EFFECTS OF PRODUCT QUALITY ON
BROWNWASHING BEHAVIOUR
An analysis of the wine industry
MAARTEN A. J. M. VAN DER MULLEN
Strategic Innovation Management
S3245454
ABSTRACT
Brownwashing is a strategy where environmental achievements are deliberately understated. This thesis aims to provide a better understanding of this phenomenon by studying it on a product level. More specifically, the effect of product quality on the amount of brownwashing is analysed. The wine industry is used as an empirical context to study the absence of environmental achievements (vegan labels) on products (wine bottles). By using a dataset of 156 vegan wines, the likelihood of brownwashing in three different quality segments is analysed. The outcome of a probit regression show that brownwashing occurs more often in the low- and high-quality segments than in the medium-quality segment. The main explanation for these findings is that customers have stronger negative associations towards vegan labels in these particular quality segments. In order to avoid these, companies purposely use fewer labels on low- and high-quality products. This thesis contributes to the research on impression management and brownwashing by providing support of this phenomenon on product level, and by identifying product quality as a factor of brownwashing.
Keywords: Brownwashing, Impression Management, Product, Quality, Green Label, Wine
Supervisor: Prof. Dr. Jordi Surroca Co-assessor: Prof. Dr. Jana Oehmichen
TABLE OF CONTENT
INTRODUCTION ... 3
Empirical setting ... 5
THEORY AND HYPOTHESIS... 8
METHODOLOGY ... 14 Measurements ... 15 Vegan label ... 15 Quality ... 16 Firm characteristics ... 17 Country ... 17 Region ... 18 Store ... 18 Analytical method... 19 RESULTS ... 20 Descriptive statistics ... 20 Probit regression ... 20 Additional analysis ... 23 DISCUSSION ... 24 Theoretical implications ... 24 Practical implications ... 25
Limitations and future research ... 27
CONCLUSION ... 28
LITERATURE ... 29
INTRODUCTION
In the last few years, the pressure on companies to behave responsibly towards the environment has vastly increased (Berrone, Fosfuri, Gelabert, & Gomez-Mejia, 2013; Kim & Lyon, 2015; Kittilaksanawong, 2019). Simultaneously, consumers have become more aware of their personal environmental impact and have bought more environmentally friendly products (Lii & Lee, 2011; Martínez & Rodríguez del Bosque, 2013). Companies anticipated to this changed situation by being more attentive of their environmental impression. By doing so, companies aim to alter the perception of the company regarding environmental friendliness, and with that increase the value of the company (Leire & Thidell, 2005; Loureiro & Lotade 2005; Eichholtz, Kok & Quigley, 2010). In order to attain this objective, some companies increasingly promote their environmental friendliness (Terrachoice, 2010), while others choose an opposite strategy.
The latter chooses to appear less environmentally friendly to their customers and other stakeholders (Rauber, 2006). When companies are deliberately understating their environmental achievements, they are brownwashing (Kim & Lyon, 2015). There are several reasons why companies choose to brownwash and prefer to avert the effects of appearing environmentally friendly. For example, when companies invest in becoming carbon neutral, and articulate this to their customers and stakeholders, the latter expects to incur the investment costs, which result in higher prices or lower profits. Consequently, they are less attracted to the company, which has a negative effect on the market value (Jacobs, Singhal & Subramanian, 2010; Fisher-Vanden & Thorburn, 2011).
Scholars found that, on a product level, customers tend to have negative perceptions towards environmentally friendly products and that they estimate these to have higher prices (Bonini & Oppenheim, 2008; Delmas & Lessem, 2017). Additionally, green claims on products (e.g. labels) might decrease the perceived quality of the product (Galarraga Gallastegui, 2002; Peattie & Crane, 2005). This effect repels customers from buying the product (Delmas & Grant, 2014; Bonini & Oppenheim, 2008). Some producers prefer to avoid these negative consequences, and decide to brownwash their products. Kim & Lyon (2015, p.705) even stated: ‘‘it may pay to be brown”.
attention to brownwashing is surprising since stakeholders (e.g. customers) are provided with incomplete and modified information (Delmas & Burbano, 2011). For example, by a selective disclosing of knowledge, customers will develop a different perception of the product and, most likely, make a different choice.
The existing literature on brownwashing has mainly focussed on a firm level (Kim & Lyon, 2015; Testa, Miroshnychenko, Barontini & Frey, 2018). No prior research has been found which focussed products, despite several suggestions of brownwashing on product level (Newman et al., 2014; Testa et al., 2018, p.3). Consequently, there is much uncertainty about this phenomenon. It is unclear in what type of situation products are being brownwashed. In order to fill this research gap, this paper will analyse brownwashing behaviour on products.
There is a division made between products based on product quality segments. Research has indicated that, throughout these quality segments, customers respond differently when faced with environmentally friendly products (Chernev & Carpenter 2001; Raghunathan, Naylor & Hoyer, 2006; Chernev 2007; Torelli, Monga & Kaikati 2012; Luchs et al., 2014; Newman et al., 2014; Delmas & Lessem, 2017). The products are divided into three different quality segments, viz. low-, medium- and high-quality. By examining products corresponding to those segments, this study will shed light on the behaviour of brownwashing. The following question guides the paper:
In which product quality segments is brownwashing more likely to occur?
The answer to the research question will be of added value to the topic of brownwashing in multiple ways. First, it would be the first article focusing on brownwashing on a product level, making it capable of delivering rich information. Moreover, it can increase the understanding of what type of products are relatively more often subject to brownwashing. Since these insights contribute to the understanding of the concept of brownwashing, it fills a part of the gap in the literature. Second, it could help target the quality segments where brownwashing is more likely to occur. These insights could guide future researchers in performing a more specific study on the concept of brownwashing.
labels might increase. As a result, the label organisations gain more clients and the overall amount of brownwashing decreases.
This paper is structured as follows: an empirical setting is identified in order to formulate an answer to the research question. Then, an elaboration on the current relevant theoretical
knowledge is provided. On the bases of existing research, a hypothesis is constructed regarding brownwashing in different product quality segments. The most essential arguments for this prediction are elaborated. Furthermore, the data on products are manually collected, and the level of brownwashing in different quality segments is analysed. The results support a particular non-linear relationship between brownwash-behaviour and the different product quality segments. Based on this, a better understanding of situations where brownwashing is more likely to occur is contributed, and a new factor regarding brownwashing is identified. Consequently, five
implications and several limitations are given. At the end of the paper, an answer to the research question is provided
Empirical setting
The wine industry is selected as the empirical setting of this study. It is possible to expect brownwashing in this industry since wines mostly rely on their visual appearance when being examined by potential customers. This aspect makes it crucial for the wineries to give an appealing impression. Additionally, previous research found evidence for brownwashing activities occurring in the wine industry (Delmas & Grant, 2014). However, there are three more reasons why this industry is suited. First, it allows the study of different quality segments due to the width of wine quality. Second, environmental achievements are relevant in the wine industry because of the enormous environmental impact of agriculture (Audsley, Brander, Chatterton, Murphy-Bokern, Webster & Williams, 2009; Garnett, 2011). Third, this product allows the generalisation of the findings to a large population, due to its wide availability and large consumption per capita (WHO, 2018).
and practicable, all forms of exploitation of, and cruelty to, animals for food, clothing or any other purpose’’. This definition also excludes any form of animal products in both the final product as well as the production process. The popularity of vegan products has vastly increased over the past few years. Between 2014 and 2017, the number of people that identified themselves as vegan increased by 500% (Report Buyer, 2017). Consequently, veganism has become a priority for supermarkets. Tesco has even started an own vegan product line (Quinn, 2017). Also, (Michelin starred) restaurants have shown an increase in vegan products and offer more vegan courses (Conti et al., 2010; Natexpo, 2018)
When producers choose to switch towards a vegan production process, their impact on the environment changes. Agriculture creates an enormous amount of greenhouse gasses that spur climate change (Audsley et al., 2009; Garnett, 2011), and the vast majority comes from animal products (Carlsson-Kanyama & Gonzalez; 2009; Committee on Climate Change 2010; Gonzalez Frostell & Carlsson-Kanyama. 2011; Scarborough et al. 2014). Producing animal products requires a large amount of water and crops, where the latter often causes deforestation (Vegan society, n.d.). The environmental impact of animal products has been one of the reasons why producers turn towards plant-based solutions (Vegan society, n.d.).
In conclusion, changing from animal products to vegan-friendly alternatives eventually decreases the amount of greenhouse gasses as well as the amount of water and soil usage. Since the vegan option is an environmentally friendlier alternative, companies that produce wine in a vegan-friendly way, become more environmentally vegan-friendly.
Producers can promote this attainment to customers by presenting it on the wine bottle. The Vegan Society is an organisation that hands out a trademark vegan label. In contrast to organic certificates where it is mandatory to present the label on products (European Union, 2012), vegan labels are not regulated. Consequently, winemakers can still choose how and if they will present it. Some wineries choose not to provide the information on their label (Delmas & Grant, 2014). When a wine is indeed made according to vegan guidelines, but the wine bottle does not show a vegan label, the product is being brownwashed. In this study, the absence of vegan labels is used as an indicator of brownwashing.
clarification process. This procedure removes small particles, which changes the wine from cloudy to transparent. By adding certain substances, these residues are clustered and sink to the bottom of the barrel. By doing so, the wine becomes visually attractive and increases its physiochemical stability, i.e. it becomes possible to store and transport the wine (Boulton et al., 1999). This procedure is especially necessary for wines that are being exported.
The two most common clarification methods are animal-based and earth-based. Traditionally, the first type of clarifiers is most often used in the wine industry (Boulton et al., 1999; Zoecklein et al. 1999). Animal-based clarifiers can be gelatine, isinglass (fish batter), casein (derived from milk) or egg white (Robinson, 2007). Earth-based clarifiers are substances of natural clays (bentonite or kaolinite), silica or microfilters (Zoecklein et al., 1999). Importantly: none of these substances end up in the final product but will be removed along with the undesired particles. However, since vegans do not use nor consume products where animals have played a part during the production process (Vegan society, n.d.), wines which are clarified with animal-based products are not suited for vegans. When there are no animal products involved in the clarification process, the wine is considered vegan-friendly and is allowed to express this through a vegan label (Hamersma, 2015, Barnivore, 2018, The Vegan Society, n.d.).
THEORY AND HYPOTHESIS
This paper builds on the theory of impression management. Impression management was first conceptualised by Erving Goffman (1959). This theory describes the behaviour where companies are either providing or withholding certain information, in order to (sub-) consciously influence the perception of an individual towards an object, event or person (Goffman, 1959). For example, during a job interview, a candidate might overemphasise or not share specific information about himself, in order to appear more favorable to the adjudicators.
According to research, there are several strategies to alter an impression, which can differ case by case (Bansal & Clelland 2004; Bolino, Kacmar, Turnley & Gilstrap, 2008). A common method is to articulate the achievements made by the company (Delmas & Burbano, 2011). However, an alternative strategy is that companies choose not to disclose such information and remain “strategically silent” (Carlos & Lewis, 2018). When companies deliberately withhold information regarding environmental achievements, they are brownwashing (Kim & Lyon, 2014). Some companies even become certified, and are consequently allowed to articulate this environmental friendliness, but choose not to (Rauber, 2006). As an illustration, IKEA decided not to disclose that a substantial percentage of the cotton they use is certified as the Better Cotton Initiative (Testa et al. 2018, p3). Given the focus of this article the following definition is used: when the product is produced in an environmentally friendly manner, but there is no label visible, the product is brownwashed.
contributing to these costs. Consequently, they expect a price premium and are repulsed from investing in or buying from the company. This behaviour negatively affects the company’s performance, and with that, it is expected that firms are less likely to disclose their environmental achievements.
When certain environmental characteristics of the product are not articulated, the image of the product towards the customers is modified (Delmas & Burbano, 2011). By disclosing information selectively, the perception of the product formed by the customer is based on incomplete information. This act of misleading the customers alters the image of the product and with that influence’s customers behaviour. Given the seriousness of brownwashing, it is surprising that this phenomenon has received little academic attention (Kim & Lyon, 2015).
As motivated before, this paper will study brownwashing on a product level. One of the primary methods how companies can influence the customer’s impression of a product is through labels. By doing so, the image of the product is influenced. Labels can be attained by companies when they comply to a particular set of criteria. For example, in order to receive a organic certificate, the product should be produced by a particular method. When a company successfully applies for a label, it can use it on their products. The official goal of a green label is to provide credible information related to the environmental friendliness of the product. Also, customers are unaware of the environmental friendliness of the product when they are faced with it in the store. Utilising green labels can reduce this information asymmetry (Crespi & Marette, 2005; Delmas & Grant, 2014). At the same time, the implicit goal of a label is to appear superior to other products (Crespi & Marette, 2005) and to persuade customers to buy the product (Leire & Thidell, 2005). The effectiveness of a label is determined by the credibility and associations attached to a particular label. When consumers are faced with unfamiliar labels, they tend to be more suspicious and will start to question their credibility (Ibanez & Grolleau, 2008). Since there is no elaborate additional information about the certifying organisation on the product, this question remains unanswered. This decreases the trust towards the label and therefore, to the product (Delmas & Grant, 2014). Because of this, consumers tend to hold back which decreases the likelihood of buying the product.
might view a vegan label as a positive aspect and are therefore drawn to it. On the other hand, resolute, carnivores might be less attracted to buy these products, as it challenges their personal views. In other words, to effectively sell products, it is essential to follow an appropriate labelling strategy (Mason, 2006). When the company believes that its customer group will not appreciate the label, they are expected not to use it.
Apart from the previously mentioned reasons for brownwashing, this paper will explain why the quality segment of the product is another estimate for brownwashing behaviour. In the following section, it will be argued how much brownwashing is expected per quality segment. Arguments are provided for the low-, medium- and high quality segments. This universal segmentation allows generalising the results to different product groups. In the end, the expectations are joint together into one main hypothesis.
Recent research found that, in the low-quality segment, labels can negatively influence the perception of products (Newman et al., 2014; Delmas & Lessem, 2017). There are two explanations for this. First, this is due to the quality of initial products that bore these labels. These were often experimental products and had an even more experimental level of quality. This unpleasant experience caused consumers to have negative associations towards these type of products (Galarraga Gallastegui, 2002; Peattie & Crane, 2005).
Nonetheless, this theory expects an adverse effect on consumer behaviour when confronted with a vegan label. This makes the product appear less attractive, and can negatively influence the sales of the product. Given this adverse effect, wineries are expected to under-emphasise the environmentally friendly aspects of their product. They will do so in order to manage the impression given by the product. In short, relatively few vegan labels are expected in the low-quality segment.
For medium-priced products, Luchs et al. (2014, p.2) found that the negative effect of labels on the perceived quality, as described in the previous section, diminishes when ‘a minimum threshold of functional performance’ is met. They argue that consumers believe that there are sufficient resources left to be invested in the quality, and therefore do not perceive the product to be of relatively lower quality.
Moreover, the environmental friendliness of the product might be an additional reason for customers to buy the product. Given the increased interest in such products (Martinez & Rodriguez, 2013; Lii & Lee, 2011), more consumers can identify themselves with a green label. Products carrying a vegan label can help to target this new customer segment. The vegan label becomes a differentiating aspect of the product which might influence sales. To conclude, a vegan label seems to become a unique selling point in the medium quality segment.
Unique selling points make the product appear more attractive and can positively influence the sales of the product (Levitt, 1986). Given the positive effect on consumer behaviour, wineries are less reluctant to reveal that the wine is vegan-friendly. They will do so in order to positively influence the impression of the product. However, since veganism is not yet fully integrated nor required by law, it is not expected that all vegan wines will carry a vegan label. In summary, relatively many vegan labels are expected in the medium-quality segment
customers take the greater good into account when choosing a product. For example, buying a locally produced product. This type of value also includes the environmentally friendly characteristics of the product, such as vegan aspects. The second type concerns self-enhancement values (e.g. status and superlative wine enjoyment) (Ahtola, 1985). These aspects are expected to give short-term emotional feelings. Customers buy this type of product for personal pleasure. Scholars found that when both values are present in one product, the perceived amount of enjoyment from the product is reduced (Torelli et al. 2012; Raghunathan et al., 2006). As an illustration, organic Fairtrade vegan (Self-transcendent) potato chips (self-enhancement) sounds less appealing for most customers than an ordinary version. Customers assume that when consuming such potato chips, there is more of a sense of doing something rightful than enjoying the product. In other words, the expected amount of enjoyment of consuming a high-quality product is decreased when it carries a label (Torelli et al., 2012; Raghunathan et al., 2006). This situation is also applicable to the wine industry. Consumers of high-quality wine often buy it for self-enhancement values such as the winery’s reputation or the ecstatic experience of tasting the high-quality wine (Workman, 2016). Consequently, the occurrence of vegan labels on high-high-quality wines is expected to have a negative effect on the perceived enjoyment of the wine. As a result, this decreases the buying behaviour of the customers (Raghunathan et al., 2006).
Additionally, from an ethical point of view, consumers who are financially capable of buying high-quality products tend to be less concerned about the environmental impact of their actions (Piff et al., 2010). This type of customer tend to prioritises their own happiness and buys products which they personally prefer. The necessity of buying a product that is produced environmentally friendly is of less priority. Consequently, they are less drawn to products that disclose their environmental friendliness.
product, will under-emphasise the environmental aspects of their product. In short, relatively few vegan labels are expected in the high-quality segment.
In conclusion, the three quality segments all have different motivations for the expected amount of brownwashing. For products in the low-quality segment, there are relatively few vegan labels predicted due to negative associations. In the medium-quality segment, vegan labels are estimated to occur relatively often. In the high-quality segment relatively few vegan labels are expected. Together, these predictions lead to the following main hypothesis.
METHODOLOGY
In the following section, the research design will be presented. In the first part, the selection of the database and the stores are explained. This is followed by, the form of data collection and additional desk search. After that, the choices for the dependent, independent and the four control variables are justified. The research design will conclude with a description of the analytical method and the type of regression.
To establish a set of vegan wines, the database of Barnivore is utilized. This organisation holds a list of vegan liquors that are sold worldwide. Based on this database, Rick Smind (2012) has made a selection of wines which are sold in The Netherlands. Some modifications were made to this list. For example, online shops were excluded from the research. The reason for this exclusion is that websites often only show the front, while most of the vegan labels are presented on the back of the bottle. This form of display would made it impossible to verify whether the wine carried a vegan label. This exclusion is not expected to skew the results, since online wine sales only represent a few percents of market share (Driessen & Morren, 2015).
Another modification to the list concerns the vintages. Some wines in the database carried the vintage 2013, while the store sold 2015. Naturally, this is not expected to have any effect on this study. However, the updated wines where doubly verified via Barnivore to ensure the wines were still suited for vegans.
Furthermore, some wines had to be deleted from the sample because they were no longer sold at the store. Stores continuously improve their product portfolio and sometimes choose to add or delete products. This selection had only a small effect on the numbers of wines in the dataset. In total, the dataset counts 156 vegan wines. To emphasise, all the wines in this study are made in a vegan-friendly manner, and they are allowed to express this on the bottle.
To collect the data, each of the stores were visited. Most of the stores did not offer all the vegan wines which were in the dataset, due to different local assortments. To overlap this problem, stores were visited at several locations and cities. This enabled the inclusion of all of the wines. When visiting the stores, each of the 156 wines were manually examined on the absence of a vegan label, which is, as explained before, the measure of brownwashing.
Additionally, a desk search was performed to retrieve the information on the size of the winery (this is motivated in the following section). This information was mostly attained via online sources. In case of missing values, the winery or the national importer has been contacted by email or telephone. Furthermore, other characteristics were recorded, such as the price of the wine, region, country and store (elaborated in the next section).
Measurements
Previous studies have identified multiple variables which are important to take into consideration when studying the wine industry. Moreover, it is essential to control several aspects. These previously defined measurements were the foundation for the following dependent and independent variable, plus four control variables.
Vegan label
The dependent variable is the occurrence of a vegan label on the wine bottle. When a bottle bears a vegan label, the product is not brownwashed, while when the label is not visible, brownwashing occurs. The amount of wines without a label determines the level of brownwashing (Delmas & Lessem, 2017).
Quality
The independent variable is the quality of the wine. There are several ways to determine the quality of the wine. Unfortunately, it was not possible to personally assess the quality of each of the wines. However, several scholars have taken a hedonic approach to determine the quality of the wine by decomposing it into observable attributes (Rosen, 1974; Bombrun & Sumner, 2001). They used the scores of wine critics to determine the quality of each wine. They found that especially price is a significant measurement for the quality of the wine. In this study, it was not possible to use critic scores in the first place, for the following reasons. First, some wines had received reviews from several non-comparable wine critics. This is a large constraint because the scores can strongly vary between critics (e.g. some wines simultaneously receive a bronze, silver and gold medal) and because most critics use a different scale (e.g. scores ranging between 70-100, 12-20, or up to three glasses, four stars five medal types). As a result, using different wine critic scores would make it very difficult to establish an objective estimate of the quality of the wine.
Second, it was not possible to use one wine critic. The person that gave the most reviews was Harold Hamersma, but he ‘only’ scored 91 of the wines. Using his scores as a measure of quality would sharply decrease the number of observations, reducing the reliability of the study. However, his scores were used differently: they were used to test whether the price of the wine is indeed a valid measurement within this particular dataset. To test the relation between scores of Hamersma (2017, 2018a) and the price of the wines, a correlation has been conducted (r = 0.5427,
p = .000). The results show that price is indeed a rightful measurement for quality. Therefore, price
is used in this study as independent variable. This measurement of quality is in line with previous research (Oczkowski, 1994, 2001; Landon & Smith, 1998; Bombrun and Sumner 2001; Thrane, 2004; Lockshin et al., 2006; Mtimet & Albisu, 2006; Delmas & Grant, 2014; Delmas & Lessem, 2017).
In the case of temporary discounts, the original price was registered. Additionally, other factors that might affect the reliability of price were removed. Sparkling wines were excluded since these wines often carry a price premium to compensate for an expensive production process (Wine & Spirit Education Trust, 2012). Furthermore, sulphur free wines were ignored since these bottles carry large R&D premiums. In order to test the inverted U-Shape character of the hypothesis, a squared version of this variable was incorporated in the analyses and called price2 (Lind &Mehlum,
Firm characteristics
In this study, a control variable is created for firm characteristics measured through firm size (Delmas and Burbano, 2011; Kim & Lyon 2014). Hartmann (2011) explained that a firm’s brand orientation could influence its engagement in the field of responsible social and environmental behaviour. Companies with strong brands (such as large wineries) tend to disclose environmental achievements more often in order to protect themselves from criticism (Luhmann & Theuvsen, 2016). Consequently, the size of the company might influence the occurrence of a vegan label. To measure the size of the company, the total amount of hectares of vines that belongs to the winery was used (including plots for non-vegan wines). Additional aspects that might influence this number, such as forests or other agricultural property, were excluded. This total number represents the size of production and organisation and therefore is an indication for firm size. The number of hectares was transferred into a log-scale to account for the non-linearities.
Country
Any effect from the country of origin is also accounted for (Hu & Baldin, 2018; Delmas & Grant, 2014). In New World wine countries (countries outside of Europe), producers are less traditional and tend to experiment more with the layout of the wine bottles, compared to Old World countries (European countries) (Wine & Spirit Education Trust, 2012). This difference could result in different labelling strategies between countries. Additionally, the varying cultures between countries might also influence whether green labels are integrated and whether brownwashing is more common.
In the dataset, there were some countries with only a few observations. This grouping would cause omitted or empty findings during the analysis, which would exclude these wines from the study. To overcome this problem, two country combinations were made. Firstly, New Zealand and Australia were juxtaposed into one category. Secondly, Chile and Argentina were merged. Both country combinations have geographical proximity, and for many years have carried a comparable wine identity and label style (Paronetto, 1985; Naudin & Flavigny, 2004; Wine & Spirit Education Trust 2012). The other wine countries had a sufficient amount of observations and differed too strongly; therefore, no further combinations were required.
Region
The region of production is also included as a control variable. This measure is coherent with the study of Oczkowski (1994, 2001) and Delmas and Grant (2014), since the type of region affects various aspects such as wine- and label style. The Wine & Spirit Education Trust (2012) classified three types of wine regions where a wine can originate from; ‘National’, where the grapes from the entire country are allowed in the production (e.g. Vin de France); ‘Province’, where the wine can only come from the province (e.g. Bordeaux); ‘Appellation’, where the grapes have to be grown in the specified region (e.g. Saint Emilion). The regulations for production are increasingly strict along the different segments. Consumers are usually more drawn to Appellation wines, due to higher quality associations. This effect makes it easier to sell them, compared to Province or National wines (Wine & Spirit Education Trust, 2012). Therefore, the differences between region types might affect the use of vegan labels. Furthermore, some regions have a sufficient image and do not want to risk spoiling it with the negative associations attached to a green label (Delmas & Grant, 2014).
Moreover, the region of the wine might affect the price of the product. The more specific a region is, the higher the production costs are. This premium is due to stricter production regulations and less room for buying wine from other farmers to supplement or dilute the wine in case of flaws (Naudin & Flavigny 2004). In some regions, consumers pay a premium for the origin of the wine. For example, Thrane (2004) found that wines from Burgundy are on average more expensive than wines from other regions. Given the possible influence of region type, this aspect is controlled for. The three region types are translated in two dummy variables.
Store
The store is added as a control. Some stores are focusing more on vegan products, e.g. Tesco has created an own vegan product line (Quinn, 2017). Additionally, supermarkets that are considered to sell organic products, e.g. EkoPlaza (Stroeken, 2015), tend to have more environmentally conscious customers. These customers usually prefer products with a lower environmental impact. In order to provide for these needs, the product portfolio of those supermarkets might have an over-representation of responsible/vegan-friendly products.
might have less impact when confronted with vegan-labeled products. Contrarily, supermarkets that do not have an overage of environmentally conscious customers might, for the same reasons, choose to have a product portfolio with fewer vegan labels. In other words, the amount of vegan labels might be influenced by the store assortment. In total there are five different stores, resulting in 4 dummy variables.
Analytical method
In order to test the hypothesis, a probit regression will be conducted. This type of regression is suitable when the dependent variable is dichotomous (Brooks; 2014; Albright, 2015). As mentioned before, the dependent variable is 1 when there is a vegan label and 0 when there is no disclosure of veganism. When the value of the probability of a vegan label decreases, the amount of brownwashing increases.
To account for the possibility of heteroskedasticity in the model, a robust standard error was incorporated during the probit regression (Giles, 2013). In order to test the relationship between vegan label occurrence and price, the following formula has been used:
Pr(Yi=1) = α + 𝛽1𝑋1 + 𝛽1𝑋12 + 𝛽1𝐶1+ 𝛽2𝐶2+ 𝛽2𝐶2+ 𝛽3𝐶3+ εi
RESULTS
This section displays the statistical results derived from this study. First, the description and correlation of the dependent variable, independent variable and the one continuous control variable are provided. Second, the results of the probit analysis are discussed, and it is thoroughly analysed whether the hypothesis holds. Last, a test to verify the joint significance of the control variables is conducted.
Descriptive statistics
Table 1 reports the summary statistics and correlation matrix. The results show that 35,3% of the
wines had a vegan label. With 156 observations, this results in 55 wines. Firm size is the number of hectares but transformed into a log scale. The real number of hectares that belong to the firm range from 9 up to 3600, with a mean of 853 hectares and a standard deviation of 911 hectares. The results show a negative correlation between firm size and price. The other control variables, including an additional analysis, are described in appendix A.
Table 1 Description and Correlations
Variable Mean Sd. Min Max Vegan label Price
Vegan label 0.353 0.479 0 1 -
Price 7.074 2.108 3.79 13.99 0.207* -
Firm Size 5.793 1.687 2.197 8.189 0.008 -0.419*
* p < .01
Since Table 1 showed that there are no high correlations, no multicollinearity is expected. Nonetheless, to ensure that multicollinearity does not play a part with the other variables, a Variance Inflation Factor (VIF) is conducted. The mean VIF denotes 3.68, and the highest VIF is 5.6. Since none of the variables exceeded the >10 threshold, no multicollinearity is present in the data (O’Brien, 2007; Robinson and Schumacker, 2009).
Probit regression
Table 3 displays the results of the probit regression. Under the variables: country, region and store
Table 3 Results Probit regression
Vegan label B Robust SE
Independent variable Price 1.251** (0.442) Price2 -0.069* (0.027) Control Variables Firm Size 0.007 (0.091) Country Spain 0.781 (0.845) Chile/Argentina 1.042 (0.819) France -0.303 (0.833) Germany 2.270** (0.878) Italy 2.023* (0.907) South Africa 0.080 (1.051) Region Province 0.504 (0.349) Appellation -0.055 (0.520) Store Ekoplaza -0.239 (0.404) Albert Heijn 0.783 (0.493) HEMA -0.716 (0.607) Jumbo -0.742 (0.645) Constant -6.575** (2.095) N 156 Wald chi2 59.80
Log pseudo likelihood -67.779
Pseudo R2 0.330
* p < .05, ** p < .01
in Stata). This test verifies if the quadratic coefficient is significant, plus it confirms whether the extreme point is within the dataset, and whether the function is significantly increasing in the lower bound and decreasing in the upper bound (U-shape or V-shape). A part from these separate aspects, the test verifies whether an inverted U-shape relation is present throughout the entire model. The results of the test are presented in Table 4. The findings for the lower bound indicate that the slope is positive and significant (p < .01), i.e. the function is increasing. The results for the upper bound show that the slope is negative and significant (p < .05 level), i.e. function is decreasing. These results confirm that the relationship is not V-shaped.
The Fieller interval estimation falls within the range of the data set. The extreme point is estimated at €9,05. Therewith, it is concluded that there is a highest point in the dataset, i.e. the function is not merely concave, but there is an extreme point. The overall test of the presence of an inverted U-shape throughout the model is also significant (t = 2.00, p = .024). Consequently, the criteria for an inverted U-shape relation are met (Lind & Mehlum, 2010).
Table 4 Results on the U-shape test Lower Bound Upper Bound Interval 3.79 13.99 Slope 0.727 -0.681 t-value 2.928 -1.998 P> ltl 0.002 0.024
95% Fieller interval for extreme point: [7.52; 13.77]
Graph 1 Vegan label occurrence and price
On the grounds of all the results mentioned above, there is enough evidence to conclude that an inverted U-shape relation between vegan label occurrence and price is present. Therefore, the main hypothesis is supported.
Additional analysis
DISCUSSION
Scholars have studied the possibility for companies to remain strategically silent about attainments (Carlos & Lewis, 2018). Kim and Lyon (2014) specifically identified the situation where companies choose to withhold information on their environmental achievements. The existing knowledge on brownwashing shows that this phenomenon occurs in order to manage the impression of the company. One of the main reasons for brownwashing is to prevent negative associations towards the products, such as additional costs (Fisher-Vanden & Thorburn, 2011). Furthermore, environmentally friendly aspects are understated when the company believes that its customers do not identify with it (Delmas & Grant, 2014). Also, brownwashing occurs in order to keep the stakeholders satisfied (Jacobs, Singhal & Subramanian, 2010). By omitting to denote environmentally friendly aspects, the company’s profile is affected. Although research has identified brownwashing on a firm level (Kim & Lyon, 2015; Testa, 2018), no articles are known to have investigated brownwashing on a product level. Consequently, there is scarce knowledge as to when products are brownwashed. By contributing to research on impression management (Carlos & Lewis, 2018) and brownwashing (Kim & Lyon, 2015; Testa, 2018) this thesis has aimed to fill this research gap. This study has been conducted to analyse in which type of quality segments brownwashing is more likely to occur. The results show that product quality has an inverted U-shape relation to vegan labels. In this section, the results of the thesis are discussed. The contribution to the literature is elaborated by presenting a new factor of brownwashing and evidence for brownwashing on a product level. Furthermore, practical implications for label organisations, organisations and customers are presented. Also, the limitations of this research along with interesting future research directions are given.
Theoretical implications
brownwashing can be expected. It seems that there is a market-based reason for companies to brownwash their products.
The motivations for producers to brownwash in the low- and high-quality segment have a different theoretical explanation. Previous work showed that low-quality products are often brownwashed due to the customers’ lay theory (Chernev & Carpenter 2001; Chernev 2007). Another reason for producers to brownwash a product is not to incite the negative influence a label has on the perceived product quality. When a low-quality product also has a green label, it will have a lower perceived quality (Galarraga Gallastegui, 2002; Peattie & Crane, 2005).
In the high-quality segment, there is a different theoretical explanation for the observed brownwashing by producers. Negative previous experiences with environmentally friendly experimental products will create the impression that the product is of lower quality (Delmas & Lesseman, 2017). Also, conflicting purchase values decrease the anticipated enjoyment of the product (Torelli et al., 2012; Raghunathan et al., 2006). This theoretical explanation shows that there are different explanatory motivations for brownwashing throughout the segments.
An additional explanation for the amount of brownwashing is that some producers might believe that a green claim should not be a valid reason for purchasing their product. They believe that the quality of the product should be the reason for purchase and that the product should speak for itself. In honour of the product, producers are reluctant to use these labels. Since product quality is a reason for purchase in the high-quality segment, in particular (Workman, 2016), this explanation is especially applicable to the results of the high-quality segment. Producers are likely to brownwash their products in order to anticipate to these unwanted effects. When a green claim is expected to decrease sales, producers are likely to omit the label from the product. In sum, the quality of the product affects the amount of brownwashing.
Practical implications
labels could be increased (Ibanez & Grolleau, 2008; Mason, 2006). By doing so, the possible reticence of customers towards products with such labels could be decreased. Consequently, the occurrence of a label could not only become a unique selling point in the medium-quality segment but the low- and high-quality segments (Martinez & Rodriguez, 2013; Lii & Lee, 2011).
Limitations and future research
This study is limited in several ways. First, the relationship between labels and purchase decisions have not been examined. The article used previous theories on label strategies to determine when customers are repulsed by or attracted to a product. However, product labels differ between each other, and some labels are better integrated than others. Especially, vegan labels are vastly increasing in occurrence (Terrachoice, 2010), and consequently also the acceptance and recognition of customers towards these labels. Possibly, the aversion to such labels might change in time. When customers are frequently confronted with vegan labels, customers might become more acquainted with the label. Consequently, the unfamiliarity and negative associations towards a label might change accordingly. As a result, producers might feel more free to use these label. The amount of brownwashing might be different for well-integrated labels than for emerging ones. Future research could focus on how different green labels affect the buying behaviour. For example, a well-integrated organic label versus an emerging vegan label. This could provide a deeper insight into the preferences of customers.
Second, the amount of observations per control variable were not representative in some cases. Since this was not part of the hypothesis, it is perceived as a minor limitation to the study. However, it does provide an interesting future research direction. As mentioned before, cultural differences between countries might affect the labelling strategy (Paronetto, 1985; Naudin & Flavigny, 2004; Wine & Spirit Education Trust 2012). Therefore, it could be possible that brownwashing differs between countries. It is not unimaginable that some countries are more conventional and less concerned about the environmental impact of products. Consequently, products might be more brownwashed in these settings. It would be an interesting future study to investigate whether this affects the likelihood of brownwashing. This proposition is not limited to countries but can be focussed on any cultural differences. Such research can verify whether the findings are universally applicable.
of high-quality wines. Therefore, it is advised to conduct a future study, including the previously mentioned aspects, to verify the findings of this study on high-quality products.
CONCLUSION
This thesis has further developed the theory on impression management and brownwashing. By studying brownwashing on a product level, a new facet of brownwashing was analysed. In order to analyse different quality segments, the following research question was formulated:
In which product quality segments is brownwashing more likely to occur?
LITERATURE
Ahtola, O.T. (1985). Hedonic and utilitarian aspects of consumer behavior: An attitudinal perspective. Advances in Consumer Research, 12, 7-10.
Bansal, P., & Clelland, I. (2004). Talking trash: Legitimacy, impression management, and unsystematic risk in the context of the natural environment. Academy of Management Journal,
47(1), 93–103.
Berrone, P., A. Fosfuri, L. Gelabert, & L. R. Gomez-Mejia. (2013). Necessity as the Mother of ‘Green’ Inventions: Institutional Pressures and Environmental Innovations. Strategic
Management Journal 34, 891–909.
Bombrun, H., & Sumner, D. (2001). What determines the price of wine? AIC Issues Brief, 18, 1-6.
Boiral, O. (2016). Accounting for the Unaccountable: Biodiversity Reporting and Impression Management. Journal of Business Ethics, 135, 751–768
Bolino, M. C., Kacmar, K. M., Turnley, W. H., & Gilstrap, J. B. (2008). A multi-level review of impression management motives and behaviors. Journal of Management, 34(6), 1080–1109. Boulton, R. B., Singleton, V. L., Bisson, L. F., Kunkee R. E. (1999). Principles and Practices of
Winemaking, New York City: Springer US
Brooks, C. (2014). Introductory econometrics for finance. Cambridge: Cambridge University Press.
Burritt, R., Christ, K. (2016). Industry 4.0 and environmental accounting: A new revolution?
Asian Journal of Sustainability and Social Responsibility. 1, 23–38.
Carlos, W. C., & Lewis, B. W. (2018). Strategic silence: Withholding certification status as a hypocrisy avoidance tactic. Administrative Science Quarterly,63(1), 130–169.
Carlsson-Kanyama A., Gonzalez, A. (2009). Potential contributions of food consumption patterns to climate change. American Journal of Clinical Nutrition 89, 1S–6S.
Chernev, A. & Carpenter, S. (2001). The Role of Market Efficiency Intuitions in Consumer Choice: A Case of Compensatory Inferences, Journal of Marketing Research, 38 (August), 349– 61.
Combris, P., Lecocq, S., & Visser, M. (1997). Estimation of a hedonic price equation for Bordeaux wine: Does quality matter? Economic Journal, 107, 390-402.
Conti, G., Ruiz-Molina, M.E., Gil-Saura, I. (2010), Instruments for wine promotion in upscale
restaurants, Journal of Foodservice Business Research, 13 (2), 98-113.
Crespi, J. M., & Marette, S. (2005). Eco-labelling economics: Is public involvement necessary?, Northampton, MA: Edward Elgar.
Daniel B. (1957). "Use of Dummy Variables in Regression Equations". Journal of the American
Statistical Association. 52(280), 548–551.
Delmas M.A., Burbano, V.C. (2011). The drivers of greenwashing. Californian Management
Review 54(1), 64–87.
Delmas, M., Grant, L. (2014). Eco-Labeling Strategies and Price-Premium: The Wine Industry Puzzle. Business and Society, 53(1), 6-44.
Delmas, A., Lessem N. (2017). Eco-Premium or Eco- Penalty? Eco-Labels and Quality in the Organic Wine Market, Business & Society 56(2), 318-356.
Doveh, E., Shapiro, A. & Feigin, P. D., (2002). Testing of monotonicity in parametric regression models, Journal of Statistical Planning and Inference, 107, 289–306.
Eichholtz P., Kok N., Quigley, J. (2010). Doing well by doing good? Green office buildings.
American Economic Review. 100, 2492-2509.
European Union EU (2012) Commission Implementing Regulation (Eu) Official Journal of the
European Union. 203.
Fisher-Vanden, K., Thorburn, K. S. (2011). Voluntary corporate environmental initiatives and shareholder wealth, Journal of Environmental Economics and Management, 62(3). 430-445, Friedman M. (1970). The social responsibility of business is to increase its profits. New York Times Magazine, 9(13), 32–33.
Galarraga Gallastegui, I. (2002). The use of eco-labels: A review of the literature. European
Garnett, T. (2011). Where are the best opportunities for reducing greenhouse gas emissions in the food system (including the food chain)? Food Policy. 36, S23–S32
Goffman, E. (1959). The Presentation of Self in Everyday Life. New York: Random house. Gonzalez, A., Frostell, B., Carlsson-Kanyama, A. (2011) Protein efficiency per unit energy and per unit greenhouse gas emissions: potential contribution of diet choices to climate change mitigation. Food Policy 36, 562–570
Gray, R. (2006). Social, environmental and sustainability reporting and organisational value creation. Accounting, Auditing and Accountability Journal, 19(6), 793–819.
Haans, R. F., Pieters, C., & He, Z. L. (2016). Thinking about U: Theorizing and testing U-and inverted U-shaped relationships in strategy research. Strategic Management Journal, 37(7), 1177-1195.
Hamersma, H. (2018a) De Grote Hamersma *Speciale Editie* De 1030 lekkerste biologische
wijnen. Amsterdam: Fontaine uitgevers.
Hamersma, H. (2018b). De Kleine Hamersma, de beste wijnen tot 7,50. Amsterdam: Fontaine uitgevers
Hamersma, H. (2017) De Grote Hamersma. Amsterdam: Fontaine uitgevers.
Hartmann, M. (2011). Corporate social responsibility in the food sector. European Review
of Agricultural Economics, 38(3), 297–324.
Hu, L. Baldin A. (2018). The country of origin effect: a hedonic price analysis of the Chinese wine market. British Food Journal 120(6), 1264-1279.
Ibanez, L., & Grolleau, G. (2008). Can ecolabeling schemes preserve the environment?
Environmental Resource Economics, 40, 233-249.
Jacobs, B. W., Singhal, V.R., Subramanian, R. (2010). An empirical investigation of
Khan, S. A. R., & Dong, Q. (2017). Impact of green supply chain management practices on firms’ performance: An empirical study from the perspective of Pakistan. Environmental Science and Pollution Research, 24(20), 16829–16844.
Kim, E. H., Lyon, T. P. (2015). Greenwash vs. Brownwash: Exaggeration and Undue Modesty in Corporate Sustainability Disclosure. Organization Science. 26(3):705-723
Kittilaksanawong, W. (2019). Institutional Perspective on the Drivers of Green Innovations in Firms. In I. Management Association (Ed.), Green Business: Concepts, Methodologies, Tools,
and Applications (pp. 1009-1019). Hershey, PA: IGI Global.
Landon, S., & Smith, C. E. (1998). Quality expectations, reputation, and price. Southern
Economic Journal, 64, 628-647.
Laursen, K., Salter, A. (2006). Open for Innovation: The Role of Openness in Explaining Innovation Performance among U.K. Manufacturing Firms Strategic Management Journal,
27(2), 131-150.
Lii, Y. & Lee, M. (2011). Doing Right Leads to Doing Well: When the Type of CSR and Reputation Interact to Affect Consumer Evaluations of the Firm. Journal of Business Ethics,
105(1), 69-81.
Leire, C., & Thidell, A. (2005). Product-related environmental information to guide consumer purchases—A review and analysis of research on perceptions, understanding and use among Nordic consumers. Journal of Cleaner Production, 13, 1061-1070.
Levitt, T. (1986). The marketing imagination (New, expanded ed.). New York: Free Press. Lind, J. T., & Mehlum, H. (2010). With or without U? The appropriate test for a U-shaped relationship. Oxford Bulletin of Economics and Statistics, 72(1), 109-118.
Lockshin, L., Jarvis, W., d’Hauteville, F., & Perrouty, J. (2006). Using simulations from discrete choice experiments to measure consumer sensitivity to brand, region, price, and awards in wine choice. Food Quality and Preference, 17, 166-178.
Luhmann, H., & Theuvsen, L. (2016). Corporate social responsibility in agribusiness: Literature review and future research directions. Journal of Agricultural and Environmental Ethics, 29(4), 673–696.
Luchs, M.G., Brower, J., Chitturi, R. (2012). Product choice and the importance of aesthetic design given the emotion-laden trade-off between sustainability and functional performance
Journal of Product Innovation Management, 29, 903-916
Martínez, P., & del Bosque, I. R. (2013) CSR and customer loyalty: The roles of trust, customer identification with the company and satisfaction. International Journal of Hospitality
Management, 35, 89-99.
Mason, C. F. (2006). An economic model of ecolabeling. Environmental Modeling and
Assessment, 11, 131-143.
Milne, M. J., Tregidga, H., & Walton, S. (2009). Words not actions! The ideological role of sustainable development reporting. Accounting, Auditing and Accountability Journal, 22(8), 1211–1257.
Mtimet, N., & Albisu, L. M. (2006). Spanish wine consumer behavior: A choice experiment approach. Agribusiness, 22, 343-362.
Naudin, C., Flavigny, L. (2004). Larousse du vin, Paris: Tous les vins du Monde.
Newman, G. E., Gorlin, M., Dhar, R. (2014). When Going Green Backfires: How Firm Intentions Shape the Evaluation of Socially Beneficial Product Enhancements, Journal of Consumer
Research, 41(3.1), 823–839.
O’Brien, R. M. (2007). A Caution Regarding Rules of Thumb for Variance Inflation Factors Quality & Quantity. 41(5), 673.
Oczkowski, E. (1994). A hedonic price function for Australian premium table wine. Australian
Journal of Agricultural Economics, 38, 93-110.
Oczkowski, E. (2001). Hedonic wine price functions and measurement error. Economic Record,
77, 374-382.
Peattie, K., & Crane, A. (2005). Green marketing: Legend, myth, farce or prophesy? An
International Journal, 8, 357-370.
Piff, P. K., Kraus, M., Côté, S., Cheng, B. H., & Keltner, D. (2010). Having Less, Giving More: The Influence of Social Class on Prosocial Behavior. Journal of personality and social
psychology, 99(5), 771-784.
Raghunathan, R., Naylor, R. W., and Hoyer, D. W. (2006) The Unhealthy, Tasty Intuition and Its Effects on Taste Inferences, Enjoyment, and Choice of Food Products. Journal of Marketing.
70(4), 170-184.
Robinson, C.,Schumacker, R. E. (2009). Interaction effects: centering, variance inflation factor, and interpretation issues. Multiple Linear Regression Viewpoints, 35(1), 6-11.
Robinson, J. (2007). The Oxford Companion to Wine. Third Edition, Oxford: Oxford University Press.
Rosen, S. (1974). Hedonic prices and implicit markets: Product differentiation in pure competition. Journal of Political Economy, 82, 34-55.
Sasabuchi, S. (1980). A test of a multivariate normal mean with composite hypotheses determined by linear inequalities, Biometrika, 67, 429–439.
Scarborough P., Appleby P.N., Mizdrak A., Briggs A.D., Travis R.C., Bradbury K.E., Key T.J. (2014). Dietary greenhouse gas emissions of meat-eaters, fish-eaters, vegetarians and vegans in the UK. Climate Change 125(2), 179–192.
Testa, F., Miroshnychenko, I., Barontini, R. & Frey, M. (2018). Does it pay to be a greenwasher or a brownwasher?. Business Strategy and the Environment. 27, 1104–1116.
Thrane, C. (2004). In defense of the price hedonic model in wine research. Journal of Wine
Research, 15(2), 123-134.
Torelli, C. J., Monga, A. S. B., & Kaikati, A. M. (2012). Doing poorly by doing good: Corporate social responsibility and brand concepts. Journal of Consumer Research, 38(5), 948-9.
Zoecklein B., Fugelsang K.C., Gump B., Nury F.S. (1999). Wine analysis and production. New York: Springer.
Webpages
Albright, J. (2015, February 25). What is the Difference Between Logit and Probit Models? Retrieved from: https://www.methodsconsultants.com/tutorial/what-is-the-difference-between-logit-and-probit-models/
Audsley E., Brander M., Chatterton J., Murphy-Bokern D., Webster C., Williams A. (2009).
How low can we go? an assessment of greenhouse gas emissions from the UK food system and the scope to reduce them by 2050. Retrieved from http://assets.wwf.org.uk/downloads/
how_low_can_we_go.pdf
Bonini, S., Oppenheim, J. (2008, September 16). Cultivating the green consumer. Retrieved from: https://faunalytics.org/cultivating-the-green-consumer/
Committee on Climate Change (2010) The Fourth Carbon Budget Reducing emissions through
the 2020s. Retrieved from
http://www.folkestone-hythe.gov.uk/webapp/lydd-airport/CORE%20DOCS/CD12/CD12.18%20%20Reducing%20Emissions%20Through%20the %202020s.pdf
Driessen, S. Morren, R. (2015, Januari 10) De wijnmarkt in Nederland: trends en cijfers, kansen
voor de horeca. Retrieved from:
https://insights.abnamro.nl/2015/01/de-wijnmarkt-nederland-trends-en-cijfers-kansen-voor-de-horeca/
Giles, D. (2013, May 8). Robust Standard Errors for Nonlinear Models. Retrieved from: https://davegiles.blogspot.com/2013/05/robust-standard-errors-for-nonlinear.html
Hamersma, H. (2015, March 26) Veganfriendly, Retrieved from https://www.nrc.nl/nieuws/2015/03/26/veganfriendly-a1472963 Hamersma, H. (2016, April 30) Dikbestofte flessen, Retrieved from:
Natexpo (2018, July). Vegan : the new standard in gastronomy, Retrieved from: https://natexpo.com/en/vegan-la-nouvelle-gastronomie/#_ftnref2
Quinn, I. (2017, May 23). Tesco plans new vegan logo for own-label products. Retrieved from: https://www.thegrocer.co.uk/buying-and-supplying/marketing/tesco-plans-new-vegan-logo-for-own-label-products/553126.article,
Rauber, C. (2006, October 22). Winemakers go organic in bottle but not on label. Retrieved from: http://sanfrancisco.bizjournals.com/sanfrancisco/ stories/2006/10/23/focus2.html Smind, R. (2012). Vegan database. Retrieved from: http://vegan.ricksmind.com/
Report Buyer (2017, June) Top Trends in Prepared Foods 2017: Exploring trends in meat, fish
and seafood; pasta, noodles and rice; prepared meals; savory deli food; soup; and meat substitutes. Retrieved from:
https://www.reportbuyer.com/product/4959853/top-trends-in- prepared-foods-2017-exploring-trends-in-meat-fish-and-seafood-pasta-noodles-and-rice-prepared-meals-savory-deli-food-soup-and-meat-substitutes.html#tab-preview
Stroeken, J. (2015, March 8). EkoPlaza bindt de strijd aan met Marqt, Retrieved from: https://www.trouw.nl/home/ekoplaza-bindt-de-strijd-aan-met-marqt~a91ca7f1/
Terrachoice, (2010). The sins of Greenwashing. Retrieved from: http://sinsofgreenwashing.com/index35c6.pdf
The Vegan Society. (n.d.). Definition of veganism, Retrieved from https://www.vegansociety.com/go-vegan/definition-veganism
WHO (2018) Global status report on alcohol and health 2018. Retrieved From:
APPENDIX A
In this section, the descriptive statistics will be presented in order to provide more insight into the control variables country, region and store. Additionally, an ANOVA or Welch test is performed.
Table A1 Descriptive statistics control variable country N With label Without label Mean Std. Deviation Minimum Maximum Australia/ New Zealand 7 1 6 5.707 1.720 3.99 7.99 Spain 31 9 22 6.313 1.602 3.79 10.49 Chile/ Argentina 37 17 20 6.326 1.485 3.99 8.49 France 40 7 33 7.002 1.693 3.99 11.89 Germany 10 8 2 8.972 1.424 6.79 9.99 Italy 15 11 4 7.643 2.118 3.79 11.99 South Africa 16 2 14 9.327 3.136 3.99 13.99 Total 156 55 101 7.073 2.107 3.79 13.99
By analysing the variable country as a factor variable for the dependent variable Price, the assumption of homogeneity of variances is violated. This is tested with the Levene test
F (6, 149) = 4.74 , p < .001. Therefore, the Welch test is performed instead of the ANOVA. Welch F (6, 38.22) = 7.01, p < .001.
Table A2 Descriptive statistics control variable region N With label Without label Mean Std. Deviation Minimum Maximum National 44 7 37 6.705 2.325 3.79 11.99 Province 85 35 50 6.666 1.669 3.99 11.89 Appellation 27 13 14 8.954 2.022 6.59 13.99 Total 156 55 101 7.073 2.107 3.79 13.99
For the control variable of region as a factor variable for the dependent variable Price, the assumption of homogeneity of variances is met. The Levene test showed F (2,153) = 1.813,
p < .167. The ANOVA test showed F (2,153) = 15.42, p < .001. Therefore, the means between the
Table A3 Descriptive statistics control variable store N With Label Without Label Mean Std. Deviation Minimum Maximum Gall&Gall 26 6 20 9.398 2.047 7.49 13.99 EkoPlaza 56 26 30 7.642 1.802 4.29 11.99 AH 36 15 21 6.065 1.464 3.99 9.98 HEMA 15 3 12 5.933 1.112 4.75 8.00 Jumbo 23 5 18 5.386 1.220 3.79 8.99 Total 156 55 101 7.074 2.108 3.79 13.99
The assumption of homogeneity of variances is violated for store as a factor variable for the dependent variable Price. This is tested with the Levene test F (4,151) = 3.853, p < .01. Consequently, the Welch test is performed instead of the ANOVA. The following results belong to the test of store and Price Welch F (4,59.45) = 24.25, p < .001.
