collectivism is expected to have an intensifying effect onCBP resulting from Green Marketing.

Third, Masculinity (vs. Femininity) is about the motivation that drives people. (Hofstede, 1980) Whereas competition and success are seen as masculine traits, feminine values include ‘caring for others and quality of life.’ (Hofstede Insights, 2021) Masculinity is expected to have a negative effect of Green Marketing through the tendency of egoism. (Leonidou et al., 2013) Studies suggest masculine people are less likely to conform to ethical standards (Vitell, Nwachukwu, and Barnes, 1993; Yoo and Donthu, 2002) Similarly,

(22)

22

Fourth, Uncertainty Avoidance centers around the idea that we cannot know the future, and how we tend to cope with this ambiguity. (Hofstede, 1980, 2001) Though Uncertainty Avoidance was initially predicted to evoke idealistic behavior, opposite results were found. (Leonidou et al., 2013) This was based on the Uncertainty Avoidance positive association with ‘perceived marketing ethicality’, as identified by Yoo and Donthu (2002).

The suggested relationship between Uncertainty Avoidance and egoism ‘implies that uncertainty-averse individuals might sacrifice their idealistic values and codes, in an attempt to safeguard their own security, prosperity, and well-being.’ (Leonidou et al., 2013, p. 542) Therefore, a buffering moderating effect of Uncertainty Avoidance on the relationship between Green Marketing and CBP is expected.

Fifth, Long-Term Orientation relates to the tradeoff between maintaining long-held traditions and introducing new uses to prepare for the future. (Hofstede Insights, 2021) Though Long-Term Orientation has not (yet) been investigated by Leonidou et al. (2013), it is presumably linked to the trait of egoism. Consequently, negative associations with Green Marketing might be expected, unless personal benefits are identified. Contrarily, Malik and Singhal (2017) found Long-Term Orientation positively influences ‘Consumer

Environmental Attitude’, which may increase ‘Willingness to Purchase Environmentally Friendly Products’. Therefore, an intensifying moderating effect of Long-Term Orientation on the relationship between Green Marketing and CBP is predicted.

H2a: Power Distance strengthens the relationship between Green Marketing and Consumer Brand Perception

H2b: Collectivism strengthens the relationship between Green Marketing and Consumer Brand Perception

H2c: Masculinity weakens the relationship between Green Marketing and Consumer Brand Perception

(23)

23

H2e: Long-Term Orientation strengthens the relationship between Green Marketing and Consumer Brand Perception

Figure 3.1: Conceptual Model

(24)

24

4. Methodology

The aim of this research is to test two main hypotheses based on the literature

background. Namely, how does Green Marketing affect Consumer Brand Perception (CBP), and how does Culture influence this relationship? Yoo, Donthu and Lenartowicz’s (2011) CVSCALE is used to measure Hofstede’s (1980, 2001) Cultural Dimensions on the

individual level. Additionally, the skincare segment of the beauty industry is used as scope, by showing two similar face creams advertisements in the survey.

4.1 Research Strategy

This study follows a positivist paradigm because it is built upon previously conducted scientific research, of which generalizability in a different environment will be examined on the basis of quantifiable data. Moreover, the assumptions of the researcher being both objective and independent of the study, hold. (Taylor and Medina, 2013) Due to a lack of available data on the influence Culture has on the relationship between Green Marketing and CBP, primary data collection was required to conduct the research.

The (statistical) analyses will mainly be performed in SPSS version 26. However, since Structural Equation Modelling as well as testing this model through Confirmatory Factor Analysis is not possible in SPSS itself, STATA version 16 will be used for these steps. Both programs are accessed through the Universal Windows Platform (UWP) of the

University of Groningen.

4.2 Research Design and Procedures

Within the research model Green Marketing is set as the independent variable and will be used to study the effect on CBP, the dependent variable. Additionally, Cultural Values (CVs) are expected to have a moderating effect on the relationship between Green Marketing and CBP. This moderating effect could be strengthening or weakening, depending on the CV. (section 3.1) Lastly, several demographical details will be collected and serve as

control variables (Blumberg, Cooper and Schindler, 2014) Justification for the specific

(25)

25 4.2.1 Data Collection Methods

An online survey was designed to collect primary, quantitative data through a publication on Qualtrics. This method is chosen because it is time efficient and eliminates any geographical limitations in terms of time and distance. Moreover, it enables scalability due to the standardized nature of the survey. All the questions are in multiple-choice format and are addressed to one individual per response. The survey contains a Neutral or Green skincare brand advertisement and is followed with questions about CBP, Cultural Values and demographic details. The survey can be found in Appendix D.

4.2.2 Sampling Approach

The research population consists of consumers of eighteen years and older, and therefore, the same applies to the sample (frame). Since the study is not focused on specific countries or comparing different ones, no conditions for nationalities are in place. The only other prerequisite is that the participant possesses adequate English skills to complete the survey while comprehending the topics discussed sufficiently. Moreover, it is suggested to ensure the inclusion of participants with various cultural background, genders, ages, incomes, etc. to avoid sampling bias (errors), which is essential to make sure the sample is

representative of the population. (Simundic, 2013) Furthermore, though simple random sampling is not attainable in this context, the non-probability technique of convenience sampling is applied. (Hancock, 2018) The open-access online survey URL is shared among

social media and university-related platforms, which lead to a minor snow-ball effect among acquaintances.

Moreover, there are several methods to calculate the minimal sample size but the minimum ‘twenty times the main variables’ rule of thumb (Mundfrom, Shaw and Ke, 2009) leads to at least 140 respondents. Moreover, a minimum of 150 is required to conduct Factor Analysis, according to the ‘five to one’ rule (DeVellis, 2012; Tinsley and Tinsley, 1987) since the highest number of questions within one construct is 30 for Brand Associations.

4.3 Measurement

The three variables of the conceptual model (3.1) as well as demographic control

variables are included in the online survey. The Likert scale is used because it is a

(26)

26

Wright, 1967) Though applying a 5-point Likert scale may be associated with ‘lower frustration levels’ (Babakus and Mangold, 1992) and better comprehensibility (Marton-Williams, 1986), this study will use a 7-point Likert scale because of its higher reliability with t-tests (Lewis, 1993) and increased ‘power to detect incorrect substantive models’. (Maydeu-Olivares, Fairchild and Hall, 2017)

Resulting, the three main variables are measured by questions accompanied by answers on a 7-point Likert scale, where 1 represents ‘strongly disagree’ and 7 ‘strongly agree.’ The control variables are collected through categorical questions that are nominal, ordinal or interval in nature.

The online survey consists of two conditions, the ‘Neutral’ and the ‘Green’ condition, to which each participant is assigned randomly. More specifically, the Green group will see a skincare brand advertisement that includes Green Marketing while the neutral group is presented a brand advertisement without this. Consequently, results from the Neutral and Green group will be compared in order to assess the effect Green Marketing has, since all other aspects of the survey are exactly the same.

4.3.1 Green Marketing of Skincare Brands

(27)

27

Figure 4.3.1: Two skincare advertisements – small versions

Brand A (Neutral) Brand B (Green)

Green Marketing is incorporated in Brand B’s marketing techniques as well as in the ‘green’ nature of the product itself. More specifically, the advertisement states Brand B is classified as ‘ecofriendly’, vegan, ‘made from recycled packaging’ and that it contains

‘natural ingredients’. However, this image merely depicts the initial advertisement as it would be displayed on social media. This means users have to click on the advertisement to view more details about Brand A or B. Resulting, the advertisements are used to capture the initial

perceptions respondents prior to the decision to click or not click on them in order to access

more information. In other words, the empirical study concerns the Beliefs and Attitudes rather than the Behavioral Intentions and actions that may follow from them. By focusing on these first phases of the Theory of Planned Behavior (Ajzen, 1985, 1987), this study aims to gain in insights the effect Green Marketing may ultimately play in the implicit stages of Consumer Behavior. Therefore, no further attention is paid to the risk of Greenwashing (section 2.1.1) and degree of trust respondents have in the stated claims.

(28)

28

while buying preferences are usually more diverse, online shopping became increasingly relevant as the coronavirus made ‘traditional shopping’ more difficult or even impossible. (Coppola, 2021) Again, though not being able to feel, smell and try skincare products prior to online purchase (without visiting a store), the focus is on initial assessment of the products based on the advertisements alone. Additionally, an online advertisement may be especially accessible and practical during the coronavirus. Consequently, the products may feel more ‘tangible’, which might lead to more quantifiable answers.

Moreover, the rather general term of ‘social media (advertisement)’ is chosen as the setting for the fictional brands. Both companies and consumer increasingly use social media (Barnhil, 2011), which might indicate Green Marketing should be adapted to these digital platforms due to its growing relevance. Moreover, the emotional and affective responses that are predicted to result from Digital Marketing seem more suitable to this study than the rather analytical or rational reactions print media may incur. (Chaudhuri and Buch, 1995; Minton et al., 2012) More details on the chosen elements and design of this measurement can be found in Appendix C1.

Furthermore, respondents are asked to assess the skincare advertisements as if they were found on ‘a social media platform’. This setting is chosen because the growing social media usage – by consumers and marketeers (Barnhill, 2011) – and increasing relevance of Green Marketing (see above) require research in which the two subjects overlap. (Minton et al., 2012) Since Digital Marketing, Green Advertisement (Zinkhan and Carlson, 1995) and its related areas are not main concepts, but useful as a setting, the general term of ‘social media (advertisement)’ will be used.

4.3.2 Consumer Brand Perception

(29)

29 A. Brand Associations

Though Aaker’s (1991) ‘Brand Associations’ was originally built upon on prior experience with a brand, it will be used in this empirical study, nonetheless. This is justified by Keller’s (1993) notion that the – closely related – concept of Brand Knowledge essentially consists of accumulated Brand Association, combined with memory nodes. Since the first Brand Associations are based on some initial experience, one may argue this empirical study could form the basis of a new memory (node). More specifically, Brand Associations could be an indication of consumers’ very first encounter with the fictional brands used in this research. Alternatively, using real brands would have been accompanied by an extensive list of subconscious biases that may be hard to distinguish. The results from the fictional brands, however, could be applied to existing brands – with caution. An overview of the CBP-related concepts can be found in Appendix A.

Furthermore, Keller (1993) used Brand Knowledge as an indirect means of measuring Customer-Based Brand Equity, which is closely related to our concept of CBP.

‘Customer-Based Brand Equity is defined as the differential effect of Brand Knowledge on consumer

response to the marketing of the brand.’ (Keller, 1993) Again, this concept requires prior experience with a brand, which is captured by the Brand Awareness component of Brand Knowledge (see Appendix C2). However, Brand Awareness is irrelevant in this research and thus omitted. The concept of Brand Image, however, is related to the sought-after data for this research as it refers to ‘a set of associations’. (Keller, 1993) Consequently, the Brand

Associations properties of Attributes, Benefits and Attitudes, as well as their Favorability, Strength and Uniqueness will be included in the survey.

Lastly, Romaniuk and Sharp’s (2003) Image Attributes may prove to be a valuable addition to the measurement of Consumer Brand Perception. Image Attributes can be defined as ‘triggers’ that have the potential to induce consumers to buy the product based on the linkage of its features.

B. Purchase Intention

Similar to Yoo and Donthu (1997), the survey of this research includes questions about Purchase Intention and ‘Product Category Involvement’ (next section) to measure CBP. This is based on the practice of using Purchase Intention and Brand Attitude as surrogates of Brand Equity (Agarwal and Rao, 1996), which in turn, can be seen as a result of CBP.

(30)

30

(Park and Srinivasan, 1994; Rangaswamy et al., 1993; Swait et al., 1993), a strong relationship between the concepts is expected. (Yoo and Donthu, 1997)

Additionally, though the intention to buy a product is more ‘behavioral’ than a mere perception (section 2.2.1), it may provide additional insights regarding consumers’ overall judgement. Meaning, one may have a positive stance towards a product but still have no intention to purchase it as positive perceptions do not necessarily lead to buying behavior. Therefore, though respondents cannot actually buy the fictional skincare products, their answers will reflect whether they would (consider to) buy them.

4.3.3 Cultural Dimensions

As previously mentioned, Hofstede’s (1980, 2001) Cultural Dimensions consists of aggregated individual responses that lead to cultural scores on a national level. Simply using these national scores as is would lead to several problems, including the ecological fallacy, which refers to the application of these national scores to individuals (Hofstede, 2001), thereby ignoring the significant ‘inter-individual variations’ (Hoffmann, Mai and Cristescu, 2013) that are included on the country level. Though there is no ‘universally accepted’ scale to measure Hofstede’s cultural dimensions yet, Yoo and Shin (2017) suggest the Cultural

Value Scale or CVSCALE (Yoo, Donthu and Lenartowicz, 2011) in order to measure the

cultural dimensions on the individual level. The ‘adequate cross-cultural measurement equivalence’ (Schumann et al., 2012) and sufficient validity and reliability (Mazanec et al., 2015) shown by studies applying CVSCALE (Yoo and Shin) infers it is a suitable

measurement for this cross-cultural research as well.

Moreover, though Beugelsdijk and Welzel (2018) argue culture remains stable over time when measured in relation to one another, Yoo and Shin (2017) emphasize ‘(…) culture needs to be measured freshly at the moment of the research to precisely describe the culture of the exact participants for the study.’ Its enablement of these conditions as well as its adequate reliability, validity, exhibition of across-sample and across-national generalizability make CVSCALE an appropriate tool for a cross-cultural research (Djamen, Georges and Pernin, 2020; Yoo and Shin, 2017), like this one.

The questionnaire in Appendix C3 depicts the dimensions Power Distance,

(31)

31 4.3.4 Product Category Involvement

In addition to Brand Associations and Purchase Intention, Yoo and Donthu (1997) included several ‘items of other constructs’. Of these, ‘product category experience’, ‘usage and ownership of product category’ and ‘product category involvement’ will be added as a so-called ‘conditional variable’ named Product Category Involvement. Though respondents have never encountered the fictional brands before, they will presumably have some

experience with skincare products, which may provide additional insights regarding their shared associations. As Keller (1993) states, the degree to which consumer are familiar and involved with the product category affects their opinion of the proposed product as one cannot have a strong attitude towards a matter they do not care about.

4.4 Research Ethics

The survey contain assurance regarding both ethical and trustworthiness

(32)

32

5. Findings

This section comprises all the steps undertaken in terms of the original findings. First, the data is prepared by cleaning it and checking several assumptions. Next, Structural

Equation Modeling will be performed to assess the relationship between variables and reduce the number of variables. Finally, after all assumptions are tested, regression analyses will be conducted in order to answer the research questions in the next section.

5.1 Data Preparation

The data collection phase lasted two weeks and resulted in 231 completed survey responses. The incomplete responses will not be considered. The corresponding data was downloaded from Qualtrics and converted to a .xlsx file on January 20th. Next, all data irrelevant for assumption checking and performing statistical analyses, was deleted.

After importing the data to SPSS, missing values, names, labels and measurement levels were verified. Appendix E summarized all variables and their key details. The market (sub)variables are marked as ‘reversed’ because they were recoded (into the same variables) in order to ensure the direction corresponds with their meaning.

Next, the descriptive statistics showed that 60.6% of the respondents were female, 38.1% was male and 1.3% preferred not to disclose their gender. The vast majority is born in The Netherlands (79.2%) and/or currently resides here (82.3%). Since the second most frequent nationality (Germany and the USA) only contained six people, and many others merely consisted of one person, the nationalities are reduced to two dummies. Namely, ‘Current nationality: not NL’ and ‘Born nationality: not NL’. Moreover, the most frequent age groups are 25 – 34 years (42.9%) and 18 -24 years (33.8%). Other demographic

categories that are notable in size are respondents without children (81%), and those who are currently not in a relationship (34.2%). A complete overview of the descriptive statistics can be found in Appendix F.

5.2 Green Marketing Binary Variable

The independent variable, Green Marketing, consists of two mutually exclusive skincare advertisements in the survey. Therefore, Qualtrics marked all responses with a letter A (Brand A: Neutral) or B (Brand B: Green), depending on the randomly assigned

(33)

33

transformed into a (binary) dummy variable named ‘BrandB’ that indicate the respondents who saw the Green Marketing condition. Resulting, the Neutral condition is the reference group by default. The descriptive statistics show that the 47.6% of the respondents saw the Neutral brand advertisement and 52.4% saw the Green Marketing brand advertisement.

5.3 Consumer Brand Perception Factor Analyses

The dependent variable is Consumer Brand Perception, and consists of Brand Associations and Purchase Intentions. Both components will be discussed and analyzed separately.

5.3.1 Brand Associations

As previously mentioned in section 4.3.2 A, dependent variable (DV1.1) Brand Associations measures connotations consumers might have with the brand. This part of the survey consisted of 30 questions that are based on different levels of Keller’s (1993) Brand Image framework (section 2.2). Each question ranges between the values of 1 and 7, in which higher scores indicate a higher, positive value on the concept. Table 5.6a later in this section provides the descriptive statistics of this variable.

(34)

34 a. Exploratory Factor Analysis: PAF

Though the 30 sub-variables and their questions (Appendix C) aim to capture

different facets of Brand Associations, this number needs to be reduced drastically in order to perform regression analyses. Since ‘Brand Associations’ is the latent variable which is

merely theorized conceptually and there is no direct way of measuring it, Exploratory Factor

Analysis will be conducted. More specifically, Principal Axis Factoring (PAF) will assess

which questions, in fact, measure Brand Associations as inspired by Keller (1993). Ultimately, several questions will be combined into new variables (‘factors’) suitable for regression analyses.

Before conducting PAF, several assumptions will be checked. First, the 231 responses and 30 Brand Association questions lead to a 7.7 responses per item ratio, which meets the ‘five to one’ requirement (DeVellis, 2012; Tinsley and Tinsley, 1987) Second, the internal

consistency and discriminative power of Brand Associations are assessed by using

Cronbach’s Alpha. (Peter, 1979) Resulting, a Cronbach’s Alpha (.900) > .6 indicates

sufficient sampling adequacy. Though the Item-Total Statistics suggest deleting some of the items would increase the Cronbach’s Alpha to .901 or .909, these differences are negligible. Moreover, Samuels (2015) suggests checking bivariate correlation matrix for any correlation scores > .8 (Field, 2018) as they could be an indicator of multicollinearity. Since this is not the case, we can proceed.

As a condition of Factor Analysis itself, the Kaiser-Meyer-Olkin measure (KMO = .909) > .6 indicates sufficient sampling adequacy due to the proportion of variance.

Moreover, Bartlett’s Test of Sphericity (p = .000) is significant, which suggests the variables are suitable for structure detection. Lastly, sampling adequacy is confirmed by the scores on the diagonal in the anti-image correlation matrix, which are all > .5 – the lowest being .764 for Image Attributes item 2.

(35)

35

Consequently, the items that are deemed most relevant will be isolated for further PAF. Namely, Symbolic Benefits, Attitudes and Association Favorability are most relevant in terms of factor loadings as well as the theoretical explanatory value when comparing the Green and Neutral condition groups. Since ‘Promax’ rotation is often suggested as an

additional, non-orthogonal alternative for this type of research (Kahn, 2006; Worthington and Whittaker, 2006), this is applied as well. Appendix G1.3 confirms its similar factor loadings. Though three items of both Attribute Benefits and Image Attributes each load on one factor, they are less relevant in assessing the effect Green Marketing has on Brand

Associations – as a component of Consumer Brand Perception. Therefore, they will not be included in the PAF of DV1. They did, however, confirm respondents’ comprehension of the brands and indicate they take a predominantly positive stance towards the advertisements.

Moving forward, Symbolic Benefits, Attitudes and Association Favorability have an ‘excellent’ Cronbach’s Alpha (.914). (Salkind, 2015) Conducting PAF on the total of 12 items leads to a KMO (.925) > .6, indicating sampling adequacy. This results in a number of two factors with Eigenvalues (6.591 and 1.216) > 1.0 that explain 60.058% (51.927% + 9.131%) of the total variance. Since the Direct Oblimin and Promax rotations show identical divisions of factor loadings (G1.3), the two factors are saved as new variables by using the regression method.

b. Confirmatory Factor Analysis

Confirmatory Factor Analysis (CFA) will be performed on the 12 items of Symbolic

Benefits, Attitudes and Association Favorability to assess whether Structural Equation Modeling based on the two factors suggested by the previous PAF is suitable.

Appendix G1.3 shows the STATA Structural Equation Model in which the squares depict the items (‘observed variables’) and the ovals represent the factors (‘latent variables’). Since there were no missing values in the data, a ‘maximum likelihood method without standardized values’ (Gould, Pitblado and Sribney, 2006) was applied.

Several measures of the Overall goodness of fit-test in the post-estimation phase are used to assess the quality of the model. First, the Chi Squared test statistic (.000) is

(36)

36

‘acceptable fit’ is in place. (Hu and Bentler, 1999) However, since the pclose (.000) is not significant, we cannot reject that the population RMSEA is, in fact, smaller than or equal to .05., which would suggest a good fit. Third, both the comparative fit index (CFI = .925) and the Tucker-Lewis index (TLI = .906) are > .9, thus indicating a good fit of the model. (Carlson and Mulaik, 1993; Rigdon, 1996) Fourth and last, the standardized root mean squared residual (SRMR = .053) is < .08, which indicates an acceptable fit. (Hu and Bentler, 1999)

Though there are two indicators of a poor fit, one of an acceptable fit and two of a good fit, the proposed model is deemed as acceptable for three reasons. First, the CFI and TLI can be seen as the most relevant measures of the overall goodness of fit-test. (Alavi et al., 2020) Second, none of the insufficient scores deviate from their thresholds considerably and thereby, third, confirms the factors suggested by PAF. Consequently, two new variables are created by ‘factor scores for latent variables’ in prediction after estimation. More specifically, the new variables ‘Overall Brand Attitude’ and ‘Self-Expression Alignment with Brand’ are shown in Appendix G1.4 and will be used to measure Brand Associations.

5.3.2 Purchase Intention

Purchase Intention consists of five statements to which respondents could provide a score between 1 (‘strongly disagree’) and 7 (‘strongly agree’). Overall, respondents agreed the least (3.65) with ‘My purchase interest for Brand X is high.’ (‘PI_2’, std: 1.738). Contrarily, they disagreed the most with ‘I am certain I would not buy Brand X.’ (‘PI_3’), which was later reversed and lead to a mean of 4.81 (std: 3.022) The descriptive statistics are summarized in Appendix F3.

(37)

37

correlations, which are all > .5 – the lowest diagonal value being (.800) for item 1. Moreover, the Bartlett’s Test of Sphericity (P = .000) indicates that the correlation matrix is likely not an identity matrix, and therefore a factor analysis may be useful

a. Exploratory Factor Analysis

Following, Principal Component Analysis is conducted in order to reduce the number of variables while minimizing value loss. Resulting, the five items load on one factor with an Eigenvalue (3.604) > 1.0, explaining 72.077% of the total variance.

Subsequently, Confirmatory Factor Analysis will be performed in STATA by using Structural Equation Modeling. The ‘maximum likelihood method without standardized values’ was applied as there are no missing values. Figure G2.1 shows the connection between Purchase Intention factor (latent variable) and its five items (observed variables) resulting from questions.

Next, since the Overall goodness of fit-test indicates the CFA of a model with four items has a better fit than the initially proposed model with five questions, this alternative is chosen. First, though the statistically Chi Squared test statistic (.035) of this model infers poor model fit, it is closer to the saturated model than the initial model (.000) with five items. Table Appendix G2.2 displays the goodness of fit-test results for both models. Next, the root mean square error of approximation (RMSEA = .104) does neither infer a ‘close’ fit (< .05) (Browne and Cudeck, 1993; Jöreskog and Sörbom, 1993) nor a ‘reasonable’ fit. (< .08) (Bentler and Bonett, 1980) However, since the pclose (.104) is not < .05, the hypothesis that the population RMSEA is larger than .05 cannot be rejected, which may indicate a good fit. Moreover, both the comparative fit index (CFI = .993) and the Tucker-Lewis index (TLI = .980) are > .9 and thus indicate a close model fit is in place. (Carlson and Mulaik, 1993; Rigdon, 1996) Fourth and last, the standardized root mean squared residual (SRMR = .0014) is < .08, which indicates an acceptable fit. (Hu and Bentler, 1999)

5.4 CVSCALE Factor Analyses

(38)

38

respondents were asked how important (1 = ‘not important at all; 7 = extremely important) certain values were to them in order to measure Long-Term Orientation. Overall, people scored the lowest (means ranging from 1.49 to 2.35) on Power Distance, which concerns the attitude towards unequal power distribution. (Hofstede, 1980) Uncertainty Avoidance, on the other hand, showed the highest mean scores (ranging from 4.35 to 5.84), indicating

respondents have tendencies aimed towards preventing ambiguous and unforeseen

circumstances. (Hofstede, 2001) Appendix F4 summarizes the key descriptive statistics of the CVSCALE.

In order to reduce the number of variables, Factor Analyses will be conducted on the CVSCALE items. In accordance with similar research (Djamen et al., 2020), several

conditions have to be verified first. First, the ratio of observations to variables is nine to one, which is more than the required five to one requirement. (DeVellis, 2012; Tinsley and

Tinsley, 1987) Then, the Kaiser-Meyer-Olkin measure (KMO = .761) > .6 indicates sufficient sampling adequacy. Furthermore, Bartlett’s Test of Sphericity (p = .000) is significant, and thus a Factor Analysis may prove useful. Next, assessing the reliability and validity of the multi-items scale will ensure whether the CVSCALE items are ‘psychometrically sound measures’. (Churchill, 1979) More specifically, the internal consistency and discriminative power of the CVSCALE are assessed by using Cronbach’s Alpha. (Peter, 1979) A

(39)

39 a. Exploratory Factor Analysis

Similar to previous studies (Djamen et al., 2020; Prasongsukarn, 2009), the 25 CVSCALE items are subjected to a Principal Component EFA (Exploratory Factor Analysis). The scree plots of the independent (CV) scales indicate that each of the five dimensions only have one factor with an Eigenvalue > 1.0.

Contrarily, conducting Principal Component EFA on all 25 CVSCALE item. leads to seven factors with Eigenvalues > 1.0 that explain 59.996% of the total variance. The KMO is .761 and Bartlett’s test of Sphericity is (p = .000) significant, indicating a Factor Analysis may be useful. Appendix G3.2 shows most of the items load on their expected factor after applying orthogonal (Varimax) rotation, which is indicated by the values > .5. PD_1 and PD_2 do not load on any factor, while COL_6 and LTO_3 load on the extra factors, 6 and 7, respectively.

The EFA has been conducted numerous times in which different combinations of items were omitted to assess optimal factor loadings. PD_1, COL_6 and LTO_3 regularly do not load on their corresponding factors while MAS_4 often does not load on any factor whatsoever. However, when omitting ‘MAS_4’ and/or setting ‘5 factors’ as factor criterium, the other three items typically do load on their corresponding factor. Therefore, the EFA variations in which combinations of factor criteria, rotation methods (Djamen et al., 2020) and including/omitting MAS_4 are summarized in Appendix G3.2. Resulting, the EFA indicates most items load on their corresponding factor when omitting MAS_4, using the ‘5 factor’ criterium and applying oblique rotation. This would lead to five factors that explain 51.260% of the total variance. Though other combinations have higher variance, their factor loadings are illogical as less items load on their corresponding factor. However, the ultimate decision will be based on the CFA following below.

Appendix G3.3 depicts how the 25 items (all ‘observed variables’, without COL_4) are inserted in the Structural Equation Model, and connected to their corresponding factors (‘latent variable’). The double-headed arrows indicate covariance between the factors, as is the case with the CVs. Since all items of one CV should load on one factor, as suggested by EFA, a ‘1’ is inserted above the first items of each factor.

The Effect of Green Marketing on Consumer Brand Perception, and the Moderating Role of Culture