DOES IT MATTER WHERE IT IS FROM AND HOW IT IS MADE?
A CHOICE-BASED CONJOINT ANALYSIS ON THE EFFECTS OF COUNTRY-OF-ORIGIN AND LABELS ON PURCHASE BEHAVIOR
Master Thesis June 2019
MSc Marketing Management Faculty of Economics and Business University of Groningen, The Netherlands
Author:
E.M. (Emile) van der Veen Student number: 2546183 emilevdveen@gmail.com
First supervisor:
F. Eggers
Second supervisor:
B.J.W. Pennink
Abstract
The purpose of this paper is to develop a better understanding of how different product attributes affect purchase behavior. In particular, this research examines the effects of the country-of-origin cue and two types of labels, respectively, Fair Trade labels and organic labels.
In order to further investigate these effects, this study additionally investigates how these attributes interact and how these effects change when consumers shop for different consumption motives. Lastly, to gain insights into whether these effects differ per type of person, this study examines the role of personal values on the effects of earlier mentioned attributes on purchase intentions.
By conducting a choice-based conjoint analysis, this paper provides significant evidence for the effects of country-of-origin, Fair Trade labels, and organic labels on purchase behavior.
However, it does not find evidence that these attributes interact. Moreover, this study provides valuable insights on how altruistic, egoistic, biospheric, and hedonic personal value orientations (de)motivate Fair Trade or organic purchases and motivate the choice between different country-of-origins.
The findings of this paper provide relevant theoretical implications, as it contributes to the existing literature by providing empirical evidence on the effects of country-of-origin, Fair Trade labels, and organic labels on purchase behavior. Moreover, the results provide valuable managerial implications, as marketeers benefit by understanding how consumers use different informational cues in their purchase decisions.
Keywords: country-of-origin, country-of-origin effects, labeling, purchase intentions, personal
value orientations, hedonic shopping motives, utilitarian shopping motives.
3
Acknowledgments
I would like to thank Dr. Eggers and Dr. Pennink for their joint supervision on this thesis. From the very first start, we had a very interesting discussion about possible topics that would build upon my previous thesis. Since then, the support and constructive feedback have been very helpful to me.
Also, I would like to thank my friends and family for their support throughout the process. Your
suggestions and feedback have been very valuable and definitely elevated the quality of this
thesis.
4
Table of Contents
Abstract ... 2
Acknowledgments ... 3
1. Introduction ... 6
2. Theoretical background ... 9
2.1 Country-of-Origin (COO) ... 9
2.2 Labeling ... 11
2.2.1 Effects of labeling... 12
2.2.2 Fair Trade labels ... 13
2.2.3 Organic labels ... 14
2.3 The interaction effect between COO and labels ... 15
2.4 Personal values as motivators ... 16
2.5 Hedonic and utilitarian consumption ... 17
2.6 Price & willingness-to-pay ... 17
2.7 Conceptual framework ... 18
3. Methodology ... 19
3.1 Research design ... 19
3.2 Research context ... 19
3.3 Data collection ... 19
3.4 Survey development ... 20
3.4.1 Choice sets ... 20
3.5 Measurements ... 22
3.5.1 COO construct ... 22
3.5.2 Labels ... 22
3.5.3 Personal values ... 22
3.5.4 Hedonic and utilitarian shopping motives ... 23
3.5.5 Control variables ... 23
3.6 Sampling ... 23
3.7 Data analysis ... 24
4. Results ... 25
4.1 Pre-test ... 25
4.2 Sample characteristics ... 25
4.3 Data analysis ... 26
4.3.1 Model 1: attributes without interaction effects ... 26
4.3.2 Willingness-to-pay for attributes ... 27
4.3.3 Model 2: attributes with interaction effects ... 27
4.3.4 Model 3: interaction effects between attributes and personal values ... 28
4.3.5 Model 4: attributes without interactions (2-classes) ... 30
4.3.6 Model 5: attributes with personal value interactions (2 classes) ... 31
4.4 Hypotheses overview... 33
5. Discussion ... 35
5
5.1 Theoretical implications ... 35
5.1.1 Main effects of attributes ... 35
5.1.2 Interaction effects ... 35
5.1.3 Hedonic and utilitarian shopping motives ... 36
5.2 Managerial implications ... 37
5.3 Limitations & future research ... 37
6. Conclusion ... 39
7. References ... 40
8. Appendix ... 51
8.1 Appendix 1 ... 51
8.2 Appendix 2 ... 52
8.3 Appendix 3 ... 60
Word count: 10.365 (excluding appendices)
6
1. Introduction
Imagine that moment in the grocery store, where you have the opportunity to choose from dozens of different types of products from anywhere in the world. From buying your Brazilian coffee beans to having your favorite Italian spaghetti. Globalization has made it more and more easier to import and export products, a phenomenon reflected in the shift of commercial reality in terms of product-distribution, marketing, and management (Levitt, 1984). As multi-national sourcing and production become increasingly important, supply chains are being scattered around the world, becoming increasingly complex, but resulting in access to new products originating around the globe (Yip, 1995).
In what way consumers cope with products originating from a variety of countries is a rather interesting field of study. Numerous studies show that information regarding the country-of- origin (COO) of products has a strong effect on its perceived quality by consumers (Steenkamp, 1990). As a result of globalization, consumers are increasingly confronted with products originated from across the world. Despite the implications of globalization, it appears that the COO effect does not significantly change when a product is designed and manufactured in different countries. In fact, numerous scholars considered multi-national sourcing and production as the end of the COO effect, but it does not appear to diminish the relevance of the COO cue (Verlegh & Steenkamp, 1999).
Despite prior research, Insch & Cuthbert (2018) conclude that there is still limited knowledge
about how the COO construct affects consumer decision-making as a quality cue. In the context
of global production, Bilkey & Nes (1982) found that products from less developed countries
(LDCs) are generally evaluated less positively than products from more developed countries
(MDCs). In particular, products from LDCs are associated with inferior quality and an increased
risk of bad performance and dissatisfaction (Cordell, 1991). Developing countries have
outperformed the western world on prices, often at the expense of quality, proper working
conditions and fair wages. Although research on the effects of globalization in developing
countries reveals that globalization has positively influenced wages and non-wages working
conditions (World Bank, 2008), media still report about poor labor conditions in developing
countries (NY Times, 2018; The Guardian, 2018). Since prior studies are outdated, this study
will reexamine whether consumers are more like to purchase products from MDCs than from
LDCs.
7 To avoid buying products that violate consumers’ own ethical values, an increasing number of consumers tend to use eco- or ethical certifications in their purchase decisions, either in terms of product characteristics or the production methods that are employed (Annunziata, Ianuario,
& Pascale, 2011; Loureiro & Lotade, 2005; J. Roberts, 1995, 1996). To illustrate, consumers that engage in responsible consumption behavior may purchase organic goods or Fair Trade goods, as one of the motivations in the decision-making process can be based on ethical principles, in terms of ecological sustainability, animal welfare or ethical working practices (Ims & Jakobsen, 2006; Pino, Peluso, & Guido, 2012). Prior studies have predominantly focused on motivations to buy Fair trade or organic products; however, little evidence was found on the relative importance of a Fair Trade or organic label. In this research, the effect of Fair Trade labels and organic labels on purchase intentions will be evaluated and will be tested on relative importance compared to other attributes.
Given the globalization of the food industry and an increased demand in organic foods, Xie, Gao, Swisher & Zaho (2016) stress the importance of examining the interaction effect between organic labels and the COO cue. Here, the question arises whether the COO effect can be mitigated by organic labels or, likewise, by Fair Trade labels. To bridge this research gap, this study will examine the interaction between COO cues and labels in relation to purchase decisions. Evidence on this interaction would help marketeers, as this relationship could implicate that consumers’ likeliness to buy a labelled product depends on the COO cue.
In prior research, personal values have been identified as strong motivators for purchase decisions; however, only a few studies have found interaction effects of personal values and Fair Trade consumption or organic consumption (Ladhari & Tchetgna, 2015). To bridge this gap, this study will evaluate the relative impact of personal value orientations on purchase decisions relating to Fair Trade, organic and COO.
In the marketing literature, the underlying motives that motivate the purchase and consumption
of goods and services can be categorized as either for hedonic gratification or utilitarian reasons
(Howard & Millan, 2007; Teller, Reutterer, & Schnedlitz, 2008). Hedonic shopping explains
the purchase of goods to satisfy a consumer’s sensory needs, enlarged by experiences of
pleasure entertainment, fantasy, and fun (Hirschman & Holbrook, 1982), utilitarian shopping
fulfills functional and non-sensory needs. Considering an increasing concern among consumers
regarding the conditions under which products are being produced and fabricated, it is expected
that consumers engaging in hedonic consumption tend to use ethical and eco-labels more often
8 to avoid feelings of guilt. Henceforth, this study aims to clarify whether the preference for labeled goods differs between hedonic or utilitarian shopping motives.
The remainder of this paper is structured as follows. The next section aims to provide a
theoretical background on the different attributes under study. Hereafter, the research design is
explained in detail. Lastly, the results are presented, followed by a discussion and a conclusion.
9
2. Theoretical background
2.1 Country-of-Origin (COO)
In academic literature, the COO construct is defined on numerous levels: country-of-assembly (COA), country-of-design (COD), country-of-manufacturing (COM) or country of a brand (COB) (Nebenzahl, Jaffe, & Lampert, 1997; Samiee, 1994; Srinivasan, C. Jain, & Sikand, 2004). A more general, commonly used definition is that of Nebenzahl & Lambert (1997), where country-of-origin (COO) is defined as “the country which a consumer associates with a certain product or brand as being its COO, regardless of where the product is produced” (p.30).
Considering earlier statements about globalization, it is important to recognize that a product may not be necessarily manufactured in that country for reasons of multinational sourcing.
Nonetheless, in line with other studies, in this research context, it is assumed that the product or brand is identified with that country (Johansson, Douglas, & Nonaka, 1985).
Although numerous definitions have been developed for the COO construct, these definitions do not cover the description of the COO cue that is used in this study design. Therefore, a slightly adjusted definition of that of Nebenzahl & Lambert (1997) is adopted, where the COO construct is defined as “the country which a consumer associates the origin of the main substance of a certain product or brand as being its COO, regardless of where the product is produced or assembled”. This adjusted definition provides a better fit with the attribute in this study, as the COO cue here is understood as the origin of the primary contents of a product.
Despite prior research, Insch & Cuthbert (2018) conclude that there is still limited knowledge about how the COO construct affects consumer decision-making as a (quality) cue. The primary reason for this lack of knowledge can be related to a difficult generalization of the effect and a lack of theoretical underpinnings (Newman, Turri, Howlett, & Stokes, 2014; Peterson &
Jolibert, 1995; Pharr, 2005).
The COO effect was first found by Dichter (1962), proposing that COO influences consumers’
evaluations of products, followed by empirical evidence, claiming that consumers evaluated
products differently when all attributes of a product were identical, except the COO label
(Schooler, 1965). According to Obermiller and Spangenberg (1998), customers tend to more
favorably evaluate brands that are associated with a positive COO; in other words, countries
10 that pertain a positive image. On the contrary, associations of countries with a negative COO tend to impair consumer evaluations (Obermiller & Spangenberg, 1998).
Some studies approach the mechanism behind the effect of COO as merely being a cognitive cue, meaning COO functions only as an informational stimulus referring to a product or brand used by customers to infer beliefs regarding product attributes, for example, size, taste and quality (Li & Wyer, 1994; Verlegh & Steenkamp, 1999).
Other studies demonstrate that the COO effect is not merely a cognitive cue, but can also have a symbolic or emotional meaning (Obermiller & Spangenberg, 1998; Xie et al., 2016). To demonstrate, Obermiller & Spangenberg (1998) categorize the country of origin effects in three categories: cognitive, affective & normative. Here, the cognitive cue is explained as either the halo construct or the summary construct (Erickson, Johansson, & Chao, 2002; Johansson, 1989;
Johansson et al., 1985; Samiee, Shimp, & Sharma, 2005; Shimp, Samiee, & Madden, 1993).
The halo construct regards that country image is being used as a halo infer the quality of an unknown foreign brand, in other words, that the perceptions of the country image directly influence attitudes towards a product or brand in case consumers know little about a country’s products (Han, 1989; Nebenzahl et al., 1997). In this theory, country image is interpreted as consumers’ general perceptions of quality for products made in a given country (Bilkey & Nes, 1982). The second construct to explain the cognitive cue is the summary construct. The summary construct assumes that the perception of a country, the country image, is influenced by the perceived attributes of products made in a given country, learned by experience.
Consequently, the consumer generalizes these perceptions to the attributes of products made in this country and sold under a particular brand name, which makes the country images affecting consumer attitudes towards the brand or product (Nebenzahl et al., 1997).
The affective mechanism entails an emotional response to country stereotypes that influence attitude directly without intervening belief changes (Obermiller & Spangenberg, 1998; p.2). An example that Obermiller & Spangenberg (1998) use, is an Arab-American that evaluates an Israeli-made product high on quality, nonetheless holds a negative attitude towards the product as a consequence of the negative attitude towards Israel.
The normative aspect of the COO effect pertains the social and personal norms that consumers
hold to specific countries of origin (Verlegh & Steenkamp, 1999). From this perspective,
customers consider a purchase for a particular country as indirect support for the economy of
that country. In this respect, buying products from countries that engage in disputable activities
11 can be regarded as an a-moral purchase and indirect support for the countries policies and practices (Smith, 1990). Another norm that relates to the normative aspect is the preference to buy domestic goods. Shimp & Sharma (1987) defined this as consumer ethnocentrism and found that the norm positively influences consumer preference for domestic products, and negatively influence the preference for foreign products.
Bilkey & Nes (1982) found that products from LDCs are generally evaluated less positively than products from MDCs. Their research proposes that the reason for this phenomenon is that products made in LDCs are associated with inferior quality. In support, Verlegh & Steenkamp (1999) found that the COO effect is significantly stronger in case products are compared characterized by COOs that differ in development level, i.e., more developed and less developed countries, where consumers are more likely to purchase products from MDC.
Based on the arguments mentioned above, it is interesting to reflect the three types of mechanisms, proposed by Obermiller & Spangenberg (1998), on the findings of Bilkey & Nes (1982). , it could be argued that negative cognitive associations of LDCs are transferred to the quality perception of products, that consumers hold negative beliefs to LDCs, or that consumers negatively evaluate LDC based on social and personal norms. Based on these considerations and support in the literature, the following hypothesis is developed:
Hypothesis 1: Consumers are more likely to purchase products from more developed countries than products from less developed countries.
2.2 Labeling
For customers to make a well-considered choice between a variety of products, labels are provided to inform the consumer about the characteristics of a product. According to a qualitative research study by the Food Safety Authority in Ireland, it appears that the two main benefits of reading food labels perceived by consumers are to know the ingredients of a product and where it originates from (Food Safety Authority of Ireland, 2009). Moreover, the study reveals that consumers find it important being able to trace the origin of the product, what the product contains, and under what conditions the product is being fabricated (Food Safety Authority of Ireland, 2009).
The main function of labeling is to contribute to a better decision-making process for customers
by providing customers with relevant information that form input in their decision-making
12 process. Here, labels enable consumers to match product choices with their individual preferences. Although numerous studies have revealed the effective use of labeling in general, the relationship between gathered (product) information and consumer behavior remains to be rather complex; therefore, scholars emphasize the need for further research (Caswell &
Mojduszka, 1996; Deselnicu, Costanigro, Souza-Monteiro, & McFadden, 2013).
In academic literature, the signaling theory explains the rationale of why consumers look for signals to distinguish products and choose between alternatives. The roots of signaling theory originate from the study of information economics, which focus on the asymmetric distribution of information between buyers and sellers in the context of market interactions (Spence, 1974).
Similar to the signaling theory, the economic theory perspective, explains the function of reducing the information asymmetry between sellers and consumer (Akerlof, 1970). In the market of consumer goods, sellers know the quality of their product, as opposed to consumers, which are often not fully informed about the offered products. In the decision-making process, consumers, therefore, search for cues and attributes that distinguish the offered products in a variety of dimensions, like quality or price. From this perspective, labels help consumers to distinguish different products from others and make a choice that is in line with their preferences. As the signalling theory offers interesting perspectives for marketeers, it is examined on a variety of attributes such as price, warranties and advertising (Boulding &
Kirmani, 1993).
2.2.1 Effects of labeling
Prior research on labeling in the consumer goods sector has led to different outcomes. Ippolito
& Mathios (1990) found evidence for the effective use of health labels, in terms of fiber labeling, where consumers changed their behavior once informed by the label. In contrast, research by Moorman (1996) reveals that nutrition labels do not function as intentioned for every consumer, as the study found that labels are not successful under levels of low motivation and nutrition knowledge.
Although studies have found evidence for the influence that labels have on consumer behavior,
other studies reveal limitations of labeling. First, limited space on product packaging and high
demand for labeling requires a deliberate choice of which labels to show. Second, the limited
space allocated for labels limits the length and form of messages. Third, consumers tend to
13 spend less time during shopping, which negatively influences the sufficient use of label information in the decision buying process (Annunziata et al., 2011).
2.2.2 Fair Trade labels
In literature, there is no precise categorization of labels used in the consumer goods sector. De Pelsmacker et al. (2006) argue that Fair Trade consumption is a particular type of consumption, that stands for purchasing goods that support farmers, or people in general, in developing countries. For this reason, Fair Trade consumption may not be interpreted as equal to other ethical or eco-labels, as it involves a specific kind of concern for consumers. For that reason, De Pelsmacker et al. (2006) question to what extent empirical research on attitudes towards ethical consumption applies to Fair Trade. To control for this bias, this study will adopt a clear distinction between the different labels under study, namely Fair Trade labels and organic labels, where organic labels differ from Fair Trade since organic labels are identified as eco- labels (Loureiro & Lotade, 2005).
To understand the construct of Fair Trade, this study adopts the common used definition of FINE (a network of Fair Trade organizations) (Becchetti & Huybrechts, 2008; Moore, 2004;
Morrell & Jayawardhena, 2010): "Fair Trade is a trading partnership, based on dialogue, transparency, and respect, that seeks greater equity in international trade. It contributes to sustainable development by offering better trading conditions to and securing the rights of, marginalized producers and workers – especially in the South” (FINE, 2001). It is a certification program related to a trading partnership that aims for sustainable development for excluded or disadvantaged producers (De Pelsmacker et al., 2006). Recent years, a large body of literature paid attention to ethical consumption and the growing importance of the Fair Trade market.
A qualitative study by Shaw & Clarke (1999) presented that Fair Trade was considered as the
most important ethical concern in purchase behavior, in comparison to other issues, for instance,
environmental issues or vegetarianism. In support, quantitative research by De Pelsmacker,
Driesen & Rayp (2005), using a conjoint choice analysis, found that Belgian consumers value
the ethical part in a product and that consumers were willing to pay a premium for Fair Trade
product. In this research, the brand was perceived as the most important attribute, followed by
flavour, and lastly the Fair Trade label (De Pelsmacker et al., 2005). Other studies often examine
Fair Trade purchasing by using the Theory of Planned Behavior (Ajzen, 1985) and generally
find that positive attitudes towards Fair Trade products are positively related to purchase
14 behavior (Hunt & Vitell, 1986; Ozcaglar-Toulouse, Shiu, & Shaw, 2006; Shaw & Clarke, 1999). In addition, studies on the influence of price and Fair Trade labels are consistent in finding that consumers are willing to pay a premium for Fair Trade products (e.g. Basu & Hicks, 2008; De Pelsmacker et al., 2006; Loureiro & Lotade, 2005; Yang, Hu, Mupandawana, & Liu, 2012), except results on WTP values are relatively heterogeneous. The results vary, mainly due to the geographical context of the study, sampling type, methodology, product type, number of attributes (Rotaris & Danielis, 2011).
Considering both qualitative and quantitative studies mentioned above, it appears that existing literature finds consisting evidence for positive evaluations of Fair Trade products. Therefore, in this study, the following hypothesis is developed:
Hypothesis 2: The presence of a fair trade label is positively associated with purchase intentions, compared to having no label.
2.2.3 Organic labels
Similarly to Fair Trade products, most consumers hold positive attitudes and beliefs towards organic products (Kihlberg & Risvik, 2007; Magnusson, Sjödén, Åberg, Koivisto Hursti, &
Arvola, 2001; Saba & Messina, 2003). Scholars found that organic products are mainly purchased for health reasons, as organic products are not, or to a lesser extent, exposed to pesticides or chemical substances (Davies, Titterington, & Cochrane, 1995; Foster & Latacz‐
Lohmann, 1997; Tregear, Dent, & McGregor, 1994). A few studies highlight that consumers purchase organic foods because of ethical or moral reasons (Morris, 1996; Worcester, 2000).
In line with research on Fair Trade labels, prior studies on organic consumption are generally conducted using the Theory of Planned behavior (Ajzen, 1985), and have found that people that tend to have positive attitudes towards organic products (e.g. Arvola et al., 2008; Chen, 2007;
Thøgersen & Ölander, 2006). An interesting paradox, is that many scholars found that the relatively higher price premium is the main barrier to purchase organic products (Batt & Giblett, 1999; Lynchehaun & Hill, 2002; Magnusson et al., 2001; Tregear et al., 1994), but on the other hand, many scholars agree that consumers are willing to pay a premium for organic products (Batte, H. Hooker, Haab, & Beaverson, 2007; Calverley & Wier, 2002; Gil, Gracia, & Sanchez, 2000; Jones, Shears, Hillier, & Clarke‐Hill, 2001; Magnusson et al., 2001; Thompson &
Kidwell, 1998).
15 Although price remains an important attribute, the existing literature is consistent in finding that consumers generally hold positive attitudes towards organic products, compared to products that are not organic. Based on this reasoning, the following hypothesis is developed:
Hypothesis 3: The presence of an organic label is positively associated with purchase intentions, compared to having no label.
2.3 The interaction effect between COO and labels
Only a few studies have examined the interaction effect between the COO cue and labels.
Dentoni, Tonsor, Calantone, & Peterson (2009) found evidence for positive effects of the
“locally grown” attribute in combination with other credence attributes, where credence attributes represent quality cues that cannot be verified before purchase, such as organic practices or nutrition contents (Darby & Karni, 1973). Similarly, Thilmany, Onozaka, & Nurse (2011) found that the negative valuation of imported food products could be mitigated by USDA organic certification standards. Consumers may consider organic food produced in foreign countries are certified under less strict standards than those produced in the United States (Friedland, 2005). In support of this claim, Xie, Gao & Swisher (2016) found that domestically produced organic food was evaluated significantly higher than that of imported organic food.
Given an increased demand in organic foods and the globalization of the food industry, Xie et al. (2016) stress the importance of further research and the importance of examining the interaction effect between organic labels and the COO cue. This study answers this research call in a wider context, and will examine the interaction between labels and the COO cue, in which both Fair Trade and organic labels will be tested.
Since the proposed interaction effect between the attributes is under-researched, this study aims to take on an exploratory approach, as prior research cannot give a decisive answer on the direction of the interaction effect. Therefore, the following hypotheses are developed:
Hypothesis 4: The development level of country-of-origin influences the relationship between Fair Trade labels and purchase intentions.
Hypothesis 5: The development level of country-of-origin influences the relationship
between organic labels and purchase intentions.
16
2.4 Personal values as motivators
Numerous scholars have found that personal values influence consumer preferences, and therefore purchase behavior (Ah Keng & Yang, 1993; Ladhari & Tchetgna, 2015; Lowe &
Corkindale, 1998; Vinson, Scott, & Lamont, 1977). Personal values are defined as motivational constructs that are related to “desirable trans-situational goals that vary in importance, and serve as guiding principles in the life of a person or other social entity” (p.21) (S. H. Schwartz, 1994).
Prior studies have researched the influence of personal values in the context of ethical behavior (e.g. de Ferran & Grunert, 2007; De Pelsmacker et al., 2005; Doran, 2009; Ma & Lee, 2012).
De Ferran & Grunert (2007) used a laddering methodology and suggested that Fair Trade consumption is either derived from taste (individualistic) or for human well-being (collectivist).
De Pelsmacker et al. (2006) identified four clusters and found that people were more likely to value Fair Trade products when they were relatively more idealistic. Doran (2009) compared the means of Schwartz’ values among three different consumer segments and found that consumers that value universalism, were loyal Fair Trade buyers, as universalism includes values like unity with nature, protect the environment, a world of beauty.
For organic labels, Van Huylenbroeck, Verbeke, Mondelaers, & Aertsens have found that egocentric values are stronger motivators than altruistic values in relation to the purchase of organic food (2009). Furthermore, the meta-analysis revealed that health, related to the value security, is the strongest motivation for the purchase of organic food (Van Huylenbroeck et al., 2009).
In the context of ethical and green consumption, numerous studies suggest that only four values jointly underlie environmental and ethical concern, respectively egoism, altruism, hedonism, and biospherism (Dietz, Stern, & Guagnano, 1998; Stern, 2000; Stern & Dietz, 1994). De Groot
& Steg (2007) validated this assumption and developed a short rating scale which measures these four values orientations and is commonly used in the context of responsible consumption (Gatersleben, Murtagh, & Abrahamse, 2014).
Based on the above-mentioned studies, this research will examine the interaction effects of the
four value orientations, proposed by De Groot & Steg (2007). Prior studies have found that
these studies influence responsible consumption, therefore, the following hypothesis will be
tested:
17 Hypothesis 6: Personal values will influence the relationship between COO, labels,
and purchase intentions.
2.5 Hedonic and utilitarian consumption
In the field of marketing, research on the purchase of goods and service is predominantly explained in two basic reasons: either hedonic or utilitarian consumption. Hedonic shopping motives explain consumer behavior as customers seek value according to pleasure, recreational consumption, and high-arousal stimuli (Babin, Griffin, & Darden, 1994). At the other side of the spectrum, utilitarian shopping motives are characterized as rational, efficient, and task- related with a focus on completion of the task (Adomaviciute, 2013; Batra & Ahtola, 1991).
Numerous scholars argue that hedonic consumption motives can result in a sense of guilt for consumers (Belk, Ger, & Askegaard, 2003; Kivetz & Simonson, 2002; Strahilevitz & Myers, 1998). The feeling of guilt can be explained as an individual’s concern for moral stands or harm done to other people or society (Eisenberg, 2000; Tangney & Dearing, 2002). Moreover, research reveals that consumers experience a sense of guilt when purchasing for an unethical alternative (Adomaviciute, 2013; Marks & Mayo, 1991).
Based on the arguments above, the question arises whether shoppers that engage in hedonic shopping prefer labeled goods over non-labeled goods to avoid the prior mentioned feeling of guilt. In order to evaluate whether the consumption motive influences the purchase decisions, the following hypothesis is developed:
Hypothesis 7: The influence of labels on purchase intentions is stronger in the case of hedonic consumption motivations rather than utilitarian consumption motivations.
2.6 Price & willingness-to-pay
In order to gain a comprehensive understanding of the attributes under study, the price attribute
is included in the research design. The academic literature provides convincing evidence on the
willingness-to-pay a premium for Fair Trade products (De Pelsmacker et al., 2005), organic
products (Gil et al., 2000), different COO labels (Ojea & Loureiro, 2007) and consumption
motives (Dhar & Wertenbroch, 2000). However, a common critic on other Fair Trade studies
is that studies focus on beliefs and attitudes, and overlook the influence of higher prices
(Browne, Harris, Hofny-Collins, Pasiecznik, & Wallace, 2000). Prices, quality, convenience,
18 and brand familiarity often remain the most important attributes in purchase decisions (Attalla
& Carrigan, 2001; Boulstridge & Carrigan, 2000). For this reason, to be able to determine the relative importance of the attributes and to calculate willingness-to-pay, it is required to include the price attribute. However, due to convincing evidence in the literature, no hypotheses were developed.
2.7 Conceptual framework
To visualize the relationships between the presented variables, hereby the conceptual framework is presented (figure 1). This model represents all the variables under study, as well as the hypothesized effects between the concepts.
Figure 1 Conceptual framework
19
3. Methodology
3.1 Research design
In order to answer the proposed hypotheses, a quantitative study is conducted to gain insights on the effects of COO and labels consumers’ purchase behavior. The research method will consist of a choice-based conjoint analysis, as this type analysis consists of simulated purchase decisions, where respondents repeatedly have to choose between products. This methodology tends to be very realistic, as choices are essential in everyday life. Furthermore, these choices are perceived as relatively simple tasks for respondents compared to ratings or rankings (Desarbo, Ramaswamy, & Cohen, 1995). Moreover, ranking or rating procedures can be problematic, as different cultures tend to respond differently to rating scales and differ in the meaning and interpretation of the values (Eggers & Sattler, 2011). The surveys that have been designed can be found in appendix 2 and appendix 3.
3.2 Research context
The study is conducted in the context of coffee products since this product category meets several criteria that are important to ensure validity. First, it must be a product where COO is visualized on the packaging. Second, the product under study must be eligible for a Fair Trade label. Third, it must be possible to produce or grow the product organically. Fourth, it should be possible to purchase the product for both hedonic and utilitarian reasons. As coffee products meet all the criteria mentioned above, it can be concluded that the product category of coffee appears to be a valid product for this research.
3.3 Data collection
This study makes use of primary data, as a survey is used to transform data in quantitative records. The software used for the survey is PreferenceLab, an online tool that allows researchers to gain insights into consumer behavior by designing choice-based surveys.
The survey was pretested before it was send-out to a larger audience, to identify potential
problems regarding the intent, clarity of the survey and navigation and so increase clean
responses and low error rates (Fanning, 2005).
20
3.4 Survey development
It is essential to decompose the presented products in attributes and levels to design the choice sets used in the surveys. It is advised to use only a few attributes, at maximum six, in order to make the provided information processable for respondents and reduce complexity (Eggers &
Sattler, 2011). In this study, four attributes were identified and assigned to different coffee products.
The first attribute is based on the COO construct, which consists of two levels, respectively, lower developed countries (LDC) and more developed countries (MDC). The second attribute is the presence of an organic label; the third attribute is the presence of a Fair Trade label.
Lastly, the fourth attribute is the price of the coffee product, which is operationalized into four levels, as this allows for tests on willingness-to-pay for the different presented products. In table 1, the attributes and levels used in this study are summarized.
Independent Variable
Attributes Level 1 Level 2 Level 3 Level 4
Country-of- origin
COO cue Less Developed
Country
More Developed Country
Fair Trade label Fair Trade label cue No Fair Trade label Fair Trade label
Organic label Organic label cue No organic label Organic Label
Price Price of the product Price level 1 Price level 2 Price level 3 Price level 4
Table 1 Attributes and Levels used in research design
3.4.1 Choice sets
In designing the various choice sets that are used in the survey, several criteria are taken into consideration. First, all attributes and attribute level pairings should appear an equal number of times in the survey. Second, the choice must not consist of equal levels within one attribute.
Lastly, the choices or alternatives must be equally attractive to the respondent in order to control
for dominating alternatives (Eggers & Sattler, 2011).
21 To determine the number of choice sets used in this study, this study adopts the methodology presented by Hair, Black, Babin, Anderson, & Tatham (2006) which holds that the following formula can calculate the minimum number of choice sets:
Minimum number of choice sets = total number of levels across all attributes – number of attributes + 1.
Using the formula, the minimal number of choice sets for this study would be 8. This is an acceptable number according to observations by Eggers & Sattle (2011), who recommend using 12 to 15 choice sets at maximum.
In the survey, three alternatives per set are presented in each choice set (see figure 2). A no- choice option is not included in the alternatives, as the situation in the survey is illustrated in a way that a “no-choice” option is not applicable. The shopping situation in the survey is primed in a manner that a respondent is urged to buy coffee, which excludes the possibility of a no- choice action.
Figure 2 Example of choice set in the survey
22
3.5 Measurements
To measure the constructs under study, the variables must be operationalized in order to measure the relationships. This paragraph elaborates on how the constructs are being transformed into measurable variables.
3.5.1 COO construct
The COO constructs are based on two categories: less developed countries (LDC) and more developed countries (MDC). Here, the distinction is adopted from Verlegh & Steenkamp (1999), in which the authors classify MDCs as all OECD-members countries, excluding Greece, Mexico, and Turkey. Next, the countries must be considered as a coffee-producing country, as this is the product under study. Based on these two criteria, Ethiopia is assigned as LDC and Italy as MDC.
3.5.2 Labels
This study will adopt a clear distinction between different labels under study, namely Fair Trade labels and eco-labels, in line with categorizations in similar studies (Loureiro & Lotade, 2005).
For eco-labels, the label will consist of an organic label, as an organic label is strongly associated with eco-labels (Loureiro & Lotade, 2005). For Fair Trade labels, the Max Havelaar certificate will be used.
3.5.3 Personal values
Personal values are measured using the adapted version of Schwartz’s value scale (S. Schwartz, 1992; S. H. Schwartz, 1994) developed by De Groot & Steg (J. I. M. de Groot & Steg, 2007; J.
de Groot & Steg, 2008). This scale measures four values orientations, namely egoistic, altruistic, hedonic, and biospheric value orientations.
For egoistic value orientations, social power, wealth, authority, being influential, and ambition
were measured. For altruistic value orientations, equality, a world of peace, social justice, and
being helpful were measured. For hedonic value orientations, pleasure, enjoying life, and self-
indulgence were measured. For biospheric value orientations, preventing pollution, respecting
the earth, unity with nature, and protecting the environment were measured.
23 All of these values were rated based on perceived importance for the respondents, defined as
“being guiding principles in their lives, on a nine-point Likert scale, ranging from -1 “opposed to the principles that guide me”, 0 “not at all important” to 7 “of supreme importance as a guiding principle in my life”.
3.5.4 Hedonic and utilitarian shopping motives
To evaluate whether the effects under study differ between shopping motives, two surveys were designed. The shopping situation in the survey was primed as either a utilitarian shopping motive (appendix 2) or a hedonic shopping motive (appendix 3). To control whether the primed situations were perceived as either hedonic or utilitarian, three control statements are included, which measure to what extent the situation is perceived as either utilitarian or hedonic, using a five-point Likert scale. Correlation analysis confirmed that both situations were perceived as intended, with a significant difference between both situations (see appendix 1).
3.5.5 Control variables
A few control variables were included as these variables might influence the constructs under study. In this study, the control variables include age, gender, education level, and income.
Studies have found that individuals engaging in socially responsible consumption are predominantly female (Eagly, 2013; J. A. Roberts, 1993), older (Hallin, 1995) and higher educated. Although studies show correlations between socio-demographic variables and pro- environmental behavior, or socially responsible behavior, the correlation, in general, remains to be weak or questionable (Hines, Hungerford, & Tomera, 1987) or contradicting (Olli, Grendstad, & Wollebaek, 2001). Thus, there is a possibility that age, gender, education, and income level influence the outcomes of this research, therefore, these variables are chosen as control variables.
3.6 Sampling
The sample of this study will be determined by probability sampling, defined as samples that
are selected in a way that every person of the population has a nonzero possibility of being
included in the sample (Henry, 1990). The sampling type will be convenience sampling, which
is a type of sampling that selects a population-based on practical criteria, in this case,
accessibility, availability and the willingness to participate for this study (Dornyei, 2007). A
24 common critic on convenience sampling is that it is a biased manner of sampling, as it would not represent the population and it would be very sensitive to outliers, due to high levels of self- selection by the researcher (Mackey & M. Gass, 2005). For this reason, a critical analysis is conducted on the sample characteristics in order to highlight eventual outliers.
3.7 Data analysis
For the data analysis, the data derived from the online questionnaire will be examined in order to answer the proposed hypotheses. The data will be analyzed by using the statistical application LatentGOLD. The repeated choices of the respondents are estimated for significant effects of the attributes on the purchase behavior.
For the data analysis, several models will be tested. First, the main effects of the given attributes
are estimated. Second, the interaction between the attributes is being tested. Third, the influence
of personal values on the main variables is estimated. Lastly, two models will estimate the
influence of hedonic and utilitarian shopping motives on the interaction between the attributes,
and the personal values.
25
4. Results
In total, 190 respondents completed the survey. No responses were detected that completed the survey in less than 2,5 minutes. To ensure reliability, only completed surveys were used for the data analysis.
4.1 Pre-test
Prior to the data collection, five pre-sets were conducted to ensure the clarity of the questions and to verify whether respondents perceived the two surveys as either hedonic or utilitarian conditions. The questions and choice sets appeared to be clear, and the difference in consumption motives seemed to be perceived as intended. Also, pre-test respondents indicated that only after a few choice sets, they were fully aware of the different attributes that were displayed. No action was taken in response to this feedback, as this process is part of the consumer decision-making process that is under study.
4.2 Sample characteristics
See table 2 for the characteristics of the sample for this study. Out of the respondents, 53,2% is male, and 46,8% is female, and the vast majority of the respondents is aged under 25.
Furthermore, the respondents were predominantly higher educated, as 66,9% had obtained a degree on a University level, and the median for income was €1 - €9,999 per year.
Variable Frequency Percentage
Gender - Male - Female
101 89
53,2%
46,8%
Age
- <25 - 25-34 - 35-44 - 45-54 - 55-64 - 65+
114 26 12 24 11 3
60%
13,7%
6,3%
12,6%
5,8%
1,6%
Education Level - No education - Elementary school - High School
1 4 35
0,5%
2,1%
18,4%
26
- Intermediate Vocational Education - University of Applied Sciences
- University (Bachelor, Masters or Doctoral degree)
23 49 78
12,1%
25,8%
41,1%
Income Level - €0 - €1 - €9,999 - €10,000 - €24,999 - €25,000 - €49,999 - €50,000 - €74,999 - €75,000 - €99,999 - €100,000 - €149,999 - €150,000+
53 66 37 11 5 5 3 10
27,9%
34,7%
19,5%
5,8%
2,6%
2,6%
1,6%
5,3%
Table 2 Sample Characteristics
4.3 Data analysis
In order to test the developed hypotheses, a number of choice-based conjoint analyses were performed to determine how the different variables influence the dependent variable. In this section, the outcomes will be presented and interpreted.
4.3.1 Model 1: attributes without interaction effects
Model 1 tests the main attributes without interaction effects. Model 1 indicates an R
2above 0.2, which is considered to be acceptable and represents a good fit with the data.. The predictive value of this model can be analyzed by the HIT-rate, which represent how well the model predicts the observed values. For this model, the HIT-rate is 68%.
First, a choice-based conjoint analysis is performed using the independent variables in order to
determine the effects on the dependent variable. Model 1 shows a significant effect of all
attributes on the purchase intention of the respondents (p = < 0.000). Also, the results on COO
show that consumers are more likely to purchase a product from an MDC than an LDC (β =
0,1139, p = 0.000), which supports hypothesis 1. Finally, the results indicate that the presence
of both Fair Trade labels (β = 0,6673, p = 0,000) and organic labels (β = 0,3860, p = 0,000)
have a significant positive influence on purchase intentions, which support hypothesis 2 and 3.
27
Model 1: Attributes without interaction effects
Attribute Level Parameter Wald-statistic p-value Relative
importance
Organic No organic -0,3860 225,5998 0.00 *** 0,1586
Organic 0,3860
Fair Trade No Fair Trade -0,6673 593,2756 0.000 *** 0,2742
Fair Trade 0,6673
Country-Of-Origin Lower Developed Country -0,1139 20,8871 0.000 *** 0,0468
More Developed Country 0,1139
Price 2,5 1,2294 1098,5688 0.000 *** 0,5204
3 0,3813
3,5 -0,3068
4 -1,3039
Model 1. R2 (Overall) = 0,3433. L2 = 4844,7786. X2= 1.22e+13, p=0.000***
4.3.2 Willingness-to-pay for attributes
Based on the parameters of model 1, the willingness-to-pay can be calculated for the attributes.
Table 3 presents the willingness-to-pay for the different attributes, based on which can be concluded that consumers are willing to pay an additional 46 cents for organic products, 79 cents for Fair Trade products and 13 cents for a product from an MDC.
Attribute Willingness-to-Pay
Organic 0,4571
Fair Trade 0,7902
More Developed Country 0,1348
Table 3 Willingness-to-pay per attribute, derived from model 1.
4.3.3 Model 2: attributes with interaction effects
Model 2 tests the influence of the independent variables on purchase intentions, including the
interaction effects. Model 2 shows an R
2above 0.2, which is considered to be acceptable and
28 represents a good fit with the data. For this model, the HIT-rate is identical to model 1, respectively 68%.
The results indicate that the level of development of the COO cue has no significant interaction with the relationship between organic labels (β = -0,0574, p = 0,56), Fair Trade labels (β = 0,0698, p = 0,50) or price (β = -0,0242, p = 0,66) on purchase intentions, therefore hypotheses 4 and 5 are rejected.
Model 2: Attributes with interaction effects
Attribute Level Parameter Wald-statistic p-value Relative importance
Organic No organic -0,400 124,3940 0,000*** 0,1573
Organic 0,400
Fair Trade No Fair Trade -0,6490 282,9110 0,000*** 0,2552
Fair Trade 0,6490
Country-Of-Origin Lower Developed Country -0,1362 4,0409 0,044* 0,0535
More Developed Country 0,1362
Price 2,5 1,2094 458,7050 0,000*** 0,4899
3 0,3739
3,5 -0,3008
4 -1,2825
Interaction COOxFairTrade 0,0698 0,4551 0,50 0,0137
COOxOrganic -0,0574 0,3312 0,56 0,0113
COOxPrice -0,0242 0,1942 0,66 0,0190
Model 2. R2 (Overall) = 0,3431. L2 = 4843,7970. X2= 9.1131E+12. p=0.000***
4.3.4 Model 3: interaction effects between attributes and personal values
Model 3 estimated the interaction effects of the different personal value orientations with the
attributes under study. Model 3 shows an R
2above 0.2, which is considered to be acceptable
and represents a good fit with the data. For this model, the HIT-rate is 70.2%.
29 First, the results indicate a significant influence of egoistic value orientations on the choice of Fair Trade products (β = -0,0297, p = 0,002), organic products (β = -0,0521, p = 0,000) and price (β = 0,0025, p = 0,033). Second, altruistic value orientations have a significant influence on the choice of Fair Trade (β = 0,0465, p = 0,003) and price (β = 0,0330, p = 0,000). Third, hedonic value orientations appear to have a significant influence on COO (β = 0,0794, p = 0,000) and on price (β = -0,0293, p = 0,001). Lastly, results indicate that biospheric value orientations have a significant influence on the choice of Fair Trade products (β = 0,0460, p = 0,000), organic products (β = 0,0534, p = 0,000), and on price (β = 0,0140, p = 0,006). In conclusion, the results present support for hypothesis 6.
Model 3: interaction effects between attributes and personal values
Attribute Level Parameter Wald-statistic p-value Relative importance
Organic No organic -0,8204 25,3905 0.000*** 0,0588
Organic 0,8204
Fair Trade No Fair Trade -0,2263
Fair Trade 0,2263 1,9081 0,17 0,0162
Country-Of-Origin Lower Developed Country
0,6165 16,0866 0,000*** 0,0441
More Developed Country
-0,6165
Price 2,5 2,0173 79,8481 0,000*** 0,1480
3 0,6569
3,5 -0,5585
4 -2,1156
Interactions HedonicxFairTrade -0,0232 1,3529 0,24 0,0175
EgoisticxFairTrade -0,0297 9,4321 0,002** 0,0414
AltruisticxFairTrade 0,0465 8,9115 0,003** 0,0467
BiosphericxFairTrade 0,0460 17,2864 0,000*** 0,0510
HedonicxOrganic -0,0275 2,0707 0,150 0,0207
EgoisticxOrganic -0,0521 31,1806 0,000*** 0,0727
30
AltruisticxOrganic -0,0279 3,3558 0,067 0,0280
BiosphericxOrganic 0,0534 24,7991 0,000*** 0,0593
HedonicxCOO 0,0794 19,0615 0,000*** 0,0597
EgoisticxCOO 0,0084 0,9441 0,33 0,0118
AltruisticxCOO 0,0142 0,9971 0,32 0,0142
BiosphericxCOO -0,0150 2,1457 0,14 0,0167
HedonicxPrice -0,0293 10,3683 0,001*** 0,0880
EgoisticxPrice 0,0025 0,3591 0,033* 0,0141
AltruisticxPrice 0,0330 21,3835 0,000*** 0,1290
BiosphericxPrice 0,0140 7,6386 0,006** 0,0622
Model 3. R2 (Overall) = 0,3876. L2 = 4542,2668. X2= 4,3780E+22. p=0.000***
4.3.5 Model 4: attributes without interactions (2-classes)
In order to determine whether the effect of the attributes differ between hedonic and utilitarian conditions, a 2-class conjoint choice analysis is performed, see model 4. In this model, class 1 represents the utilitarian condition, and class 2 represents the hedonic condition. Again, this model shows an R
2above 0.2, which is considered to be acceptable and represents a good fit with the data. For this model, the HIT-rate is 69,1%.
First, the results indicate a significant difference between the effect of a Fair Trade label
between the two conditions (p = 0,000). The effect is more positive for the utilitarian condition
(β = 0,8790, p = 0,000), than for the hedonic condition (β = 0,5831, p = 0,000) and the relative
importance of the Fair Trade attribute is more important for the utilitarian condition (0,3488)
than for the hedonic condition (0,2410). In addition, there is a significant difference for the
effect of price between the utilitarian condition (β = -1,1710, p = 0,048) and the hedonic
condition (β = -1,3776, p = 0,048), where price was perceived as more important for the hedonic
condition (0,554) than for the utilitarian condition (0,4432). No significant difference between
the two classes was found for the effect of the organic attribute. Based on the results, hypothesis
7 is rejected.
31
Model 4: attributes without interactions (2-classes)
Attribute Level Parameter Wald-
statistic (=)
p-value Relative importance
Class 1 Class 2 Class 1 Class 2
Organic No organic -0,4163 -0,3770 0,4875 0,49 0,1652 0,1558
Organic 0,4163 0,3770
Fair Trade No Fair Trade -0,8790 -0,5831 21,9424 0,000*** 0,3488 0,2410
Fair Trade 0,8790 0,5831
Country-Of- Origin
Lower Developed Country -0,1078 -0,1156 0,0200 0,89 0,0428 0,0478
More Developed Country 0,1078 0,1156
Price 2,5 1,0629 1,3099 7,8979 0,048* 0,4432 0,5554
3 0,3889 0,3849
3,5 -0,2808 -0,3171
4 -1,1710 -1,3776
Model 4. Conjoint Choice Analysis. R2 (Overall) = 0,3511. L2 = 4795,0950. X2= 1,539E+13, p = 0.000.
4.3.6 Model 5: attributes with personal value interactions (2 classes)
In order to determine whether the interaction effects of personal value orientations and the choice for attributes differ between hedonic and utilitarian conditions, a 2-class choice-based conjoint analysis is performed, see model 5. In this model, class 1 represents the utilitarian condition, and class 2 represents the hedonic condition. As for previous models, this R
2value shows a good data fit with the data, as this value as is above 0.2. The HIT-rate of this model 71,7%.
This model is beyond the scope of this research but is included for exploratory purposes, as it
presents valuable insights that were not anticipated for. In addition to the results of model 4,
model 5 finds significant differences for interactions between personal value orientations and
attributes between the hedonic and utilitarian consumption motives.
32 For altruistic value orientations, significant differences between the conditions were found for the interaction with Fair Trade labels (β = -0,0379; β = 0,0903, p = 0.000), organic labels (β = 0,0497; β = -0,064, p = 0.000), the COO attribute (β = 0,0477; β = 0,0455, p = 0.003) and price (β = 0,0046; β = 0,00510, p = 0,003). For biospheric value orientations, significant differences were found for the interaction with organic labels (β = 0,0102; β = 0,0614, p = 0.034) and price (β = 0,0399; β = 0,0059, p = 0,0051). For egoistic value orientations, significant differences were found for the interaction with organic labels (β = -0,0185; -0,0641, p = 0,016) and price (β = 0,0185; β = -0,0079, p= 0.004). Lastly, for biospheric personal value orientations, significant differences were found for the choice of organic labels (β = 0,0102; β = 0,0614, p = 0.034) and price (β = 0,0399; β = 0,0059, p = 0.0051).
Model 5: attributes with personal value interactions (2 classes)
Attribute Level Parameter Wald-
statistic (=)
p-value Relative importance
Class 1 Class 2 Class 1 Class 2
Organic No organic -0,1783 -1,3159 12,3183 0,000*** 0,0118 0,0720
Organic 0,1783 1,3159 0,0659 0,0040
Fair Trade No Fair Trade -0,9993 0,0739 9,2272 0,002** 0,0106 0,0612
Fair Trade 0,9993 -0,0739 0,1083 0,1392
Country-Of- Origin
Lower Developed Country -0,1613 1,1190 15,7028 0,000*** 0,0290 0,0292
More Developed Country 0,1613 -1,1190 0,0798 0,0243
Price 2,5 1,5566 2,5024 4,7337 0,19 0,1083 0,1392
3 0,5674 0,7882
3,5 -0,3954 -0,6999
4 -1,7285 -2,5906
Interactions HedonicxFairTrade 0,0418 -0,0509 4,4135 0,0290 0,0292
EgoisticxFairTrade -0,0621 -0,0228 3,3527 0,067 0,0798 0,0243
AltruisticxFairTrade -0,0379 0,0903 12,8882 0,000*** 0,0350 0,0692
BiosphericxFairTrade 0,0442 0,0417 0,0084 0,93 0,0452 0,0353
33
HedonicxOrganic -0,0273 -0,0412 0,1200 0,73 0,0189 0,0236
EgoisticxOrganic -0,0185 -0,0641 5,8610 0,016** 0,0237 0,0684
AltruisticxOrganic 0,0497 -0,0626 12,2251 0,000*** 0,0459 0,0480
BiosphericxOrganic 0,0102 0,0614 4,4765 0,034** 0,0452 0,0353
HedonicxCOO 0,0676 0,0965 0,5576 0,46 0,0468 0,0554
EgoisticxCOO -0,0137 0,0149 2,2459 0,13 0,0176 0,0159
AltruisticxCOO -0,0477 0,0455 9,0250 0,003** 0,0441 0,0348
BiosphericxCOO -0,0019 -0,0169 0,3742 0,54 0,00019 0,0143
HedonicxPrice -0,0470 -0,0206 1,8573 0,17 0,1302 0,0473
EgoisticxPrice 0,0185 -0,0079 8,3400 0,004** 0,0951 0,0338
AltruisticxPrice 0,0046 0,0510 9,0200 0,003** 0,0164 0,1519
BiosphericxPrice 0,0399 0,0059 7,8373 0,0051** 0,1633 0,0201
Model 5. Conjoint Choice Analysis. R2 (Overall) = 0,4036. L² = 3498,3216. X² = 1,277E+24. P = 0,000***.
4.4 Hypotheses overview
Based on the results of the choice-based conjoint analyses, the proposed hypotheses can be examined. In table 3, an overview of the hypotheses is presented, whereby is stated whether the hypotheses are supported or rejected.
Hypotheses Supported/Rejected
Hypothesis 1: Consumers are more likely to purchase products from more developed countries than products from less developed countries.
Supported
Hypothesis 2: The presence of a fair trade label is positively associated with purchase intentions, compared to having no label.
Supported
Hypothesis 3: The presence of an organic label is positively associated with purchase intentions, compared to having no label.
Supported
Hypothesis 4: The development level of country-of-origin influences the relationship between Fair Trade labels and purchase intentions.
Rejected
34
Hypothesis 5: The development level of country-of-origin influences the relationship between organic labels and purchase intentions.
Rejected
Hypothesis 6: Personal values will influence the relationship between COO, labels, and purchase intentions.
Supported
Hypothesis 7: The influence of labels on purchase intentions is stronger in the case of hedonic consumption motivations rather than utilitarian consumption motivations.
Rejected
