• No results found

Consumer Preferences influenced by Health Trends Product Attributes determining Willingness-To-Pay for Fast Moving Consumer Goods

N/A
N/A
Protected

Academic year: 2021

Share "Consumer Preferences influenced by Health Trends Product Attributes determining Willingness-To-Pay for Fast Moving Consumer Goods"

Copied!
78
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Consumer Preferences influenced by Health Trends

Product Attributes determining Willingness-To-Pay for Fast Moving Consumer Goods

(2)

2

Consumer Preferences influenced by Health Trends

Product Attributes determining Willingness-To-Pay for Fast Moving Consumer Goods

Anne Marije Tammenga

March 2012

Master Thesis

MSc BA Specialization Marketing Management and Research

University of Groningen, Faculty of Economics and Business

Lageweg 28, 9472 TE Zuidlaren

(06)10439850

anne_marije_tammenga@hotmail.com

Student number 1915983

(3)

3 MANAGMENT SUMMARY

Because of the increasing consciousness of consumers when making food choices, and preferences of consumers for more natural products, it is interesting to know what those trends imply for a brand. In particular, for brands operating in the market of Fast Moving Consumer Goods (FMCG) that can have a health claim the influence of those trends would be interesting, because of the possible unnatural additives in several products. Consequently, the purpose of this study is to answer the following main research question: How can a brand use different product attributes to increase customers Willingness-To-Pay (WTP) for Fast Moving Consumer Goods (FMCG) that can have a health claim in a market affected by the current health trends? To answer this question the product category ‘chilled orange juice’ has been investigated.

In summary, five investigated product attributes were proven to be important when considering WTP for Fast Moving Consumer Goods (FMCG) that can have a health claim, namely health claim, brand, packaging material, product ingredients and product taste. More specifically, consumers are willing to pay more for a product with a Healthy Choice logo. In addition, consumers clearly prefer the familiar brand. Furthermore, packaging material influences WTP for almost all consumers, with a strong favour for glass and a dislike for plastic. Moreover, for product ingredients a clear increase in WTP was shown for products with healthier and more natural product ingredients. Lastly, for taste no significant differences are shown.

In addition, within the market three segments could be determined. Segment one consists of around 50 percent of the total market and those consumers can be named as health conscious consumers. Those consumers clearly prefer glass, with Healthy Choice logo and with healthier and more natural ingredients. Moreover, they are willing to pay more for products branded by the familiar brand. For this product combination health conscious consumers are willing to pay € 4.76. Segment one is the most beneficial segment to target on. For segment two, the convenience conscious consumers, the same product attributes are preferred; however, the WTP is only € 1.55 for the ideal combination of attribute-levels. Lastly, segment three is the price sensitive group. This segment buys products based on price. The only shown preference is that when prices are equal they prefer the familiar brand over private label. Those consumers are willing-to-pay €1.50. Note that for all segments no significant taste preferences are detected.

(4)

4 PREFACE

To finalize my Master degree in Marketing Management and Marketing Research, I wrote this Master thesis. I enjoyed improving my Marketing knowledge over the past years and I am proud of the current Master thesis which reflects a part of this knowledge. To be able to finalize this thesis in such a satisfying way, I need to thank several people for their help.

Firstly, I would like to thank Marielle Non for her supervision during the whole writing process of my Master thesis. Due to her structured and critical feedback and the time she reserved to provide me with feedback in short time periods, I was able to write this Master thesis within seven months. In addition, I would like to thank my second supervisor, Sonja Gensler. Her feedback helped me to optimize my thesis in the final phase of the writing process. Furthermore, I would like to thank Edward Bonselaar, the brand manager of Appelsientje, who also provided me with feedback on the report during the writing process. Additionally, I would like to thank all respondents who fulfilled the questionnaire. Lastly, I would like to thank Jeroen and my family for their support.

I wish you a pleasant time reading my Master thesis.

Kind regards,

(5)

5 TABLE OF CONTENTS

1. INTRODUCTION ___________________________________________________________ 7

2. LITERATURE REVIEW ______________________________________________________ 11

2.1 (Un)healthy Eating Choices ___________________________________________________ 11 2.2 Willingness-To-Pay (WTP) ____________________________________________________ 13 2.3 Product Attributes __________________________________________________________ 15 2.3.1 Health Claim ______________________________________________________________________ 16 2.3.2 Brand ___________________________________________________________________________ 18 2.3.3 Packaging Material ________________________________________________________________ 19 2.3.4 Product Ingredients ________________________________________________________________ 21 2.3.5 Product taste _____________________________________________________________________ 21 2.4 Moderators ________________________________________________________________ 22 2.4.1 Age _____________________________________________________________________________ 23 2.4.2 Gender __________________________________________________________________________ 24 2.4.3 Health Consciousness ______________________________________________________________ 26 2.5 Conceptual Model___________________________________________________________ 27

3. RESEARCH DESIGN _______________________________________________________ 29

3.1 Research Method ___________________________________________________________ 29 3.2 Data Collection _____________________________________________________________ 30 3.3 Measurement Development __________________________________________________ 31

4. RESULTS ________________________________________________________________ 34

4.1 Demographics ______________________________________________________________ 34

4.1.1 Extreme Response Behaviour ________________________________________________________ 34 4.1.2 Descriptive Statistics _______________________________________________________________ 35 4.1.2.1 Factor Analysis ________________________________________________________________ 36

4.2 Choice Based Conjoint Analysis ________________________________________________ 38

4.2.1 Model Specification ________________________________________________________________ 38 4.2.2 Aggregate Model __________________________________________________________________ 38 4.2.3 Segment Model ___________________________________________________________________ 40 4.2.3.1 Segment Model with Covariates __________________________________________________ 41 4.2.3.2 Three Segment Solution ________________________________________________________ 42 4.2.4 Predictive Validity _________________________________________________________________ 46

(6)

6

5. CONCLUSIONS AND RECOMMENDATIONS ____________________________________ 51

(7)

7 1. INTRODUCTION

Nowadays, many consumers make their food choices more conscious than in the past (Trendystyle, 2011). Still, 41 percent of all consumers like to eat healthier in the Netherlands (Paulussen and Temminghoff, 2010), and those consumers are aware that they are responsible for their own consumption pattern. Nevertheless, they also blame manufactures (32 percent) and supermarkets (13 percent) for their unhealthy eating behaviour. Consequently, consumers expect supermarkets to offer a larger assortment containing healthier products. In addition, more than 60 percent of all consumers find it important that producers make their products healthier. Still, in general, 43 percent of all Dutch consumers have a difficult time to fulfil healthy eating goals. For this reason, increasingly practical information is requested about healthier eating, especially for younger women. Additionally, while shopping, almost 50 percent of the Dutch consumers demand information about which product is healthy and which is not. A tool which can be used to communicate this information is a Healthy Choice logo. Currently, a new Healthy Choice logo has been developed, where two individual logos merged (Pas, 2011). For the new logo two different versions are developed, namely a green and a blue logo. More specifically, the green one reflects the healthier alternative within basic products like bread and milk, where blue reflects a conscious choice within other food categories like soup and juices. Interestingly, more than 60 percent of the Dutch consumers are willing to pay more for healthier products.

Another trend is that consumers prefer more natural food, for example by using Stevia instead of Sucralose to accomplish sugar reduction (Innofood, 2010). Nowadays, when consumers consume enough fruit, vegetables and nuts, they are aware that they will meet the set standards for antioxidants without consuming products with added antioxidants, which are unnatural additives (Robbie, 2011). Thus, consumers prefer natural antioxidants instead of added antioxidants. Furthermore, consumers prefer to prepare their own food again (Irving, 2011). Consequently, they know that this food is fresh and the origin of the food is clear.

(8)

8 bacterium (Opzeeland, 2010). As a consequence, this product was very expensive. However, after investigations of Foodwatch, it has been found out that all yoghurt products include this bacterium. Accordingly, for consumers it is more beneficial to consume cheap yoghurt than to consume ‘Actimel’ products. Another product which was proven to have an untrue health claim about health benefits was ‘KnorrVie’ (Opzeeland, 2010). This product claimed to contain half of the recommended fruit and vegetable intake for one day. However, after research this claim was proven to be untrue. Therefore, the product is not available in the market any more. To conclude, fake claims will be discovered by consumers or/and organisations nowadays, because consumers are more critical.

However, it is not clear what the natural health trend and more critical customers mean for brands and their products. Also, expenditures on food decreased with 0.3 percent compared to previous year in the Dutch market (Hillen, 2011). Consequently, it would be of major interest to gather knowledge about attributes of food products where consumers are willing to pay for. In fact, while healthy eating trends have been heavily investigated in the past, previous research mainly focused on outdated health trends where consumers wanted to eat healthier by consuming food with unnatural additives, for example, added vitamins. Therefore, it is both academic and managerial relevant to identify product attributes for Fast Moving Consumer Goods (FMCG) that can have a health claim, which have an effect on WTP influenced by the current health trend. With those results brands can adapt to critical customers and the decrease in food expenditures. To be able to identify relevant product attributes for FMCG the effects of the current health trend on a specific product category are investigated in this report, namely chilled orange juices. This food category would be of major interest, because of added vitamins in fruit based juices, while this product can also be consumed freshly pressed. A distinction can be made within this product category between chilled and no-chilled fruit based juices. In fact, the market of chilled juices would be most interesting, because those products are more often compared to fresh pressed juices and inferred to be more natural than no-chilled juices. Within this research the most general fruit juice taste to investigate, namely orange juices.

(9)

9 product containing specific attribute-levels. Investigated product attributes are: health claim, brand, packaging material, product ingredients and product taste.

While investigating the relation between product attributes and WTP, the effect of the health consciousness of consumers on this relation is taken into account. In fact, because the health trend is rising in importance and is changing to a more natural healthy eating trend, this moderator is recognized as main influencer on the relation between product attributes and WTP. In addition, the influence of age and gender on the relationship are also examined.

(10)

10 where Healthy People and AH Sinaasappelsap do not. On the other hand, Healthy People contains the word ‘healthy’, which probably results in some health associations for consumers. Indeed, previous research shows that a healthy product name is interpreted as a healthier product than an unhealthy product name, for consumers who are on a diet (Bower, Saadat, and Whitten, 2003; Desai and Ratneshwar, 2003).

To conclude, the goal of this report is to increase our knowledge about the influence of several product attributes on WTP for FMCG that can have a health claim in a market shaped by healthy eating trends. With WTP information can be gathered about whether the current products should be redesigned, replaced or even eliminated. In addition, ideal combinations of product attributes can result in new products. Furthermore, pricing strategies can be developed. It is very important for brands to ask the optimal price. To be more specific, when a firm asks a price which is too low, profit and revenue levels will be lower than they could be. On the contrary, when a company charges too much, products will not be sold. In addition, the gathered knowledge can be used for marketing purposes. For instance, brands can use this knowledge to design products which meet customer preferences better, and where consumers are willing to pay more money for. Based on the results, managerial implications are provided, about how to set up products by using the examined product attributes influenced by health consciousness.

(11)

11 2. LITERATURE REVIEW

To find out which product attributes might influence WTP for FMCG that can have a health claim in a market affected by the current health trends, this literature review is written. The chapter starts with some general findings concerning food choices and healthy and unhealthy eating. Those findings may help to indicate important attributes that influence WTP. Secondly, WTP in general and influences on WTP in relation to the research question are taken into account. Next, five product attributes, which are expected to influence WTP, are described. Then information about the moderators is provided. Last, the conceptual model is shown, with an overview of relationships described in this chapter.

2.1 (Un)healthy Eating Choices

First of all, some general findings concerning food choices can be named. In fact, many factors can be named which influence food choices (Sobal and Bisogni, 2009). Within the food choice process model all those factors are captured within three dimensions, namely consumer experiences, culture and personal drivers. An example of consumer experiences is that in times of disruption consumers prefer familiar things, namely comfort food (Wood, 2010). Besides, consumers do not only prefer specific food in dynamic times, but also familiar logos and supermarkets (Cutright, 2012). In addition, personal drivers include the satisfaction created by food, the suitability of food in different situations, the healthy image of products and the price asked for it (Franchi, 2012). Other researchers added an additional factor to the three dimensions, namely the availability of the food (Delaney and McCarthy, 2011).

(12)

12 Secondly, several motives can be named for consumers to eat unhealthy. Namely, consumers who over consume unhealthy food are unwilling to change their unhealthy eating behaviour, they are unwilling to give up their favourite food, and/or they believe that the taste of unhealthy food is better (Lappalainen, Saba, Holm, Mykkanen and Gibney, 1997; Raghunathan, Naylor and Hoyer, 2006). The first two arguments are driven by internal motives. In other words, those motives do take place outside the consciousness of consumers. On the contrary, food decisions based on taste can be categorized as internally driven or conscious decisions. For instance a conscious decision is when consumers decides that the ‘taste’ aspect is more important than the ‘health’ aspect in making food choices (McDaniel and Baker, 1977). This assumes that, to reach those consumers, products which are made healthier should taste the same. In addition, the degree of healthy eating can be influenced by the knowledge of consumers about risks caused by the consumption of unhealthy food. However, for this argument some contradicting findings are shown. More specifically, while some investigators proved that consumers are unaware of the unhealthy impact of some products (Raghunathan, et al., 2006), others state that barriers to eat healthy have nothing to do with consumer’s knowledge about healthy eating (Lappalainen, et al., 1997; Dickson-Spillmann and Siegrist, 2011).

Next, guided by the causes of unhealthy eating, several solutions can be indicated to stimulate healthy eating. Firstly, consumer themselves should change their eating pattern to less unhealthy food which is still tasty (Raghunathan, et al., 2006). However, changing eating patterns asks for a lot of effort, which consumers are not always willing to put in (Lappalainen, et al., 1997). Secondly, changing the assembly of products to a healthier alternative also leads to the consumption of healthier food. Still, for those products it is essential that they taste as the original, because of the major importance of the ‘taste’-aspect when making food decisions. For instance, while considering the distinction between fresh and pasteurized fruit based juices, it is shown that when consumers taste something unfamiliar, this can lead to negative purchase intentions (Sabbe, Verbeke and van Damme, 2009). A reason for those negative purchase intentions is that consumers believe that unfamiliar tasting products are less healthy and less nutritious. Moreover, another way to stimulate healthy eating could be by providing clearer information about ‘health’. This can be fulfilled, for example, by informing the consumer about how to use the food pyramid (Dickson-Spillmann and Siegrist, 2011). That is, the consumer needs to eat more fruit and vegetables and consume a balanced diet.

(13)

13 Another proven driver to eat healthy is being future focused as consumer. More specifically, when consumers experience a positive emotion, like hopefulness, they eat healthier food and prefer less unhealthy food than when consumers are past or present focused, like pride or happiness (Winterich and Haws, 2011). Moreover, a negative future focused emotion leads to the consumption of unhealthy food.

Furthermore, the government and companies are influencers of (healthy) eating behaviour. This impact can be positive when it stimulates healthy eating, as well as negative when it stimulates unhealthy eating. Due to problems caused by unhealthy eating, it is desirable that the government and companies promote healthy eating. As a result, future campaigns of companies and the government should focus on implementing healthy eating in the daily consumption pattern. This implies that consumers need to be stimulated to eat healthy, without a total refusal of their favourite food. One tool to help fulfilling this goal is by placing sufficient role models in specific campaigns. In fact, role models give consumers the opportunity to identify themselves with the behaviour in the campaign (Dalea and Hanburyb, 2010). Hence, in campaigns promoting healthy eating, role models should not been shown as perfectly healthy consumers while acting unhealthy, because when this happens the campaign will have a contradicting effect (Capella, Webster and Kinard, 2011). In addition, implementation intentions can be used by companies to promote healthy eating. Implementation intentions refer to a specific plan about how to act to meet a desired end state (Adriaanse, Vinkers, De Ridder, Hox and De Wit, 2011). In brief, it was shown that implementation intentions have no effect when discouraging the consumption of unhealthy food, while they happen to be effective for promoting healthy eating (Verplanken and Faes, 1999). In fact, when promoting healthy eating, implementation intentions are effective for consumers with and without healthy eating habits.

The findings above may help to indicate important attributes that influence WTP. However, before suitable product attributes are named, the influences on WTP in relation to the research question should be taken into account.

2.2 Willingness-To-Pay (WTP)

(14)

14 Several approaches can be used the measure WTP divided into real WTP versus hypothetical WTP. In fact, real WTP reflects an actual commitment and hypothetical WTP reflects a commitment that does not have financial consequences (Voelckner, 2006). Popular examples of hypothetical WTP are conjoint analysis and contingent valuation. With conjoint analysis consumers are asked to rate or rank products and with contingent valuation consumers are directly asked to state their WTP for a product. Examples of methods for real WTP are auctions, lotteries and preference data, like scanner data. Which method should be used to measure WTP is often investigated in the past. As a result, some guidelines are provided. At first, some reasons are provided why hypothetical methods are preferred. Firstly, it is cheaper to perform a hypothetical method. In addition, the use of real products may be difficult, especially for new product developments. Thirdly, gathering respondents may be much easier for hypothetical method, because the respondents do not need to pay anything when indicating their WTP. However, when choosing hypothetical methods it should be kept in mind that WTP is often overestimated. Furthermore, it should be kept in mind that competition and understanding of the method influences the WTP. Especially WTP measured through auctions is influenced by those factors (Sattler and Völckner, 2002). To conclude, within this research conjoint analysis is used, because of the advantages named for measuring WTP with a hypothetical technique. In addition, conjoint analysis is chosen over contingent valuation, because for respondents contingent valuation is difficult due to the lack of comparison material.

Past literature did not look into the effect of several product attributes on WTP for FMCG that can have a health claim. For this reason, in these review general price implications related to healthy eating, and the influence of organic products on WTP are discussed. Most of all, this section provides insights about potential product attributes, who influence WTP for FMCG that can have a health claim.

(15)

15 Luxembourg. Another pricing issue related to healthy eating is that when a product contains a health claim and is sold at a low price, consumers are sceptical about the reliability of the claim (Aschemann-Witzeland and Hamm, 2010). This could imply that consumers are willing to pay more for products with a health claim.

Furthermore, WTP in relation to organic products has been investigated. First of all, when consumers care about the environment, they are willing to pay more for sustainable products (Royne, Levy and Martinez, 2011). Indeed, in line with this finding, it would be interesting to investigate whether health consciousness of consumers influences WTP for healthy products. In addition, WTP for sustainable products clearly differs per demographical group. For example, African Americans are willing to spend more for sustainable products than Caucasians. Moreover, younger consumers are willing to spend more than older consumers for organic food. As a result, it would be interesting to investigate whether age influences the effect of product attributes on WTP for FMCG that can have a health claim and whether segments could be abstracted from the market.

Furthermore, quality and health functions influence WTP for organic products. Before considering those influences, a distinction can be made between products, namely ‘wants’ versus ‘shoulds’ (Milkman, Rogers and Bazerman, 2009). In short, ‘wants’ refers to products which only lead to a short-term goal fulfilment and are not favourable for the long term. On the other hand, ‘shoulds’ refers to products that fulfil long-term goals. Nevertheless, those products are less appealing for the short term. In conclusion, quality positively affects WTP for organic products, where the effect is stronger for ‘shoulds’ than for ‘wants’ products (van Doorn and Verhoef, 2011). In addition, products with an organic claim belonging to the ‘wants’ category are interpreted as products with lower quality, but healthier. However, health perceptions do not affect WTP for organic products. Nevertheless, in other findings health is seen as the main purchase motivation for organic food (Cicia, Del Giudice and Ramunno, 2009). An explanation for those contradicting findings can be because influencers determining WTP are strongly dependent on the product category.

2.3 Product Attributes

(16)

16 a product and ‘how the calorie-information is made generally known’, which can be communicated by a health claim. As a result, a health claim or logo is assumed to have an effect on WTP for FMCG that can have a health claim. Moreover, it is predicted that a healthier assembly of a product can influence WTP.

In addition, it is expected that three other variables influence WTP. To start with the brand, this is expected to influence the WTP. That is, companies can influence eating patterns of consumers by using their brand as important tool. Next, the material of the packaging is included in this investigation. This attribute is included because under time pressure visual elements of products are of major importance for the decision whether to buy or not to buy a product (Silayoi and Speece, 2004). Because time pressure is common for the Western world, this attribute is assumed to have an influence on WTP in this research. Lastly, it is expected that taste influences the WTP, because of the importance of taste when choosing healthier product assemblies.

To conclude, product attributes investigated in the current study are health claims, brand, packaging materials, ingredients of product and product taste. Below each of those product attributes are discussed in depth.

2.3.1 Health Claim

A health claim denotes a claim on a product which shows that this product is perceived as more healthy than similar products without that claim (Aschemann-Witzeland and Hamm, 2010). In the Netherlands, an example of such a claim is the Healthy Choice logo. This logo shows that a product meets a set standard for added sugars, satiated fat, trans fat, fibres, energy and sodium and that this product is the best choice in a particular product group (Vyth, Steenhuis, Heymans, Roodenburg, Brug, and Seidell, 2011). And even though it has been investigated that health perceptions do not affect WTP for organic products (van Doorn and Verhoef, 2011), other findings show that health is seen as the main purchase motivation for organic food (Cicia, Del Giudice and Ramunno, 2009). That is why the health claim is considered as relevant for this research as well.

(17)

17 reasoning that packaged products are often considered as high involvement products. Consequently, for FMCG that can have a health claim it is assumed that the product information is considered. Moreover, the health trend increases the attention for packaging information. Namely, by using product information, consumers can make well informed decisions which meet their health goals. However, this information should not be conflicting. In fact, when consumers receive conflicting information or no information at all about healthy eating, they are not willing to change their eating pattern (Lappalainen, et al., 1997). As a result, it seems to be beneficial to mention the Healthy Choice logo on the packaging, due to the clear message communicated by the logo. In addition, product labels should be understandable, otherwise consumers still do not know which products are healthy and which nutritional value the food has (Dalea and Hanburyb, 2010). Nevertheless, the current label information is often not simplified enough (Silayoi and Speece, 2004). Especially in situations where time pressure occurs, simplified product information is required. In fact, a simple message communicating the health of a product increases the sales for healthy products (Thorndike, Sonnenberg, Riis, Barraclough and Levy, 2012). Therefore the Healthy Choice logo is a solution, because the logo shows simplified information with a clear message about the healthiness of a product.

Based on previous investigations several conclusions can be identified concerning health claims. Overall, products with a health claim are preferred. Indeed, a rise in sales volume is shown for products containing a health claim, and brands using a low-fat claim are more successful than their high-fat competitors (Krystallis and Chrysochou, 2011). In line with this finding, consumers seem to be more loyal to products containing a low fat claim. In addition, a health claim is most effective when shortly shown on the front of the packaging and explained in depth on the back of the packaging (Wansink, 2003). Note that those conclusions can differ per food category. For instance, for hedonic products, which are products consumed for sensory pleasure, a health claim is shown to be more effective than for functional products (Kim, Cheong, and Zheng, 2009). A functional product is a product consumed for healthy or convenience reasons (Walker Naylor, Droms, and Haws, 2009). In contrary, a taste claim positively influences purchase intentions for functional food products. In fact, FMCG that can have a health claim can be implied as hedonic as well as functional. To be more specific, they are suitable for sensory pleasure when consuming it for pleasure, and as a functional product when it is consumed for health reasons.

(18)

18 organic products are more expensive than ‘other’ food products. Additionally, the effect of the claim depends on words used to inform a consumer about an organic claim. For example, when a packaging state ‘without chemical additives’ the impact of the claim on purchase intentions is higher, than when the claim state ‘organic’ (Abadio Finco, et al., 2010). This might suggest that not all consumers are aware of the meaning of the word ‘organic’.

Furthermore, the influence of a Healthy Choice logo has been investigated. More specifically, the logo has been examined by looking into the effect of a health claim for fast-food restaurants (Vyth, et al., 2011). In brief, consumers increasingly choose high calorie side dishes when the main dish contains a health claim (Chandon and Wansink, 2007). However, for a fast food restaurant, the positive effect of a health claim can simply be levelled out by asking consumers whether the health claim is reliable or not. In conclusion, putting the logo on menus of cafeterias has only an effect for health conscious consumers or consumers who really pay attention to the information on products (Vyth, et al., 2011). An explanation why health unconscious consumers do not pay attention to the logo can be, that they do not know what the logo means. Hence, before consumers can use the logo actively, they should be educated about the usefulness of healthy eating.

Based on the gathered information above, the following hypothesis is formulated:

Hypothesis 1: A health claim positively influences Willingness-To-Pay for FMCG that can have a health

claim.

2.3.2 Brand

(19)

19 consumers, which can be expressed through using particular brands. On the other hand, the brand also influences the product involvement. In fact, brands may steadily act to awaken involvement of customers (Silayoi and Speece, 2004). In greater detail, specific dimensions can be named which stimulate brand trust and brand affect (Sung and Kim, 2010), where ‘brand affect’ refers to: ‘the positive effect of using the product for the average consumer’. In conclusion, brand trust is strongly influenced by sincerity (honestly) and ruggedness (macho) and brand affect is more related to excitement and refinement.

Furthermore, to claim a healthy brand image the product itself is most important (Chrysochou, 2010). In addition, advertising and packaging can create a link between the product and health. However, the physical product is the base for a healthy image of a brand. To be more specific, a healthy slogan only leads to a more positive impression of a product when it promotes a healthy product (Adams, and Geuens, 2007). Consequently, an unhealthy product can better be promoted by an unhealthy slogan.

In addition, especially brand familiarity seems to be important for an increased WTP. Brand familiarity occurs when buyers already have purchase-experiences with a brand (Monroe, 1976). For example for coffee, it has been proven that brand is the most important attribute determining purchase intentions (Pelsmacker, Driesen and Rayp, 2005). In fact, for a familiar brand consumer are willing to pay more. A similar pattern has been shown for milk products, wherefore brand is the most important external cue when choosing a product (Schnettler, Viñuela and Sepúlveda, 2008). More specifically, consumers are willing to pay less for private labels and more for national brands. Furthermore for steaks the WTP has been proven to be the highest for branded steaks (Froehlich, Carlberg and Ward, 2009).

Guided by this information and the information above the following hypothesis is formed:

Hypothesis 2: Brand familiarity positively influences Willingness-To-Pay for FMCG that can have a

health claim compared to less familiar brands or private labels.

2.3.3 Packaging Material

In this research ‘packaging material’ denotes where the packaging is made from.

(20)

20 consumers’ intention to purchase the product (Abadio Finco, et al., 2010). Furthermore, packaging is a very important tool for communicating product elements (Silayoi and Speece, 2004). In particular, in the purchase state the packaging and information on the packaging seem to be main influencers of consumer buying. Additionally, packaging affects judgments of consumers about the volume the product contains (Silayoi and Speece, 2004). Interestingly, even when consumers encounter less volume than they expected upfront, they still buy the same packaging again. Furthermore, the quantity a consumer consumes of a product is influenced by the shape of the packaging (Wansink and Huckabee, 2005). Besides, the material of a package can make the product more/less appealing. Moreover, for consumers convenience and ease to carry the product are very important elements of the packaging. Besides, in the post-purchase state consumers also evaluate the product, namely by checking whether the packaging keeps the product in the desired condition (Estiri, et al., 2010).

Moreover the assumed quality of a product is influenced by the packaging. Namely, in supermarkets consumers can not base their quality decisions on intrinsic attributes as taste, but have to make them on extrinsic attributes like price. In fact, the quality of a product is not only affected by the price-level (Gerstner, 1985), but also by other marketing-mix variables (Kirmani and Rao, 2000). Namely, as quality signal it has been proven that brand and the packaging are equally or even more important than price for packed goods (Zeithaml, 1988).

In this research is assumed that for packed food the material of the packaging guides the form of the packaging. For example, for chilled orange juice, packaging materials which are mainly used are bottles of glass, bottles of plastic and carton packs. Consequently, no other packaging material-form combinations are taken into account in this research, due to a decline in WTP for non-familiar packagings. In short, consumers are willing to pay more for a packaging which is familiar in form (van Rompay and Pruyn, 2011). Namely, the form of a product influences perceptions created by a product. For example, a study proved that the form of products (shoes and speakers) creates different perceptions about functional performance and features about that products (Hoegg and Alba, 2011). Thus, the functionality of a product can be communicated through design efforts.

However, past literature looked not into the influences of different materials on the WTP for FMCG that can have a health claim. Consequently, the following hypothesis is phrased:

Hypothesis 3: The material of the packaging influences Willingness-To-Pay for FMCG that can have a

(21)

21 2.3.4 Product Ingredients

Product ingredients denote the assembly of the product. To be more specific, a product can be produced with its original ingredients or with healthier and/or more natural ingredients. To meet the current health trend, opportunities lie into producing original products with healthier and more natural ingredients instead of the current ingredients. A successful example is less fat containing milk or yoghurt. Those successfully launched products are produced with a healthier composition of ingredients without added vitamins (Bower, et al., 2003). However, most of the time those new produced products are more expensive due to adjusted ingredients. Thus, it is very interesting to examine WTP for products with healthier and more natural ingredients in FMCG that can have a health claim. In general, low-fat milk and yoghurt products perform equally or even better than their full-fat competitors and have a higher market share than full-fat products (Krystallis and Chrysochou, 2011). In addition, consumers are more loyal to low-fat products. Hence, this makes it even more interesting to take WTP into account for FMCG that can have a health claim produced with healthier and more natural ingredients.

In addition, investigations show that purchase intentions rise for products which contain ingredients with a positive effect on the health of consumers (Bower, et al., 2003). Still, the elements ‘health’ and ‘price’ both influence the intention to buy or not buy products with positive health effects. Nevertheless, it was proven that consumers WTP increases for products with a healthier composition. In addition, positive sensory aspects increase the intention to buy, even when a high price was asked in combination with the health advantage. Additionally, the communication about product ingredients seems to be of major importance (Abadio Finco, et al., 2010). Hence, in order to increase WTP, clear information about ingredients processed in a product should be provided.

Based on the information above, the following hypothesis is created:

Hypothesis 4: Products with healthier ingredients positively influence Willingness-To-Pay for FMCG

that can have a health claim compared to products with its original ingredients.

2.3.5 Product taste

Taste reflects the flavour of the product.

(22)

22 affected by taste (Sabbe, et al., 2009). Especially for functional products the communication about a good taste is very important (Kim, et al., 2009). In addition, when the product was not as tasty as expected, the product was interpreted as less healthy. As a result, taste is a precondition for the success of healthy food (Baker, Thompson, Engelken and Huntley, 2004).

Furthermore, snacks which are healthier than the original product have been investigated. In brief, the product taste is inferred to be better when a familiar brand positions a product as a healthier alternative, then when an unfamiliar brand positions a product as a healthier alternative (Desai and Ratneshwar, 2003). More specifically, this has to do with associations linked to a particular brand. Namely, consumers assume new products of a particular brand to meet associations they already built up about a brand. Those associations could be about the shape and size of a product as well as about ingredients and taste. As a result, those associations can guide a negative response to atypical and new product characteristics. For instance, low fat snack products compared to the full fat reference product. However, familiar brands can afford to change attributes, when the change is not too extreme.

Based on the information above a hypothesis is formulated. Within this hypothesis the term ‘natural taste’ is indicated. Natural taste reflects a various taste depending on the harvest. In other words, the taste will not be corrected to meet the general known taste in the market. This means that a product can taste slightly different from the same product bought a week later from the same brand (more sour/sweet).

Hypothesis 5: Products with the original taste positively influence Willingness-To-Pay for FMCG that

can have a health claim compared to products with a natural taste.

2.4 Moderators

(23)

23 2.4.1 Age

Age denotes the age of consumers. It has been shown that age has no effect on the relation between products with a health claim and purchase intentions (Aschemann-Witzeland and Hamm, 2010). Nevertheless, other articles infer that age has an effect on the influence of a health claim on WTP as well as on other relevant relationships.

At first, involvement is moderated by age. Namely, for younger consumers their lack of time is indicated frequently as a barrier to eat healthy (Lappalainen, et al., 1997). As mentioned before, lack of time results in less involvement while purchasing products, meaning that younger consumers increasingly need simplified product information, which could be, for example, a Healthy Choice logo. However, even though younger consumers encounter less time, older consumers indicated to be in stronger need for the Healthy Choice logo (Vyth, Steenhuis, Mallant, Mol, Temminghoff, Feunekes, Jansen, Verhagen and Seidell, 2009). This could be explained by the health concerns older consumers encounter.

Moreover, differences in importance levels of different attributes in purchase-stages per age group exist. In fact, a large difference exists between the age groups below 20 and above 40 years old. Indeed, for below 20 the colour and shape are of major importance in the pre-purchase stage, while for above 40 the shape and information is more important (Estiri, et al., 2010). Furthermore, during the actual purchase-stage information is important for the group consumers above the age of 40, while for under the age of 20 colour, shape, and size of the packaging are denoted as most important. Only the post-purchase stage both age groups are highly comparable, because both groups indicate size of the packaging as important. Nevertheless, consumers under the age of 20 also think shape is important in the post-purchase stage.

Additionally, age has an effect on the relation between sustainable products and WTP (van Doorn and Verhoef, 2011; Royne, et al., 2011). To be more specific, the younger the consumer, the more willing to pay higher amounts of money for sustainable food. This can be explained by the lack of information about sustainable food older consumers encounter (Dickson-Spillmann and Siegrist, 2011). On the other hand, older consumers have a higher WTP and a higher intention to purchase products with a health benefit compared to the regular product (Bower, et al., 2003).

(24)

24 consumers are familiar with specific tastes for a longer time than younger consumers, who could easier adapt their taste.

Besides, past investigations did not take into account the influence of age on the relationship between brand and WTP. Nevertheless, in all age categories some consumers switch brands a lot, while others stay loyal to one brand for a long time period (Lambert-Pandraud and Laurent, 2010). Still, overall, older consumers are more brand loyal than younger consumers. However, it is not clear whether older consumers stay loyal to a familiar brand or to an unfamiliar brand. For this reason, it is assumed that age influences the relationship between brand and WTP, while it is unclear whether this effect is positive or negative.

Accordingly, the following hypotheses are set up:

Hypothesis 1a: The positive influence of a health claim on Willingness-To-Pay for FMCG that can have

a health claim is stronger for older than for younger consumers.

Hypothesis 2a: The positive influence of brand familiarity on Willingness-To-Pay for FMCG that can

have a health claim versus less familiar brands and private label is moderated by age.

Hypothesis 3a: The influence of the material of the packaging on Willingness-To-Pay for FMCG that

can have a health claim is stronger for younger than for older consumers.

Hypothesis 4a: The positive influence of products with healthier ingredients on Willingness-To-Pay

FMCG that can have a health claim compared to products with its original ingredients is stronger for older than for younger consumers.

Hypothesis 5a: The positive influence of products with the original taste on Willingness-To-Pay FMCG that can have a health claim compared to products with its a natural taste is stronger for older than for younger consumers.

2.4.2 Gender

Gender refers to the sex of consumers.

(25)

25 WTP and intention to purchase products with a health benefit ingredient compared to the regular product (Bower, et al., 2003). However, when the price is too high, health benefits are outweighed, and not a major stimulant anymore to buy the product. Furthermore, men are less likely to change their eating patterns in a healthy eating pattern than women are (Lappalainen, et al., 1997). Namely, men see irregular working hours as a barrier to eat healthy, while this is not a mentioned barrier for women. In line with this, women perceive a Healthy Choice logo as more credible than men and buy more products with a Healthy Choice logo (Vyth, et al., 2009).

Moreover, purchase-stages are affected by gender. More specifically, where females consider colour and shape in the pre-purchase stage, males consider information and shape in this stage (Estiri, et al., 2010). Additionally, during the purchase stage a difference exists. That is, females focus in this stage on size and shape, where males focus on colour and information. In the end, also in the post-purchase stage, women care about information and size and men about colour and shape.

In addition, the influence of gender on the relationship between brand and WTP and taste and WTP has not been examined before. Nevertheless, differences between men and women concerning brands have been investigated before. For instance, women are known as more sensitive to brand images and personality dimensions linked to a brand than men are (Lau and Phau, 2010). Also, women are more likely to be brand lovers as men are (Pelsmacker, et al., 2005). Also taste differences occur between the genders, namely women are more taste-sensitive than men are (Simpson, et al., 2012).

As a result, the following hypotheses are formulated:

Hypothesis 1b: The positive influence of a health claim on Willingness-To-Pay for FMCG that can have

a health claim is stronger for women than for men.

Hypothesis 2b: The positive influence of brand familiarity on Willingness-To-Pay for FMCG that can

have a health claim versus less familiar brands and private label is stronger for women than for men.

Hypothesis 3b: The influence of the material of the packaging on Willingness-To-Pay for FMCG that

can have a health claim is stronger for women than for men.

Hypothesis 4b: The positive influence of products with healthier ingredients on Willingness-To-Pay

(26)

26

Hypothesis 5b: The positive influence of products with the original taste on Willingness-To-Pay for

FMCG that can have a health claim compared to products with a natural taste is stronger for women than for men.

2.4.3 Health Consciousness

Health Consciousness expresses the degree of awareness consumers experience concerning healthy eating. In short, for organic products was shown that the effect on WTP is moderated by health consciousness (Cicia, et al., 2009). As a result, it is expected that this moderator influences the effect of several product attributes on WTP for FMCG that can have a health claim.

First of all, the effect of health consciousness of consumers on the health claim is described. Indeed, products with a health claim are chosen more often when a consumer is health conscious (Walker Naylor, Droms, and Haws, 2009; Vyth, et al., 2009; Aschemann-Witzeland and Hamm, 2010). That is why it is expected that WTP for products with or without a Healthy Choice logo is also moderated by health consciousness. For example, it turns out that when health conscious consumers have the opportunity to choose their food in a cafeteria based on a choice logo, they use this opportunity (Vyth, et al., 2011). However, some studies detect no differences between health conscious versus unconscious consumers concerning health claims (Chandon and Wansink, 2007). For instance, when a product contains a health claim in a restaurant, calorie estimations are equal for health conscious and less health conscious consumers.

(27)

27 consumers see those products as more tasty than health unconscious consumers (Bower, et al., 2003; Desai and Ratneshwar, 2003).

Unfortunately, the effect of health (un)consciousness of consumers on the effect of brand and packaging material on WTP is not investigated before. It is assumed that the health consciousness of consumers has a positive effect when the brand is positioned as healthy, so a familiar brand. This assumption is made because consumers are also affected by ‘unhealthy’ product names (Irmak, et al., 2011). Furthermore, it is also assumed that health consciousness has an effect on the relation between packaging material and WTP. For this reason, in general, a product packed in a bottle of glass is perceived as more natural and healthier than products packed in a carton box. Consequently, consumers who are health conscious probably prefer a bottle of glass, based on the perception that the product in the bottle is healthier and more natural.

Thus, the following hypotheses are composed:

Hypothesis 1c: The positive influence of a health claim on Willingness-To-Pay for FMCG that can have

a health claim is stronger for health conscious than for health unconscious consumers.

Hypothesis 2c: The positive influence of brand familiarity on Willingness-To-Pay for FMCG that can

have a health claim versus less familiar brands and private label is moderated by health consciousness of the consumer.

Hypothesis 3c: The influence of the material of the packaging on Willingness-To-Pay for FMCG that

can have a health claim is moderated by the health consciousness of the consumer.

Hypothesis 4c: The positive influence of products with healthier ingredients on Willingness-To-Pay for

FMCG that can have a health claim compared to products with its original ingredients is stronger for health conscious than for health unconscious consumers.

Hypothesis 5c: The positive influence of products the original taste on Willingness-To-Pay for FMCG

that can have a health claim compared to products with a natural taste is stronger for health unconscious than for health conscious consumers.

2.5 Conceptual Model

(28)

28

Figure 1: Conceptual Model

Figure 1 reflects the conceptual model containing relationships described in the literature review. In the figure also the three moderating effects are included.

Based on the described literature and the formulated hypotheses, in the following chapter the research design is described.

H1c-H5c Health (un)consciousness Product Attributes H1 Health Claim H2 Brand H3 Packaging Material H4 Product Ingredients H5 Product Taste Willingness-To-Pay (WTP)

(29)

29 3. RESEARCH DESIGN

In this chapter the research design is discussed. Firstly, the chosen research method is explained. Secondly, the way of data collection is elaborated. Lastly, the measurement development is described.

3.1 Research Method

(30)

30 sometimes consumers can eliminate a product based on one particular attribute (non-compensatory). This can lead to unrealistic utility scores for other attributes. Lastly, in general, conjoint studies provide insights about consumer preferences and not about actual buying behaviour.

3.2 Data Collection

The hypotheses are tested in the Dutch market. An online survey was conducted with more than 200 respondents. Generally, the amount of respondents needed for conjoint analysis varies from 150 to 1200 respondents, depending on the method and purpose of the study (Orme, 2010). However, also an amount of 50 respondents could be enough to provide insights in preferences and variation in respondents. Still, to have an acceptable margin of error, 200 respondents is an acceptable sample size (Hair, Black, Babin and Anderson, 2010). Because of monetary and time restrictions of this study, 200 respondents is inferred to be an acceptable amount.

The choice for a survey is based on the possibility to reach many respondents in a short time period. The survey was made available online by using the website ‘www.thesistools.nl’. The choice for online instead of offline is, because of the high response rate for online surveys. Furthermore, online surveys are cheaper and it is easier to send a reminder. Some researchers state that offline surveys are more reliable. However, recent studies show that the quality of respondents and the amount of finished questionnaires is similar compared to offline surveys (Deutskens, de Ruyter and Wetzels, 2006). Awareness for this survey is gained through email messages, and by generating snowball sampling. The survey is send by email to friends, family and fellow students of the University of Groningen. In addition, the survey is made generally known through social network sites like www.facebook.com, www.twitter.com and www.hyves.nl. The reason for creating snowball sampling is, that this results in many respondents in a short time period. The survey has been conducted in November and December 2011.

(31)

31 3.3 Measurement Development

The attributes and levels included in the questionnaire are based on past investigations as shown in the literature review and are evaluated by an expert, namely the brand manager of a leading brand in orange juice in the Netherlands, called ‘Appelsientje’.

To investigate the influence of the product attributes on WTP, it is also necessary to include a price attribute.The chosen pricing levels are based on average prices asked for chilled orange juices in the largest Dutch supermarket, namely Albert Heijn (November 2011). While looking into prices for brands packed in carton packs, the price for CoolBest is €1,30, for Healthy People €1,28 and for private label the price is €1,08. Based on those prices the average price per litre is € 1,22 for chilled orange juice in carton packs. However, chilled orange juice in plastic bottles varies much more in price. For example, the C-brand ‘Euro shopper’ is sold for €0,95 per litre and the A-brand Hero costs €1,84 per litre. Because effects of packaging material and brand on WTP are taken into account, varying prices are taken into account to choose most suitable price levels. As a result, levels which are considered are €1,22 and this price at -15% and +15%. Consequently, the prices included in the study are €1,04, €1,22 and €1,40. It is chosen to have three instead of more price levels. The reason for this decision is, because it is better to have more measures per price point, than to have many levels with just a few measurements per price point (Johnson, and Orme, 2003).

As a result, five product attributes are included in the analysis, which is a preferable amount based on the software used (Sawtooth ISS Web). Note that product ingredients and product taste are measured as one variable, to ease the evaluation task for respondents. In fact, more than six attributes would result in difficulties for respondents to analyse profiles (Sawtooth Software, 2008). For each choice set the full profile method is used. This means that all product concepts are described with one level of each attribute. Compared to the trade-off method, where only two attributes are depicted a time, advantages of the full profile method are that fewer judgments per respondent need to be gathered, it is more realistic, and various presentation formats are possible. However, disadvantages are that it is difficult for respondents to evaluate this amount of information per choice task, and the order representation of attributes may affect the importance.

(32)

32

Table 1: Attributes and Levels

Attribute Attribute level Model specification

Health claim - Without Healthy Choice logo (base) Partworth model - With Healthy Choice logo

Brand - CoolBest (base) Partworth model

- Healthy People - Private Label

Packaging material - Carton pack (base) Partworth model - Bottle of plastic

- Bottle of glass Product ingredients

and product taste

- Original product ingredients (base) Partworth model - Healthier ingredients and original taste

- Healthier ingredients and natural taste

Price per litre - € 1.04 (base) Linear vector or Partworth model - € 1.22

- € 1.40

Each of these levels is mutually exclusive. In addition, because the attributes all contain about an equal amount of levels, it is expected that the number-of-levels effect does not influence the results of the gathered data. The number-of-levels effect means, that an attribute with a high amount of levels is interpreted as more important than an attribute which only contains a few levels.

Based on the information above the utility function is defined as follows:

U = (health claim, brand, packaging material, product ingredients, price) = constant + HealthyChoiceLogo

 xHealthyChoiceLogo + HealthyPeople  xHealthyPeople + PrivateLabel  xPrivateLabel + PlasticBottle  xPlacticBottle + GlassBottle 

xGlassBottle + HealthierCurrentTaste  xHealthierCurrentTaste + HealthierNaturalTaste  xHealthierNaturalTaste + Price  Price

For all x applies x = 1 means the product contains this level, and x = 0 means the product does not contain this level.

(33)

33 reliable results, manageable in a short time period. The amount of time to complete a survey containing 20 choice sets is about 5 minutes. This is assumed to be acceptable by respondents. Meanwhile, each choice set can contain about three to five product concepts. In the current study each given task contains two choice sets and a none-option (Miller, Hofstetter, Krohmer and Zhang, 2011). This amount is chosen because it is easier for respondents to choose between two than three or more products. To find out which choice sets should be asked to have a reliable reflection for the entire amount of choice sets, ‘Sawtooth SSI Web’ is used. This program helps to generate a questionnaire with 20 choice sets. For the 20 choice sets which were generated, the level balance for all attributes is sufficient in the design. Furthermore, the relative efficiency of the design is calculated by the efficiency scores. By looking at the efficiency scores can be abstracted that those are all close to 1.0, which is the best score. In conclusion, the design is efficient concerning the main effects.

Based on the information above the questionnaire is designed. The questionnaire starts with four statements, which are used to abstract the health consciousness of the respondent. Those statements are based on previous findings about how to examine the health consciousness of consumers (van Doorn and Verhoef, 2011). For determining the health consciousness the 7-point Likert-type scale is used (Nysveen, Pedersen and Thorbjørnsen, 2005). Secondly, twenty choice sets were asked to the respondents, with hold-out sets asked as choice set 5, 10 and 15. In fact, it is important to spread hold-out sets over the questionnaire, because respondents tend to learn about their preferences during the fulfilment of the questionnaire (Sawtooth Software, 2008). Besides, respondents were invited to choose one profile a time or to choose for the none-option. In fact, they were asked to indicate their first choice only for each choice set, because when asking for the second choice as well, those results are biased (Johnson and Orme, 1996). Besides, the choice sets were represented as verbal cue cards. Those are quicker to read than paragraph descriptions, where real sentences are shown. Multimedia representations are not included in the questionnaire, because all attributes and levels can be clearly explained in words and it is assumed that multimedia for some attributes may result in a favour for those attributes, due to their more attractive appearance (Sawtooth Software, 2008). However, to clarify the attributes and their levels, in the introduction images of all attribute levels are shown. After the choice sets, respondents were asked to indicate their age (Dutta-Bergman, 2004) and gender. Lastly, respondents have the opportunity to enter additional notes. In appendix 1 the questionnaire is included.

(34)

34 4. RESULTS

In this chapter the results of this research are discussed. Firstly, extreme response behaviour is discussed. Next, respondents are described based on their demographics, where a factor analysis is conducted, resulting in one variable for health consciousness. Next, the CBC analysis is conducted at two levels, namely aggregate and segment level. After that, based on those results the predictive validity is calculated followed by the equalization price and WTP. During this chapter, the formulated hypotheses are judged.

In total 267 respondents answered the questionnaire. However, after data cleaning respondents who only answered a few questions are excluded from further analysis. Nevertheless, respondents who randomly forgot to answer one or two questions, are used for further analysis. The unanswered questions are indicated as missing values. Based on this screening, the sample used consists of 209 respondents.

4.1 Demographics

Before demographics of the sample are discussed, first possible extreme response behaviour is taken into account.

4.1.1 Extreme Response Behaviour

To detect extreme response behaviour, a table is provided in appendix 2. This table shows how often the none-option is chosen. When a consumer (almost) always or never indicated the none-option while answering the questionnaire, extreme response behaviour is detected.

The table shows that no one chooses the none-option all the time. Furthermore, it is interesting that almost 60 percent of respondents never choose the none-option. Hence, respondents who answered the questionnaire buy chilled orange juice from concentrate and are willing to pay at least €1.04. However, this could also be interpreted as the offered pricing levels where too low. More specifically, consumers are probably willing to pay more for chilled orange juice. On the other hand, an advantage of a high number of chosen stimulus is that more information per parameter is available, than when the none-option was chosen all the time.

(35)

35 effect in the CBC analysis is measured as well as their arguments for their behaviour. Those additional arguments are incidentally mentioned in the box where additional notes could be provided in the questionnaire. In short, one respondent indicates that she prefers fresh pressed juice (the respondent who chooses the none-option 16 times). Therefore, this outlier is representative for the population (Hair et al. 2010). Namely, the respondent consumes chilled orange juice, only incidentally, because she prefers fresh pressed juice. The other outlier did not specify his underlying reasoning to choose the none-option frequently. Though, this specific respondent indicated a choice situated in the middle of the questionnaire and not for the first question. This could reflect a well considered choice. Thus, to ensure generalizability for the entire population, outliers are chosen to be included in the analysis. This decision is confirmed after testing the sample without outliers compared to the sample with outliers in the CBC-analysis, because differences in outcomes are minimal.

4.1.2 Descriptive Statistics

The sample (n = 209) consists for 39.7 percent out of men and 60.3 percent out of women. Thus the moderator gender is well balanced. Secondly, the age of the respondents is described (n = 207). Note that two respondents did not indicate their age. Figure 2 below shows the age division of the sample.

Figure 2: Age division of the sample (n = 207)

From figure 2 can be abstracted that the largest group of respondents belongs to the group in the age of 21 to 25 years old. This could be explained by the sampling method chosen, namely approaching friends of the researcher and creating a snowball effect with those friends as base. Thus, this could result in an age bias.

(36)

36 The last descriptive variable measured is the health consciousness of respondents. This variable is measured by asking respondents to agree/disagree on four statements on a 7-point Likert scale. The statements are: ‘Despite the risk of health problems, I do not adapt my consumption pattern’, ‘I would rather enjoy live than to worry about healthy problems’, ‘I try to prevent health problems before I experience any symptoms related to health problems.’ and ‘I do not think health problems will happen to me’. To create a more manageable number of variables out of those four statements, factor analysis is used. With this data reduction technique highly correlated variables are replaced by a smaller number of uncorrelated variables, representing health consciousness. Moreover, factor analysis is able to detect uncorrelated variables which can be excluded from the dataset. Before measuring the internal consistency between statements, one statement is recoded. More specifically, for three statements health consciousness was met when respondents indicated disagreement with the statement, and health unconsciousness when they agreed with the statement. Despite for one statement, namely ‘I try to prevent health problems before I experience any symptoms related to health problems’. To prevent misinterpretations of this statement while comparing the statement to the other statements, this statement is recoded. As a consequence, for all statements, 1 reflects health consciousness and 7 health unconsciousness.

4.1.2.1 Factor Analysis

Before the factor analysis is conducted, the internal consistency between the four statements is measured to see how closely related variables are as a group. As a result, the Cronbach’s alpha score is 0.56, which is below the acceptable standard of 0.6 (Hair et al, 2010).

(37)

37 this statement. This means that this statement is not represented well within the one component solution.

As a result, the measurement of the moderator ‘health consciousness’ can be improved by deleting ‘I do not think health problems will happen to me’. This leads to a valid interpretation of the variable ‘health consciousness’ after creating one component out of all statements. This is also confirmed by calculating the Cronbach’s Alpha when using the three remaining statements (0.617) compared to the four statement solution (0.566). Interestingly, this outlying statement is not the recoded statement. Hence, respondents actively participated in answering statements correctly to show their interests. A reason why the deleted statement differs from the other three statements could be, because the other three statements ask how consumers behave conserving their health, and the deleted statement reflects the respondent’s perceived confidence in his/her health for the future. Appendix 3 provides an overview of the data described above.

Consequently, it is sufficient to make one variable out of the three statements. Due to the interpretability of the statements is chosen to compute a new variable, instead of using the designed factor for further analyse purposes. The new variable is called ‘health consciousness’ and reflects an average of the responses on the three statements. More specifically, 1 reflects a very health conscious respondent and 7 a very health unconscious respondent. With the new created health conscious variable, the health consciousness of the respondents can be described. Figure 3 represents the degree of health consciousness of the respondents.

Figure 3: Degree of Health Consciousness (n = 208)

Based on figure 3 can be concluded that the respondents are more health conscious as health unconscious. More specifically, the mean is 3.56. In fact, only one of the respondents indicates him or herself as very health unconscious (0.5%).

0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 Percentage

Very Health Conscious Health Conscious Partly Health Conscious Neutral

Referenties

GERELATEERDE DOCUMENTEN

The three primary research objectives are: to appraise evidence-based clinical guidelines referring to ankle sprains (published within the past five years); to determine

Specifically, the aim of this study was to expand the literature on the topic of trivial product attributes, by investigating consumers’ willingness to pay, including the

Wessels het die Volksraad ten einde laaste besluit om Van Hoytema as sekretaris te ontslaan, terw yl die Raad hom nie kon verenig met die beskouing van die

Since CSR and sustainability are integrated into the core business strategy of the organization, subsidiaries are allowed and do initiate CSR related activities or

4b A robot with facial expressions and body movement strengthens the influence of active social interaction on purchase intention, compared to a robot that looks like a machine

In the analysis of inventory control it became clear that the relatively high inventory level, the lacking of an inventory control system and clear inventory rules and the lacking

El- ders langs onze kust zijn vooral de (oude) platte oesters schelpen (Ostrea edulis) vaak aangeboord door deze boor- spons.. Hoeksema (1983) heeft uitgebreid de boorspons Cli-