UNIVERSITY OF AARHUS AARHUS SCHOOL OF BUSINESS Department of Marketing and Statistics

Masters thesis

EFFECTS OF PACKAGE DESIGN ON CONSUMER EXPECTATIONS OF FOOD PRODUCT HEALTHINESS

Author: Thomas Sioutis Advisor: Joachim Scholderer

July 2011

ACKNOWLEDGEMENTS I would like to thank my supervisor Prof. Joachim Scholderer for his patience with me and his great support.

Also I would like to thank all the participants to my survey. All these people spent some time to participate to my experiment and fulfill the questionnaire, without their help it would be impossible for me to conduct this survey.

Special thanks also to my cousin Fotini Siouti for her psychological support all the last months.

Finally I would like to dedicate this work to my parents, to thank them for their lifelong support and tell them how much I love them.

EXECUTIVE SUMMARY Recently there is a trend for a healthier diet and healthy food consumption. This trend has generated an intensive competition on healthy food products within the food market. This fact along with the tough regulation rules about food health claims has made food marketers to seek new ways to communicate the health message of their product. A vehicle of communication in marketing is the product package itself. The visual non verbal attributes of a food package are of low even no regulations and offer a field of unlimited creativity. This study aims to examine if different nonverbal attributes of food and beverage packages affects consumers expectations of food and beverages product healthiness. Theory suggests that healthiness is a food quality dimension and that package can act as an extrinsic quality cue. This study based on the Total Food Quality Model that links the intrinsic and extrinsic quality cues with quality expectations and experience. Limited research has been conducted that links package design and food healthiness expectations. But research about the general preferences of consumers suggests as important food package design attributes colour, shape, graphics, branding and size. For this study, an experiment was conducted. 73 consumers participated in it. The stimuli of it were real package prototypes of cereals and fruit juice. The attributes that was examined were: colour (red-green), shape (square-curvy), graphics (image of the product-landscape) and the visibility of the product (existence of transparent part-not). Results indicate that the most important attributes are the shape and the visibility. The preferences for the shape are strongly product oriented. Also consumers prefer packages that offer visibility to the product. The other two attributes are of low importance, with the colour to be the last one. As it seems there is no significant difference between product involved consumers and uninvolved ones in terms of package design preferences for the healthiness expectations that it generates.

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ............................................................................................................................. 2 EXECUTIVE SUMMARY .............................................................................................................................. 3 TABLE OF CONTENTS ................................................................................................................................. 4 1. INTRODUCTION ........................................................................................................................................ 5 1.1 BACKGROUND .......................................................................................................................................... 5 1.2 OBJECTIVES .............................................................................................................................................. 8 1.3 LIMITATIONS ............................................................................................................................................ 8 1.4 OVERVIEW ................................................................................................................................................ 8 2. THEORETICAL FRAMEWORK .............................................................................................................. 9 2.1 HEALTH AS FOOD QUALITY DIMENSION ................................................................................................. 9 2.2 THE TOTAL FOOD QUALITY MODEL ....................................................................................................... 10 2.3 THE ROLE OF FOOD PACKAGE AS EXTRINSIC QUALITY CUE. .................................................................... 12 2.4 SUMMARY OF THEORY ............................................................................................................................ 15 3. PREVIOUS RESEARCH .......................................................................................................................... 15 3.1 FOOD PACKING DESIGN ATTRIBUTES AND CONSUMERS PREFERENCES ................................................... 15 3.2 THE SPECIAL ISSUE OF COLOUR ............................................................................................................... 20 3.3 SUMMARY OF PREVIOUS RESEARCH ........................................................................................................ 22 4. RESEARCH QUESTIONS........................................................................................................................ 23 4.1 HYPOTHESES .......................................................................................................................................... 23 5. METHOD .................................................................................................................................................... 24 5.1 PARTICIPANTS ........................................................................................................................................ 24 5.2 PROCEDURES .......................................................................................................................................... 26
5.2.1 The stimuli ................................................................................................................................................. 26 5.2.2 Survey procedures ...................................................................................................................................... 31 5.4.1 Conjoint analysis ........................................................................................................................................ 33 5.4.2 Cluster analysis .......................................................................................................................................... 34 5.4.3 Statistical tests ............................................................................................................................................ 35

5.3 MEASURES.............................................................................................................................................. 31 5.4 ANALYSIS ............................................................................................................................................... 33

6. RESULTS .................................................................................................................................................... 36 6.1 THE IMPORTANCE OF THE PACKAGE DESIGN ATTRIBUTES AND THE PREFERRED FEATURES OF THEM ...... 36 6.2 THE EFFECT OF PRODUCT INVOLVEMENT LEVEL ..................................................................................... 38 6.3 THE EFFECT OF HEALTHY EATING LIFESTYLE .......................................................................................... 46 7. DISCUSSION.............................................................................................................................................. 47 8. CONCLUSIONS AND RECOMMENDATIONS.................................................................................... 50 8.1 KEY RESULTS.......................................................................................................................................... 50 8.2 CONCLUSION .......................................................................................................................................... 51 8.3 STRATEGIC AND TACTICAL MARKETING RECOMMENDATIONS ................................................................ 51 8.4 RESEARCH IMPLICATIONS ....................................................................................................................... 52 REFERENCES ............................................................................................................................................... 53 APPENDIX I: THE QUESTIONNAIRE ..................................................................................................... 56 APPENDIX II: THE STIMULI CARDS...................................................................................................... 59

1. INTRODUCTION 1.1 Background During the last decades in food consumption have been observed several trends that have to do with changes in social and economic environment as well as in lifestyle. The working patterns, the increased employment of women, the lack of time, the income growth and several other factors led to the trend for fast food, convenience foods and food out of home (Petzoldt, Joiko & Menrad 2008). But the same time for many decades now, the citizens of Western developed countries have moved away from primary production and they procure the necessary food from supermarkets grocery stores and various kinds of large and small food stores, in small or large cities where they live. Thus urbanization has resulted in increased consumption of packaged and processed food; consumers are now seeking their food in packages on the shelves of food stores. Packaged foods have the largest market share of food in developed countries and there is a trend for big increase in their market share in developing countries as the income of consumers increases (USDA 2011). The distance between the production of raw material and food consumption, as well as the various food crises that have arisen recently, has made consumers to feel more or less insecure and suspicious for what they eat. This fact and the recommendations of scientists and nutrition experts have led to a new trend in food consumption, the trend for a healthy diet. Health has been named as the most significant trend and innovation driver in the global food and drinks market (Meziane 2007). Health plays a dominant role in contemporary discussion of food. Healthy eating is regarded as one of the most important means of health promotion not only in political programmes and strategies, but also in public discussion. According to many studies, consumers are also increasingly reflective in matters of health and willing to adopt health oriented changes in their eating habits (Niva 2007). Actions to further improve healthy eating have followed two major avenues. The first, and more traditional one, has focused on providing information about what constitutes healthy eating. Promotion efforts have created awareness and understanding of healthy eating: to eat a varied diet, more fruit, vegetables and fish and less fatty and sugary food, calories and salt. The second, and more recent one, comprises attempts to improve the

healthiness of products. This is done by either adding functional components or removing dysfunctional ones, resulting in functional food products, or by adding or reducing the content of certain nutrients, such as reducing fat or sugar. These products have been marketed on their health benefits using nutrient content claims or health claims to the extent possible under the legal constraints and have experienced high growth rates (Bech-Larsen & Grunert 2003; Bech-Larsen & Scholderer 2007; Chrysochou et al. 2010). Therefore consumers are not only aware about the composition of a healthy diet but also can find on the shelves of supermarkets plenty packaged food products that claim they can improve consumers health in various ways. As the factor health has become very important for the diet of modern consumers, food companies are asked to meet this need. So the competition around the factor health is becoming more and more intense among food manufacturers. Marketers of food companies are asked to overcome this competition using every possible marketing tool and trying to create the best possible marketing mix. They have to pass to the consumers the message of how healthy is their food product. Among other means to communicate this message is the food package itself. Packing is a communication device providing details about the product, including price, contents, ingredients and nutritional value as well as cooking instructions and recommended use by dates (Ahmed, Ahmed & Salman 2005), (the role of food package is going to be discussed more analytically in the next chapter). The aim of the of the communication functions of packing is to inform the consumers about the product it contains and assist them in making their decisions carefully. An example of such significant information is food labeling. The trend towards healthier eating has highlighted the importance of labeling, which allows consumers the opportunity to cautiously consider alternatives and make informed food choices. Package layout is important for information presentation. However, packaging information can create confusion by conveying either too much information or misleading and inaccurate information (Silayoi & Speece 2007). To maximize the information carried on products, manufacturers often use very small fonts and very dense writing styles. This reduces readability and sometimes causes confusion. To overcome this problem food industry and the regulation authorities have suggested that nutritional information panels should be laid out in the same way for all food products so that they are easy to understand

quickly. Also different ways of more visible and more comprehensive information panels have been implied, like GDAs (Guideline Daily Amounts). But food companies that like to promote their health products cannot base their communication efforts upon these generic information labels, they want to overcome competition, to differentiate their product. That is why they use very short health claims at the front side of the package, they try to attract consumers with an easy to understand message. Nutrition and health claims are strong marketing incentives for the food industry (Bech-Larsen & Scholderer 2007), providing opportunities for product differentiation based on a health-related positioning. But these claims have become subject of firm regulations by authorities in their effort to protect consumers against unsubstantiated or untruthful statements. Also research has shown that the value of health claims is depends on the type of claim, the category of product the culture and the origin of consumers (Saba et al. 2010; van Trijp & van der Lans 2007; Verbeke, Scholderer & Lhteenmki 2009) and sometimes they could have negative effects on other attributes of a food product like the perceived naturalness (Lhteenmki et al. 2010). But marketers have also another strong tool they can use in their effort to pass their product as a healthy one. This is the package design itself. Packaging has long been recognized as the silent salesperson and has been the focus of much recent regulation. For the most part, however, this regulation ignores the nonverbal package label components. The verbal elements of a package are accurate reflections of the products characteristics. Package graphics, however, can be used to strengthen or weaken the marketers explicit verbal claims or to generate inferences that could conflict with the verbal information (Bone & France 2001). Package design attributes like colour, shape, pictures etc carry a message about the product as well. Maybe this nonverbal message is an undercover one but not less important. Thus it is critical for food companies and their marketers who try to place a food product as a healthy one to be able to manipulate the package design attributes in a creative manner that can support the healthy message. They have to design food packages in a way that these products to be attractive for consumers who seek healthy food products.

1.2 Objectives This study aims to examine if different nonverbal attributes of food and beverage packages affects consumers expectations of food and beverages product healthiness. We will try to find out how important are some design attributes of food and beverage packages for their healthy image. Also we are going to find out not only how important are these attributes but furthermore in which way they affect consumers preferences. There will be an effort to link consumers attitudes towards the package design attributes with their involvement in the kind of food product as well as their healthy eating habits. The final goal of this study is to help food industry, food marketers and food package designers to understand consumers preferences about the design of healthy food and beverage packages.

1.3 Limitations For the purposes of this study an experiment took place with particular food products (cereals and fruit juice). The package design includes four attributes which vary in two particular levels. Other products, other attributes and different levels may lead more or less in different results. Furthermore the design of the experiment as well as the illustration of the packages which been shown to the participants, done with the limited skills of the author. Also the sample of the study comes from the urban population of Athens Greece. Athens is a big city of the European South with its particular social, economic and cultural characteristics. Therefore the results cannot be over-generalized

1.4 Overview This thesis is structured in 3 main parts. Chapters 1 to 4 lay the foundation of the work to be undertaken. This is followed by the empirical methodology and results in chapters 5 to 7. Chapter 8 highlights the findings of this study where conclusions are drawn and recommendations are presented. The content of each part of this thesis is summarized in the following short paragraphs: Theoretical Framework defines the role of health as a food quality factor, the role food packing as an extrinsic quality cue and a theoretical model the Total Food Quality Model that can link the expectations for quality with extrinsic and intrinsic quality cues. 8

Previous Research in chapter 3 covers research studies on consumers preferences for food package. Past research can indicate which the important package design attributes are for the consumers attention and preferences. Research Questions in chapter 4 are derived from the theoretical framework as well as the findings of previous research studies. Hypotheses are developed for each one of these questions based on existing literature. Chapter 5 describes the Method of this survey. It starts with a description of the sample which was recruited for this experiment. The experimental stimuli are described as well as the analysis methods used for hypotheses testing. Chapter 6 contains the results of the experiment. Actually is the outcome of the analysis section of the previous chapter. In Chapter 7, findings of this study are discussed with reference to their implications and their limitations in the experimental setting. Chapter 8 concludes with recommendations to food marketers and food package designers and suggestions for future research.

2. THEORETICAL FRAMEWORK 2.1 Health as food quality dimension The concept of health is very broad and can be approached from different scientific perspectives, including medical, nutritional, social and psychological. Here we are going to discuss health from a consumer point of view. Consumers view aspects like nutrition as an important aspect of their own health but the way consumers subjectively perceive nutritional effects may be different from the perspective of a nutritionist. From a consumer point of view, health involves two main dimensions: eating healthily and avoiding unhealthy foods. The first dimension, eating healthily, is related to nutritional aspects such as a healthy diet, functional foods, less fatty foods and other factors related to health and nutrition. The second dimension, avoiding unhealthy foods, deals with concerns about food safety. Food safety has been the subject of intensive public debate lately, and includes such diverse phenomena as BSE, pesticide residues in food, salmonella, and risks arising from novel production methods like genetic modification. The common

denominator is the potentially adverse impact of the consumption of food on health. Both health dimensions (eating healthily versus avoiding unhealthy food) express qualities of the food that consumers cannot evaluate or judge by themselves, and are thus credence characteristics. Consumers do not usually, and do not expect to, feel healthier because they have eaten a product that is supposed to be good for them at least not in the short run (Bruns, Fjord & Grunert 2002). The health quality dimension raises two types of questions. The first refers to how consumers evaluate the health quality dimension of food, i.e., which cues do they use to infer healthiness and safety, and how do they perceive the health quality of different food products? The second refers to how the evaluation of health aspects enters the buying decision, to what extent is healthiness an ultimate condition of purchase (Bruns, Fjord & Grunert 2002)?

2.2 The Total Food Quality Model The health dimension seems to be one of many attributes that influence the perceived quality of food products and finally the purchase decision of consumers. The Total Food Quality Model (Total Food Quality Model), originally proposed by Grunert, Larsen, Madsen and Baadsgaard (1995), integrates the multi-attribute and the hierarchical approaches to quality perception. In addition, it integrates two other major elements of consumer behaviour theory, namely the explanation of intention to purchase, as a trade-off between give and get components and the explanation of consumer satisfaction, as the discrepancy between expected and experienced quality. The model is shown in figure 1. It should be noted that a number of similar models have been proposed in the literature (Andersen 1994; Poulsen et al. 1996; Steenkamp & van Trijp 1996). First of all, the Total Food Quality Model distinguishes between before and after purchase evaluations. As already mentioned, many characteristics of a food product cannot be ascertained before purchase, ie most food products have only search characteristics to a limited degree. In order to make a choice, the consumer will develop expectations about quality but it is only after consumption that experienced quality can be determined (and even this is limited in the case of credence characteristics like healthiness). The distinction between before and after purchase thus forms the basis of the Total Food Quality Model.

10

Figure 1: The Total Food Quality Model

In the before purchase part, the model shows how quality expectations are formed based on the quality cues available. Cues are pieces of information used to form quality expectations. The intrinsic quality cues cover the physical characteristics of the product and are related to the products technical specifications, which also include its physiological characteristics, ie characteristics which can be measured objectively. The extrinsic quality cues represent all other characteristics of the product, such as brand name, price, distribution, outlet, packaging, etc. The way consumers use quality cues to infer expected quality can be quite intricate and, at first sight, sometimes appear to be quite irrational. For example, consumers use the colour of meat to infer tenderness, the consistency of yoghurt to infer taste, and packaging in bottles (compared with cartons) to infer wholesomeness. Of all the cues consumers are exposed to, only those which are perceived will have an influence on expected quality. The cues consumers are exposed to and those they perceive are affected by the shopping situation: the amount of information in the shop, whether purchases are planned or spontaneous, the pressure of time while shopping, etc. According to the Total Food Quality Model, quality is not an aim in itself, but is desired because it helps satisfy purchase motives or values. The model therefore includes motive or value fulfillment, ie how food products contribute to the achievement of desired

11

consequences and values. Extrinsic cues such as a label and its information may, for example, generate expectations about exceptionally high eating quality giving the consumer a feeling of luxury and of pleasure in life. The values sought by consumers will, in turn, have an impact on which quality dimensions are sought and how different cues are perceived and evaluated. The sequence from cues, through quality, to purchase motives forms a hierarchy of increasingly abstract cognitive categories. In this way, the Total Food Quality Model integrates the means-end model of consumer behaviour. Expected quality and expected fulfillment of the purchase motive constitute the positive consequences consumers expect from buying a food product and are offset against the negative consequences in the form of various (mostly monetary) costs. The trade-off determines the intention to buy. After the purchase, the consumer will have a quality experience, which often deviates from expected quality, especially when it is based on quality cues with a low degree of predictive power, as mentioned above. The experienced quality is influenced by many factors. The product itself, especially its sensory characteristics (in an objective sense, as measured by a sensory panel), is obviously one determinant, but there are many others: the way the product has been prepared, situational factors such as time of day and type of meal, the consumers mood, previous experience, etc. And the expectation itself may also be an important variable in determining the experienced quality of the product. The relationship between quality expectation and quality experience (eg before and after purchase) is commonly believed to determine product satisfaction, and consequently the probability of purchasing the product again (Bruns, Fjord & Grunert 2002).

2.3 The role of food package as extrinsic quality cue. Food package is the container that holds, protects, preserves and identifies the product, and which also facilitates its handling, storage and commercialization. Packaging also plays a major role in attracting consumer attention and influencing consumer purchase decisions. In the context of current self-service food retailing, packaging provides food companies the last chance to persuade consumers to buy the product before brand selection. Therefore, all packaging elements have to be combined to attract the consumer when purchasing the product (Ares & Deliza 2010b).

12

Apart from influencing consumer purchase decision, food package may also create expectations in the consumer (Ares & Deliza 2010b; Deliza & MacFie 1996). If the hedonic expectations created by the package are high, the consumer may be interested in the product and choose to buy it. However, sensory and hedonic expectations could also affect consumer response when tasting the product. When the product is tasted, the expected sensory characteristics of the product are compared with the products real characteristics, leading to confirmation or disconfirmation (Deliza & MacFie 1996). If the consumer confirms his/her expected sensory characteristics he/she would likely repeat product purchase. However, if the expected sensory characteristics are not perceived, the consumer will probably not buy the product again. Therefore, manufacturers should use food package to attract consumers attention in order to increase their interest in buying the product; but also to generate sensory and hedonic expectations that match the products real characteristics. The packages overall features can underline the uniqueness and originality of the product. Quality judgments are largely influenced by product characteristics reflected by packaging, and these play a role in the formation of brand preferences. If the package communicates high quality, consumers frequently assume that the product is of high quality. If the package symbolizes low quality, consumers transfer this low quality perception to the product itself (Silayoi & Speece 2004; Underwood, Robert L. , Klein & Burke 2001). The package becomes the symbol that communicates favorable or unfavorable implied meaning about the product. Underwood et al. (2001) suggest that consumers are more likely to spontaneously imagine aspects of how a product looks, tastes, feels, smells, or sounds while viewing product pictures on the package. A review of the relevant literature indicates that there are four main packaging elements potentially affecting consumer purchase decisions. They can be separated into two categories; visual and informational elements. The visual elements consist of graphics and size/shape of packaging. Informational elements relate to product information and information about the technologies used in the package (Silayoi & Speece 2007). For the purposes of this study we are going to cope only with the design-visual elements of the package.

13

The effect of colour is the most obvious and well studied. Consumer perceptions of an acceptable colour are associated with perceptions of other quality attributes, such as flavor and nutrition, and also with satisfaction levels. Positive effect can be achieved by manipulating one or more packaging variables, including packaging colour, clear packs that allow viewing food colour, incident light, and nomenclature and brand name appearance (Imram 1999; Silayoi & Speece 2007). Visual imagery on the package is another essential attribute. To be noticed at the point of sale, pictures on the package can be a strategic method of differentiation, which will enhance access to consumer consciousness. This is because pictures are extremely vivid stimuli compared to words (Underwood et al. 2001) and also is quicker and easier for consumers to process in a low involvement situation. Visual packaging information may attract consumer attention and set expectations for content. A well-produced product image is likely to evoke memorable and positive association with the product. Size and shape also emerges as a crucial dimension. One way in which consumers appear to use these things is as a simplifying visual heuristic to make volume judgments. Generally, they perceive more elongated packages to be larger, even when they frequently purchase these packages and have experience using them. Disconfirmation of package size after consumption may not lead consumers to revise their volume judgment sufficiently in the long term, especially if the discrepancy is not very large (Raghubir & Krishna 1999). Different packaging sizes potentially appeal to consumers with somewhat different involvement. For example, for some low involvement food products, such as generics, low price is made possible through cost savings created by reduced packaging and promotional expenses. Since generics are usually packaged in large sizes, this directly caters to the needs of consumers from larger households, who are more likely to be specifically looking for good deals. They find the low price of the generics, in larger packaging, is an attractive offer with excellent value for money. In addition, this could imply that when product quality is hard to determine, the effect of packaging size is stronger. Thus, elongating the shape, within acceptable bounds, should result in consumers thinking of the package as a better value for money and result in larger sales generally (Silayoi & Speece 2007). We are going to discuss more about the design-visual elements of food package and the research that have taken place about it at the next chapter of this study. At moment it

14

should be clear the role of food package as an extrinsic quality cue and therefore its role as a communication vehicle for food marketing.

2.4 Summary of theory In this chapter we discuss the health factor as a food quality dimension, the role of food packing as a quality indictor and we present the Total Food Quality Model a theoretical framework that can explain the food quality perception by consumers and how it can lead to particular buying behaviour. The Total Food Quality Model is an augmented and holistic theoretical model but for the purposes of this study we are going to use only the part of it that link the package design as extrinsic quality cue, with the expectations about the quality dimension of healthiness (Figure 2).

Fingure 2: Part of the TFQM that is used for this study

3. PREVIOUS RESEARCH 3.1 Food packing design attributes and consumers preferences There are extensive research and the relevant literature about consumers preferences of food package design and part of this research link package design attributes with consumers expectations. Nevertheless limited research has been conducted about food package design and its effects on consumers expectations of food product healthiness. Bone and France shown in an article of 2001 that the graphical component of a food label (colour and image) can significantly influence attribute beliefs and purchase

intentions even when very concrete verbal information is used (health claims. This suggests that graphical information could be misleading and affect the consumers buying behaviour. This effect is valid for both higher motivation and lower motivation subjects (Bone & France 2001).

15

Ares and Deliza have conducted interesting research about consumer attitudes towards food package design characteristics. Along with Besio and Gimenez they study the influence of different package attributes on consumer willingness to purchase regular and functional chocolate milk desserts and assess if the influence of these attributes was affected by consumers level of involvement with the product. Consumers level of involvement with the product affected their interest in the evaluated products and their reaction towards the considered conjoint variables, suggesting that it could be a useful segmentation tool during food development. Package colour and the presence of a picture on the label were the variables with the highest relative importance, regardless of consumers involvement with the product. The importance of these variables was higher than the type of dessert indicating that packaging may play an important role in consumers perception and purchase intention of functional foods. Brown colour instead of black and the presence of a milk dessert picture on the label showed positive part-worth utilities, indicating that they significantly increased consumers purchase intent. Furthermore, package shape (square or round) did not significantly affect willingness to purchase for any of the identified consumer segments (Ares et al. 2010). Also in another article, they present word association and conjoint analysis to study the influence of package shape and colour on consumer expectations of milk desserts. Both colour and shape significantly affected consumers associations, expected liking and willingness to purchase. Consumer associations regarding package colour were mainly related to flavor. On the other hand, differences in consumer associations due to differences in package shape were mainly related to sensory expectations related to texture characteristics (e.g. runny and thick) and to specific types of desserts (e.g. egg custard or low-calorie desserts). Apart from influencing expected flavour, package colour also affected texture expectations regarding the creaminess and softness of the desserts. Yellow round packages were more frequently associated with creamy and soft desserts than black or white ones. These results show the relevance of package characteristics, such as colour and shape, in creating sensory expectations on consumers, which could affect their product perception and acceptance. The same researchers conducted a free list research in order to indentify the most important food package design attributes and they find out that the most important for consumers are colour, shape, picture and the design of the label (Ares & Deliza 2010a).

16

Another recent study by Mizutani et al. shows that images on juice packages influence flavor evaluation. It demonstrates that pleasant images provide positive effects on palatability and the perceived freshness of juice even when incongruent images that are not directly related to the juice are presented. It was also revealed that juices presented with congruent images were rated as having a better aroma than those presented with incongruent images. These findings provide experimental evidence for the efficacy of pleasant images on commercial food products: if the package image is pleasant and congruent, it has positive effects on the consumers perception of the product (Mizutani et al. 2010). Underwood and Klein carried out an empirical research and shown that placing a picture of the product on food packages significantly improve brand beliefs and has positive effects on attitudes towards the package (Underwood, Robert L. & Klein 2002). Empirical results from a virtual reality simulation that they have carried out along with Burke, show that package pictures increase shoppers attention to the brand. However this effect is contingent, occurring only for low familiarity brands (private-label brands) within product categories that offer a relatively high level of experiential benefits. These results suggest that package pictures may be especially useful for private label brands and/or lesser tier national brands whose strategic objectives are to improve consumers perceptions of the brand and enter the consideration set. Silayoi and Speece (2007) perform a conjoint analysis in order to investigate what is the importance of different packing design attributes for consumers. Then they use the results to segment the market of Bangkok, Thailand. Results show strong segmentation in response to packaging. The three segments, convenience oriented, information seeking, and image seeking, follow patterns common worldwide. To some extent, this suggests that on a broad level, middle class urban consumer behaviour in Thailand is becoming similar to other developed countries. Within each of the three segments, none of the importance weights becomes negligible for any element. In other words, these consumers view the package as a coherent whole, stressing one aspect or another, but not completely ignoring any element. There may not be a single ideal design for the whole market, but the most effective single package would probably need to have a technology image which clearly

17

conveys convenience and ease of use; list clear product information, and have more classic, traditional graphic design, colours, and shape. The conjoint results indicate that packaging technology (which conveys a message of convenience and ease of use in this study) plays the most important role in consumer likelihood to buy. The other attributes included in this study were not much different from each other in importance. Packaging shape had a slight edge, followed by product information, colour and graphics and finally layout of graphics and information, but these are actually minor differences which are not statistically significant. Precise product information has a positive utility score, while the vague presentation of information had a negative utility. Straight shape has a positive utility compared to curvy, as does classic design on the package compared to colourful. This suggests that, overall the respondents may be more attracted to a package that seems familiar and reliable, rather than exciting. The design of food products packaging should be able to convey healthiness and safety, rather than excitement. Classic and calm graphics may better indicate the quality of the product inside. Also the shape should not be too fancy. Consumers seem to rely on traditional shapes that they are familiar with. Layout of graphics and information utility scores indicate that the position of graphics on the right and product information on the left is more effective. This result is particularly interesting for packaging developers because it is not consistent with the findings in psychology research in the West (Silayoi & Speece 2007). Rettie and Brewer have shown that recall is better for verbal stimuli when the copy is on the right-hand side of the package, and better for non-verbal stimuli which are on the left-hand side (Rettie & Brewer 2000). In another paper that they published, Silayoi and Speece (2004), utilized a focus group methodology to understand consumer behaviour toward packaging design of food products and how packaging elements can affect buying decisions. Visual package elements play a major role, representing the product for many consumers, especially in low involvement, and when they are rushed. Most focus group participants say they use label information, but they would like it if simplified. The results of this focus group study suggested that in general, visual elements of the package influence choice of the product to a great extent, and graphics and colour are frequently the major influence. Attractive

18

packaging generates consumer attention by breaking through the competitive clutter. Picture vividness has the most positive impact for products with lower levels of involvement. However, informational elements are becoming increasingly important and influence choice. The participants tended to judge food product performance by reading the label if they were considering products more carefully. Appropriately delivered information on packaging generates strong impact on the consumers purchase decision. Consumer evaluation of packaging elements changes as the perceived risk of the consumption situation increases. Visual elements, graphics and size/shape, positively influence choice more in the low involvement situation, while informational elements tend to play a key role in higher involvement decision-making. Time pressure similarly changes how consumers evaluate products at the point of sale, partly by reducing ability to give attention to informational elements(Silayoi & Speece 2004). Across a series of three studies, Madzharov & Block (2010) demonstrate that the number of product units displayed on a package biases consumers' perceptions of product quantity (i.e., the number of snack items the package contains) and actual consumption. Specifically, we demonstrate that consumers use an anchoring heuristic to infer that packages that display a greater number of product units (e.g., 15 pretzels vs. 3 pretzels) have a higher product quantity inside. Importantly, we demonstrate that actual consumption of the food product follows this anchor judgment. The studies demonstrate that these effects are moderated by level of visual processing and that they are robust even in the presence of verbal information (Madzharov & Block 2010). The main objective of a Marshalls, Stuarts & Bells study (2006) was to determine the role of packaging colour in product selection among preschoolers, by age and gender, across three product categories: cereals, biscuits and drinks. The three product categories, with logo and brand information obscured, were presented with a range of nine colours. The children were asked to choose one package from each category for themselves, one package from each category for a boy, and one package from each category for a girl. They were then asked why they had chosen the packages and asked about their favorite colour. The results showed a high correlation between favorite colour and choice of product across the total sample, with lower correlations for individuals. Favorite colours were pink (24%), purple (11.4) yellow and blue (both 9%) and most popular colours were pink

19

(40.9%), followed by purple (15%) and yellow (15%). Correlations were lower when selecting for boys and girls, with younger children more likely to select colours that matched their own preferences (Marshall, Stuart & Bell 2006).

3.2 The special issue of colour Colour is an integral part of products, services, packaging, logos, and other collateral and can be an effective means of creating and sustaining brand and corporate images in customers minds. Thus colour and its meanings have been well examined by marketing and psychological research. That is why we devote a special part of this chapter for the colour attribute An associative learning framework can be used to explain human physiological response to colour. Researchers have suggested that colour associations may have been formulated early in human history when man associated dark blue with night, and therefore, passivity and bright yellow with sunlight and arousal. To this day, cool colours, such as blue and green, are considered calming and warm colours, such as red and orange, are considered arousing (Grossman & Wisenblit 1999). Because colours have specific meanings associated with them, colours are important image cues. The meanings of colours may also have implications for the associations that consumers make about a brands position in the marketplace. Certain colours manifest pan cultural meaning associations. Strategic use of these colours affords opportunities for products, packages, logos, and the like to convey specific images associations across national markets (Madden, Hewett & Roth 2000). Moriarty (1991) discusses colour, mentioning that colour in advertising serve a variety of specific purposes. Colour is used for a variety of specific purposes. It creates moods, it draws attention, it emphasizes, and it intensifies memorability. Morarity further states that colour can be used as a cue, to either associate with or symbolize something else. Morarity also claims that the primary function of colour in advertising is to help create mood and emotional responses. In packaging, the first objective of the colour is to command the eye. It has to be seen, to jump of the shelf, if it is to survive the intense competition of the self-service

20

environment. Next the package colour is chosen for its ability to be associated with certain desired qualities such as elegance, naturalness, softness, and so on (Moriarity 1991). The notion that colour preferences are formulated through associations is a potentially important finding for marketing practitioners interested in determining colours for products. Rather than examine general colour preferences among consumers, it may be preferable to learn consumers colour associations as a basis for understanding the emotional aspects of colour. For example, Marketers can also use the theory of associations to create meanings for particular colours or to develop a brand image around a colour (Grossman & Wisenblit 1999). Here below the basic colours and their what they commonly are associated with will be presented. Red: Red is perceived to be generally exciting, cheerful, disobedient, and powerful. It is also associated with heat, anger, passion, war, and blood. It is considered stimulating (Moriarity 1991). Other characteristics associated with red are expensive, premium, high quality, good tasting. Young, warm, fun, loud, playful and happy. Dangerous, adventurous, luxurious and exciting. Life, love, passion, power and aggression (Grimes & Doole 1998). Blue: Often considered a happy colour. It means peace, calmness, loyalty, security, and tenderness. It is also associated with intellectual appeals as opposed to emotional (red). It can be identified with cold, ice, distance, and infinity as well as calm reflection (Moriarity 1991). Blue is also perceived heavy, reliable, high quality and expensive. Male mature, quiet, subdued calm and thoughtful (Grimes & Doole 1998). Yellow: Associated with the sunlight and openness as well as radiance and vividness. Because of its brightness, it is highly attention getting. (Moriarity 1991) High quality, expensive, reliable, light and good tasting are other characteristics associated with yellow. Yellow is also associated with luxury, sophistication and to some extent safety, but also life, happiness, tenderness and warmth. (Grimes & Doole 1998) Green: Like blue, also associated with serenity and calmness as well as nature. It is a quiet colour, used symbolically to suggest hope, meditation, and tranquility. (Moriarity 1991) It is also perceived inexpensive, light, reliable and good tasting. Old, traditional, trustworthy, life, tenderness, health, and happy, environment, natural, pure and fresh are all characteristics associated with green. (Grimes & Doole 1998)

21

Black: Associated with distress, hopelessness, and defiance. Used to suggest hate and death, but can also be used to express power and elegance, especially if it is shiny. (Moriarity 1991) Black is also associated with expensive, high quality, hi-tech and premium products. Mysterious, luxurious, sophisticated and dangerous, dignity, power and aggression are all associated with black. (Grimes & Doole 1998) White: In Western cultures, white means purity as well as sanitary and clean. By its lack of apparent colour, it also conveys emptiness, infinity, and the incomprehensible. White is used visually to express total silence. (Moriarity 1991) Madden, Hewett and Roth (2000) conclude that an interesting pattern of colours forming a spectrum of meanings is evident across all countries. The meaning associations along this spectrum run from active, hot, and vibrant (associated with red) to calming, gentle, and peaceful (associated with the blue-green-white cluster). The remaining colours tend to locate approximately equidistant between these endpoints (Madden, Hewett & Roth 2000).

3.3 Summary of previous research The review of relevant literature that took place in this chapter indicates that many attributes of package design influence the preference of consumers for food products in different ways. Colour, shape and graphics seem to be the most important attributes, variance of which influences the preferences of consumers respectively. Vivid colours like red and yellow may be more attractive for the eye at the place of purchase but calming ones like green or blue have been related with naturalness and health. Also the existence of a graphic on the package seems to have positive effect on consumers preferences especially if it is pleasant, congruent, classic and calm. About the shape, it seems that consumers preffer straight shape than curvy one, because it looks more familiar and reliable. As it seems all these packing design attributes which are the visual elements of the package (colour, graphics and size/shape), positively influence choice more in the low involvement situation, while informational elements tend to play a key role in higher involvement decision-making. Also seems that consumers level of involvement with a particular food product affect their interest in the evaluated products and their reaction towards the different package design attributes.

22

4. RESEARCH QUESTIONS As it is stated in the introduction this study aims to examine if different nonverbal attributes of food and beverage packages affects consumers expectations of food and beverages product healthiness. The study is based on the part of the Total Food Quality Model that links the food package as extrinsic quality cue with healthiness as a food quality factor. The most important packing design attributes, as they come from the literature and the intuition of the author will be examined for the expectations of food product healthiness that they create to consumers. Also it is going to take place an effort to link consumers attitudes towards the package design attributes with their involvement in particular food product as well as their lifestyle about healthy eating. The particular research questions that this study will try to answer are the followings:

Q1: How important are particular package design attributes (colour, shape, graphics and visibility through the package) for the expectations of food product healthiness that they create to consumers?

Q2: How the variance of the above mentioned package design attributes affects consumers expectations of food product healthiness?

Q3: Does the consumers level of involvement with particular food products affect their reaction towards the different package design attributes for the healthiness expectations that they create?

Q4: Does the consumers healthy eating lifestyle affect their reaction towards the different package design attributes for the healthiness expectations that they create?

4.1 Hypotheses Literature review as well as authors intuition suggests the following hypothesis for the research questions that have been raised:

23

H1: All the food package design attributes are important for the healthiness expectations that they create but the colour is the most important one.

H2: Calm and cold colours, straight and classic shapes, the image of the product in use on the package as well as the existence of transparent part on the package affect positively the healthiness expectations of it.

H3: Consumers level of involvement with particular food products affects their reaction towards the different package design attributes for the healthiness expectations that they create.

H4: Consumers healthy eating lifestyle affects their reaction towards the different package design attributes for the healthiness expectations that they create.

5. METHOD 5.1 Participants For the purposes of this study primary data were used. The study was conducted in the city of Athens (Greece) using a convenience sample. All participants were recruited at university campus and public places as well as within the social network of the author. Participants were recruited based on their willingness to participate and whether they are responsible for the food shopping of their household or with the help of another member of their household. Consumers who were not responsible for the food shopping at all were excluded by the sample. A field experiment took place, participants were shown pictures of food packages and then were asked to fulfill a questionnaire. In total seventy three people participate in the survey (N=73). Of them a number of 43 (58,9 %) are females and the remaining 30 (41,1 %) males (Table 1). Females are the majority of the sample because females used to be responsible for food shopping more often than males.

24

For better reading of the results seven age categories were generated 18-25, 26-30, 31-35, 36-40, 41-45, 46-50, and >50. The majority of the participants belong to two age categories 26-30 and 31-35. In these two categories belong the 75,5% of the participants. This is explained by the fact that the sample was a convenience one based on the authors social network. The other age categories are underrepresented (Table 1).
Table 1: Basic data of the sample Frequency Gender Male 30 Female 43 Total 73 Education level No school Primary school Secondary educ. 1st level Secondary educ. 2nd level Post-secondary educ. Technical higher educ. University Postgraduate studies Total

% 41,1 58,9 100,0 Age 18 25 26 30 31 35 36 40 41 45 46 50 >50 Total

Frequency 6 29 26 6 4 1 1 73

% 8,2 39,7 35,6 8,2 5,5 1,4 1,4 100,0

0 0 0 3 4 11 20 35 73

0,0 0,0 0,0 4,1 5,5 15,1 27,4 47,9 100,0

Responsibility for food shopping Me 25 Me and another 48 Another 0 Total 73

34,2 65,8 0,0 100,0

About the education level of participants, 8 levels were predefined. The majority of the sample belongs to the three upper end categories. More than 90% of participants have at least a higher education degree. As in the case of age the reason for the underrepresentation of the other categories, is the way the sample has been formed. But in any case we can claim that the sample represents the most dynamic segment of the Greek consumers. As has been mentioned above people who are not responsible at all for food shoping in their household have been excluded from this survey. So our sample consists by consumers that are responsible for their food shopping themselves, 34,2% as well as by consumers who are partly responsible for this 65,8%.

25

5.2 Procedures For the purposes of this study an experiment was conducted. Participants were called to evaluate models of food packages as per the healthiness impression that they create. Also their product involvement level as well as their healthy eating lifestyle was investigated through special variables. The healthiness expectations that are created by packing design attributes, were investigated at the sample as a whole, as well as at segments of it which were formed based on different involvement and healthy eating lifestyle levels. We are going to discuss the analysis procedures later on this chapter. 5.2.1 The stimuli At the designing phase of the experiment it should be decided three very critical issues. Firstly, what kind of food products will be used for the experiment? Secondly, which packing design attributes will be examined for the healthiness expectations that they create? Thirdly, between which levels these attributes will be varied?

The products The food products which are going to be used in an experiment like that should have some particular features as about their interaction with the consumers. They should be well known and be of wide consumption. They should not be considered as healthy or unhealthy by default. For instance a candy product is generally considered as unhealthy, so someone can expect that its packing plays a very limited role in generating healthiness expectations. The opposite can be implied for the fresh milk, usually it is considered as healthy by default. Of course the products for this experiment should be met well packaged on supermarkets shelves. Having all the above in mind, it was decided that the experiment will make usage of two products, one food and one beverage. These two products are: whole grain cereals and fruit juice (peach). Both of them seem that comply the above mentioned features.

26

Attributes and levels The next important decision that should be taken for the experiment was about the attributes of the package design that have to be examined. The literature was the guide for this, along with the inspiration of the author. Undoubtedly colour is a very important factor in packing design. This fact is not only obvious in the literature (Ares et al. 2010; Ares & Deliza 2010a, 2010b; Grossman & Wisenblit 1999; Madden, Hewett & Roth 2000; Marshall, Stuart & Bell 2006; Moriarity 1991; Silayoi & Speece 2007) but can be considered as a common place. Shape seems also to play an important role in consumers preferences on food package design (Ares et al. 2010; Ares & Deliza 2010b; Rettie & Brewer 2000; Silayoi & Speece 2004; Silayoi & Speece 2007). Another factor that influences consumers preferences is the existence of graphics (images) on the package (Ares & Deliza 2010a; Madzharov & Block 2010; Nancarrow, Wright & Brace 1998; Rettie & Brewer 2000; Silayoi & Speece 2004; Silayoi & Speece 2007; Underwood, Robert L. & Klein 2002). The issue of graphics/ images has been examined in different ways, what kind of image is preferred, where the image should be placed etc. Literature insists also some other attributes of package design that may influence consumers preferences, like the package size (Bellizzi & Hite 1992; Raghubir & Krishna 1999; Silayoi & Speece 2007) or the indication of packing technology that is used (Silayoi & Speece 2007). According the opinion of the author these attributes are not of such worth to be examined for the healthiness expectations that they create. On the other hand another attribute of package design which had not be examined widely till now, is the existence on the package of a transparent part through which one can see the content of the package. According my opinion, this attribute is worth to be examined for the healthiness expectation that creates. So finally four package design attributes are going to be examined. These are the colour, the shape, type of graphics and the existence of transparent part. The question now is about the variation levels of each attribute. For the ease of experimental and analysis procedures, it is decided each attribute varied within two levels. The two levels of colour will be red and green. Literature suggests that these two colours have a significant distance on the spectrum of colours meaning (Madden, Hewett & Roth 2000). Green is associated with serenity and calmness as well as nature, while red is perceived to be generally exciting, cheerful, disobedient, and powerful (Moriarity 1991). So

27

it is considered that it is worth to be examined the influence of these two colours on consumers expectation of packaged food products healthiness. The shape is examined as square/box or curvy/round (Ares et al. 2010; Silayoi & Speece 2007). It must be mentioned that consumers are accustomed to buy certain foods in particular types/shapes of packages which dominate the market. That means it depends on the type of product if one or the other shape is considered as classic and straight or the other. The attribute of graphics was decided to vary between these two levels: a picture of the product in use (a bowl with cereals, a glass of juice) or a picture with a natural landscape. About the fourth attribute, the transparent part in the package, it is obvious that this will vary between the existence or not of this part. Table 2 shows all the package design attributes that examined in this experiment and their levels.

Table 2: Attributes and Levels Attributes Levels Colour Red Green Shape Square Curvy/ round Graphics Image of the product in use Image of a landscape Visibility Existence of a transparent part on the pack Not existence of a transparent part on the pack

Designing the stimuli After the products, the attributes and their levels have been defined, the next step is the creation of real models to be evaluated by the participants in this experiment. The variance of all the four attributes in two levels each, gives sixteen possible combinations, in other words sixteen models that have to be evaluated by each participant (2 levels amd 4 attributes, 2x2x2x2=16 combinations). This number is considered rather high for the ability of each participant to evaluate with consistency all the given models, furthermore if we take into account that each participant evaluates two kind of products, cereals and juice. So it was decided the number of the models under evaluation to be reduced. For this a fractional

28

factorial design took place, so the number of the models reduced in eight for each product (cereals and juice). The resulting concepts of the fractional factorial design are shown in Table 3. The final models were given a code, A1 to A8 for the cereals and X1 to X8 for the juice.

Table 3: All the models presented to participants and coreponding attribute levels. Product Colour Shape Graphics Visibility Cereals A1 A2 A3 A4 A5 A6 A7 A8 Juice X1 X2 X3 X4 X5 X6 X7 X8 Red Green Red Green Red Green Red Green Square (carton box) Square (carton box) Square (carton box) Square (carton box) Curvy/ round (bottle) Curvy/ round (bottle) Curvy/ round (bottle) Curvy/ round (bottle) Image of the product Landscape Landscape Image of the product Landscape Image of the product Image of the product Landscape Transparent Transparent Not transparent Not transparent Transparent Transparent Not transparent Not transparent Green Red Green Red Green Red Green Red Square (carton box) Square (carton box) Square (carton box) Square (carton box) Curvy (nylon bag) Curvy (nylon bag) Curvy (nylon bag) Curvy (nylon bag) Image of the product Landscape Landscape Image of the product Landscape Image of the product Image of the product Landscape Transparent Transparent Not transparent Not transparent Transparent Transparent Not transparent Not transparent

For the purposes of this experiment real models of product packages were constructed by the author. The construction of the models was based to the above table (Table 3). All the graphics that printed on the models were created with the GIMP (GNU Image Manipulation Program) software. All the sixteen models were photographed in black background and the photos were printed in high resolution, in A4 illustration paper (210 x 297 mm). These photos were the cards that finally shown to the participants. These cards are shown in total in Figure 3 and 4. All of the cards are demonstrated in Appendix II in higher resolution. 29

Figure 3: The 8 packages of cereals of the experiment

Figure 4: The 8 juices of the experiment

30

5.2.2 Survey procedures All the participants were asked to evaluate all the eight cereals packages as well as all the eight juices packages for the healthy image of them. Initially the eight cards of cereals were demonstrated in front of them in a random way and orally they were explained what they have to do: You have to examine all of these packages carefully and then to evaluate them according to the healthiness perception that each one creates to you. Then the questionnaire of the survey was given to them to fill it in. After the evaluation of the cereal packages was completed, participant fulfilled some additional question about his/ her involvement in cereals (more details in the next section of this chapter). Then the same process was followed about the juice packages. At the end the participant was asked some supplementary questions about his/ her healthy eating lifestyle as well as demographic questions (see next section). The whole procedure for each participant lasted about ten to twelve minutes.

5.3 Measures The evaluation of the perceived healthiness of the presented packages was conducted by asking participants the question: How healthy it appears to you the cereals/ juice you see in the picture? This was the dependant variable of the experiment. The answer of the above question was a number from 1 to 10 (Likert scale) where 1 means not healthy at all and 10 means absolutely healthy. The evaluation of the eight packages of cereals (codes A1 to A8) according the above mentioned question was the first part of the questionnaire. The next part of it had questions which aimed to examine the product (cereals) involvement level of participants. For the measurement of the product involvement level a reduced version of the Personal Involvement Inventory (PII) was used. PII is a well established research tool for the measurement of involvement levels in consumer behaviour research, but its twenty items were considered as too many for this survey. That is why the reduced version of it was preferred. Zaichkowsky (1994) suggests that a reduced version of the PII can also be reliable including ten items instead of twenty. He estimates the reliability of the ten items PII to has a Cronbach Alphas > 0,9 in all cases. Thus participants asked to answer the ten items of the reduced PII about their involvement level in cereals, in 31

a 7 point Likert scale. The average answers are shown in Table 4.The reliability of the reduced PII was re-tested. The Cronbach Alphas for cereals was 0,85 and for juices was 0,89. That shows high reliability. The third and the forth part of the questionnaire repeat the first two parts of it, but this time for juice packages and involvement. Means are shown also in Table 4. The fifth part includes three questions that aimed to measure the healthy eating lifestyle of the participants, These particular items were suggested for the measurement of the healthy eating lifestyle by Grunert, Bruns and Bisp (1993) when they developed the Food Related Lifestyle (FRL), a tool that measures the lifestyle about food (Grunert, KG, Bruns & Bisp 1993). In this study the reliability of these items tested and be found with a Cronbach Alphas 0,86. The sixth and last part of the questionnaire consists of the demographic questions, analysis of which have been shown above in Table 1. The questionnaire was in Greek language. A translated version in English is presented in Appendix I. It contains in total 39 closed questions plus 4 demographic ones. Means and standard deviation of answers of all items (except demographic ones) are shown in Table 4.
Table 4: Means and Standard Deviation of the questionnaire items Question/ item Mean St. D. Question/ item How healthy it appears to you the How healthy it appears to you the cereals you see in the picture? juice you see in the picture? A1 6,67 1,77 X1 A2 6,74 1,82 X2 A3 6,05 2,23 X3 A4 5,36 2,11 X4 A5 6,08 1,94 X5 A6 5,47 1,95 X6 A7 4,42 1,99 X7 A8 4,52 2,08 X8 To me wholegrain cereals are: To me fresh packaged juices are: important/ unimportant* 2,77 1,58 important/ unimportant* boring/ interesting 4,48 1,59 boring/ interesting relevant/ irrelevant* 3,29 1,80 relevant/ irrelevant* exciting/ unexciting* 3,42 1,47 exciting/ unexciting* means nothing/ means a lot to me 4,97 1,45 means nothing/ means a lot to me appealing/ unappealing* 3,52 1,51 appealing/ unappealing* fascinating/ mundane* 4,11 1,47 fascinating/ mundane* worthless/ valuable 5,53 1,28 worthless/ valuable involving/ uninvolving* 3,04 1,69 involving/ uninvolving* not needed/ needed 5,78 0,98 not needed/ needed I prefer to buy natural products, ie products without preservatives. To me the naturalness of the food that I buy is an important quality. I try to avoid food products with additives. Mean St. D.

4,21 4,62 5,62 5,23 7,15 6,77 4,93 5,19 2,22 5,45 2,55 2,67 5,52 2,78 3,10 5,85 2,47 5,99 6,21 6,15 5,95

2,32 2,41 2,25 2,04 2,07 1,93 2,09 2,25 1,41 1,27 1,50 1,30 1,29 1,36 1,17 1,25 1,50 1,17 1,34 1,34 1,55

32

5.4 Analysis Two main types of data analysis were conducted for the purposes of this study. The basic one was Conjoint Analysis which was used to identify the importance of each one of the package design attributes, as well as the preferred level of each attribute. The other type of analysis is Cluster Analysis which was used to define consumers' segments based on their product involvement levels and their healthy eating lifestyle. Then Conjoint Analysis was repeated on the different segments that had been defined, in order to detect any difference in preferences between the different segments. 5.4.1 Conjoint analysis Conjoint analysis is a widely used technique in consumer research, where respondents are presented with product descriptions generated according to a factorial design of product attributes. Respondents evaluate the various product descriptions (by ranking or rating) or make a choice between them, and the responses are used to infer how the various attributes contribute to the overall evaluation, using a variety of statistical techniques ranging from traditional ordinary least squares to ordered logit models. In the context of the Total Food Quality Model, conjoint analysis provides a rigorous way of analyzing the vertical relationships between cues and expected quality and/or purchase intent (Bruns, Fjord & Grunert 2002). In this study we use a part of the Total Food Quality Model in order to investigate the linkage between the extrinsic cues of the package design with expectations of the quality dimension of healthiness. Thus conjoint analysis looks like an ideal choice. The dependent variable of the analysis is the expectations of healthiness of cereals and juices packages, while the independent ones are the package design attributes. Y= X1 + X2 + X3 + X4 Y = Expectations of healthiness by the products packages X1 = Colour of the package X2 = Shape of the package X3 = Graphics type X4 = Visibility

33

The traditional conjoint analysis approach was selected along with the additive model. That means that respondents evaluate stimuli constructed with selected levels of each attribute (full profiles) and the total value for a combination of attributes is got by adding up the valued for each attribute (part-worths). The total utility of any defined stimulus can be calculated as the sum of the parts (Hair et al. 2006). The preferences of responders collecting with a rating system as it has described at the previous section of this chapter. Rating was preferred instead of ranking because it is more practical for use in online surveys like this. Given that eh preference measure used was a metric rating, the traditional regression-based approach could be employed as the part-worths estimation model. The estimation of part-worths of each attribute was first performed for each respondent separately and the results were then aggregated to obtain an overall result. The estimation of part-worths can provide us with the importance of each package design attribute for the expectation of the product healthiness. The part-worths estimates are in common scale, so we can compute the relative importance of each factor directly. The importance of each factor is represented by the range of its levels divided by the sum of the range across all factors. The calculation provides a relative impact or importance of each attribute based on the size of range of its part-worths estimates. Factors with a large range for their part-worths have a greater impact on the calculated utility values and thus are deemed of greater importance. The relative importance scores across all attributes will total 100 percent (Hair et al. 2006). Also the comparison of the part-worths of the levels for each attribute indicates the proffered level, the higher the part-worth the higher the preference for the corresponding level. The conjoint analysis was carried out with Microsoft Excel 2007 software and the estimation model was designed under the guidance of the Multivariate Data Analysis of Hair et al. (2006).

5.4.2 Cluster analysis The above mentioned \Conjoint analysis initially will be employed to the total number of participants in order aggregated results to be generated. Then the sample of the survey will segmented in terms of product involvement and healthy eating lifestyle and conjoint analysis is going to be implied to each segment separately. This segmentation will take place through a Cluster analysis. This kind of analysis is a common way for market 34

segmentation. In this case we are going to segment the market according the answers of the participants in the reduced type of Personal Involvement Inventory questions as well as the healthy eating lifestyle questions. Applying cluster analysis we are trying to separate the market through our sample into two groups each time. These groups will be respectively: those who are cereals involved and cereals uninvolved, those who are fruit juices involved and fruit juices uninvolved and finally those who are healthy eating fans and not so much healthy eating fans. The aim is to find out if there is any significant difference about the package design preferences between the different segments. For all three cluster analysis we employed, the similarity measure that was chosen was the squared Euclidean distance since all the clustering variables are metric. Also because all the clustering variables are measured in the same scale (1 to 10) the variables need not to be standardized. The Cluster analysis will be conducted using the Hierarchical method. The algorithm that has been chosen for the hierarchical procedure is the Wards method. A stopping rule of 5 clusters has been set. Although we would like to use two clusters, we choose to examine more cases to have an overall view and be sure that the solution of two clusters is a good one. The ideal cluster number is based on the rapid change of the heterogeneity of clusters solutions. When the agglomeration coefficient has the largest percentage change, means that the solution prior the change is the most appropriate one. After the clusters have defined we create their profile examining the demographics of each one. Also we are going to examine whether these profiles are distinctive applying a chi-square test to indentify if the demographic viarables are significant different between the clusters. Cluster analysis was conducted using the SPSS 17, Statistical package software. 5.4.3 Statistical tests The next step of the analysis process is to examine whether there is significant difference between the segments in terms of consumers preferences about the package design attributes. Also we have to examine whether there is any significant difference between the consumers preferences about the different products, the cereals and the juice. This will be done by applying t-tests and One-way Analysis of Variance (ANOVA). The data which will feed the tests will be the mean score that each level received by the 35

respondents. A paired-sample t-test will be applied for the investigation of significant differences between the products and the ANOVA tests will be conducted in the case of the different segments that will have been identified.

6. RESULTS 6.1 The importance of the package design attributes and the preferred features of them The first research question of this study as it has been formulated in Chapter 4, was about the importance of each one of the package design attributes in terms of healthiness expectations that they cause to consumers. The second one was about the variation of these attributes, how the different levels of these attributes influence consumers expectation of food product healthiness. The Conjoint analysis that employed, gave us some answers on these questions through the estimation of the part-worths of each attribute level and the range of them. The estimation of part-worths of each attribute was first performed for each respondent separately and the results were then aggregated to obtain an overall result. These estimations became separately for cereals and juices.

Cereals About the cereals, Conjoint analysis indicates that the most important package designing factor in terms of healthiness expectations is Visibility at a rate of 39,81%, followed by the Shape 37,44%. Graphics and \colour have a rather low importance with 12,80% and 9,95% respectively. Consumers prefer to see the product itself through a transparent part of the package. They find this attribute the most important one and the existence of this transparent part has a positive part-worth of 1,397, while the absence of a transparent part has a part-worth of -1,397. Also consumers prefer the square shape for the cereals packages since the part-worth of it is 1,314. The same the green colour and the image of a landscape have positive part-worths, 0,349 and 0,449 respectively. So consumers expect that cereals that are packed in square boxes, with a transparent part on them and have green colours and images of natural landscapes, are more healthy. These results are shown in Table 5.

36

Table 5: Attribute importance and preferred levels of cereals packages Attribute Importance Preferred level Part-worth of preferred level Visibility 39,81 % Transparent 1,397 Shape Graphics Colour 37,44 % 12,80 % 9,95 % 100,00 % Square (carton box) Landscape Green 1,314 0,449 0,349

Juices The same procedure for juices indicates that in this case the most important attribute is the shape (56,79 %), and visibility come next with an importance at 23,21 %. The other two attributes are much less important, Graphics with 18,57 % has the third place, while Colour comes last with 1,43 %. About the most important attribute, shape, consumers appear to prefer curvy/round shapes (part-worth of 1,771), something that is in contrast with the case of cereals. But for the next two attributes, their preferences are the same with the case of cereals; they prefer packages with transparent parts (part-worth 0,714) and images of natural landscapes (part-worth 0,579) on the packages. The red colour is more preferred for juice packages (part-worth 0,045) unlike the cereals case. The profile of the ideal juice package in terms of healthiness expectations can be extracted from these results as they shown in Table 6.

Table 6: Attribute importance and preferred levels of juice packages Attribute Importance Preferred level Part-worth of preferred level Shape 56,79 % Curvy/ round (bottle) 1,771 Visibility Graphics Colour 23,21 % 18,57 % 1,43 % 100,00 % Transparent Landscape Red 0,714 0,579 0,045

Next a paired-sample T-test was employed in order to identify any significant difference to the way the different levels have been evaluated for the two products. The feed data for the test were the average score of each level, as it results from the evaluation of each stimulus. The results of the test are shown in the Table 7.

37

Table 7: Mean score for each design attribute level and product Paired-sample T-test Attribute Levels Mean score in Mean score in t Cereals Juice Colour Green 5,81 5,46 2,070 Red Shape Square Curvy/round Visibility Transparent Not Transparent Graphics Image of the product Landscape 5,52 6,21 5,12 6,24 5,09 5,48 5,85 5,48 4,92 6,01 5,68 5,25 5,29 5,64 0,269 6,840 -4,079 2,958 -0,813 1,148 1,220

Sig. (2-tailed) 0,04 0,79 0,00 0,00 0,00 0,42 0,26 0,23

The test indicates that in a significant level of 0.05, both levels of the shape are differently approached between the two products. Square shape has significant higher score for cereals and low score for juices and the opposite. Also the existence of transparent part on the package is significantly more important for cereals rather than juices. The last level that has a significant different evaluation between the products is the green colour. It receives significantly higher score in cereals than in juices.

6.2 The effect of product involvement level The third research question as it is mentioned in the fourth Chapter is about the levels of consumers product involvement and whether these levels have any significant effect on preferences about package design. In order we to be able to answer this question, firstly we ask consumers about their product involvement, using the reduced Personal Involvement Inventory tool. Then we analyze the data through Cluster analysis to segment the market in terms of the particular products (cereals and juices) involvement levels.

Cereals Cluster analysis for the case of cereals produced two clusters, Cluster 1 consist by 35 participants or the 47,9% of the market and Cluster 2 with 38 participants or the 52,1% of the market. Although we have already considered a solution of two clusters as an appropriate one for the aims of this study, hierarchical analysis that conducted suggested the same. The Agglomerate coefficient change from 1022,47 to 1612,87 (57,74%) between

38

the one and the two clusters solution. The percentage change is the biggest one while the next one is between the two and the three clusters solution (21,90%). Cluster 1 consists of consumers who evaluate with higher score the positive meanings of the PII about cereals, while Cluster 2 by them who evaluate with higher score negative meanings. This fact leads us to name Cluster 1: Involved in cereals and Cluster 2: Uninvolved in cereals Mean scores for each item and cluster are shown in Table 8. It must be mentioned that some items are reversed scored, i.e. higher rating scores pond to low level of involvement with the product. Table 8 presents also the results of the ANOVA test that indicates that both clusters are discrete in terms of all PII items.

Table 8: Reduced PII average score for the cereal market segments ANOVA test Cluster 1 (N=35) Cluster 2 (N=38) Important - unimportant* Boring - interesting Relevant - irrelevant* Exciting - unexciting* means nothing - means a lot to me Appealing - unappealing* Fascinating - mundane* Worthless - valuable Involving - uninvolving* not needed - needed 1,57 5,54 2,42 2,45 5,88 2,74 3,31 6,34 1,85 6,42 3,86 3,50 4,07 4,31 4,13 4,23 4,84 4,78 4,13 5,18

F 82,315 50,839 19,218 48,104 41,511 23,368 26,817 42,066 60,488 49,724

Sig 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00

*Items that are reversed scored, i.e. higher rating scores pond to low level of involvement with the product.

After the two segments have been indentified, the profile of them in terms of demographics was created. This profile is shown in Table 9. The only one demographic variable that differs significantly between the two clusters is the gender of consumers (chisquare test with p<0,05). The Involved in cereals segment is dominated by women, while in the uninvolved segments men are the majority. Another comment that could be made is that involved consumers are more normally distributed within the age categories, while uninvolved ones are grouped mainly in younger age categories.

39

Table 9: Demographic profile of the two defined cereals involvement segments. Cluster 2: Cluster 1: Involved in cereals Uninvolved in cereals 38 (52,1%) 35 (47,9%) N % N % Gender Males: 8 22,9 22 57,9 Females: 27 77,1 16 42,1 Age 18 25: 26 30: 31 35: 36 40: 41 45: 46 50: >50: Secondary educ. 2nd level: Post-secondary educ. : Technical higher educ.: University: Postgraduate studies: Me: Me and another: Another: 2 19 7 4 2 1 0 0 2 8 10 15 15 20 0 5,7 54,3 20,0 11,4 5,7 2,9 0,0 0,0 5,7 22,9 28,6 42,9 42,9 57,1 0,0 4 10 19 2 2 0 1 3 2 3 10 20 10 28 0 10,5 26,3 50,0 5,3 5,3 0,0 2,6 7,9 5,3 7,9 26,3 52,6 26,3 73,7 0,0

Chisquare

Sign.

9,240

0,002

11,561

0,073

Education

5,874

0,209

Food shopping

2,214

0,137

As the two cereals involvement segments have been denitrified in terms of size as well as in demographic profile, we can run conjoint analysis to each one of them to investigate the particular preferences of the consumers of them about cereals package design and the healthiness expectations that it generates. Thus Conjoint analysis indicates that the most important package designing factor for cereals involved consumers in terms of healthiness expectations is the Shape at a rate of 48,06%, followed by Visibility 40,29%. Graphics and \colour have a rather low importance with 6,80% and 4,851% respectively. Cereals uninvolved consumers assess Visibility as the most important attribute with 39,35%, followed by the shape with 27,31%. Graphics and Colour are the less important attributes for them too, but they do not evaluate them with such a low score as the involved ones. For them Graphics has an importance at 18,52% amd Colour at 14,81%. Table 10 presents the importance of each attribute for each cluster. Also the impotence levels for each cluster as well as for the cereals case as a whole are presented graphically in Chart 1.

40

Table 10: Attribute importance and preferred levels of cereals packages for the two defined segments Attributes Visibility Shape Graphics Colour Cluster 1: Involved in cereals Importance Preferred Part-worth of level pref. level 40,29% 48,06% 6,80% 4,851% Transparent Square (carton box) Landscape Green 1,274 1,519 0,215 0,153 Cluster 2: Uninvolved in cereals Importance Preferred Part-worth of level pref. level 39,35% 27,31% 18,52% 14,81% Transparent Square (carton box) Landscape Green 1,472 1,022 0,693 0,554

Consumers preferences for particular designing levels appear to be identical for both segments; therefore they are the same as in the unsegmented case that examined before. Consumers of both segments prefer to see the product itself through a transparent part of the package. The existence of this transparent part has a positive part-worth of 1,274 and 1,472 in the two clusters respectively. Also consumers prefer the square shape for the cereals packages since the part-worth of it is 1,519 for the involved segment and 1,022 for the uninvolved. The green colour and the image of a landscape have positive part-worths for both segments too. Green colour has part-worths of 0,153 and 0,554 respectively, while the landscape image 0,215 in cluster 1 and 0,693 in cluster 2. These results are shown in Table 10.

Chart 1: Importance of cereals package design attributes for each segment and the market as a whole

41

Table 11: Mean score for each cereals package design attribute level and cluster - ANOVA test Attribute Levels Mean score F Cluster 1 Cluster 2 Colour Green 5,91 5,71 0,370 Red 5,77 5,28 1,983 Shape Square Curvy/round Visibility Transparent Not Transparent Graphics Image of the product Landscape 6,55 5,13 6,43 5,25 5,74 5,94 5,88 5,11 6,05 4,94 5,23 5,76 3,510 0,003 1,078 0,591 2,222 0,271

Sig. 0,54 0,16 0,06 0,95 0,30 0,44 0,14 0,60

The question now is whether the two defined segments differ significantly each other, in terms of cereals package design preferences and the healthiness expectations that this design can generate. In order statistically significant differences between the preferences of the two segments to be identified, an ANOVA test was employed on the data of the average evaluation score for each level (see Table 11). Generally involved consumers evaluated with higher score all the levels. But the ANOVA test indicates that there are no significant differences between the two segments in a significant level of 95% in terms of design levels evaluation. Only for the square shape the evaluation of the two segments is significant different in a significant level of 90%. Cereal involved consumers appreciate the square/box shape of the cereals package more than the uninvolved ones.

Juice The same procedures as with the cereals case will be followed for the investigation of the relationship between the fruit juice involvement level and the preferences of juice package design in terms of healthiness expectations that it generates. Cluster analysis for the case of juices produced two clusters, Cluster 1 consists of 46 participants or the 63,0% of the market and Cluster 2 with 27 participants or the 37,0% of the market. Although we have already considered a solution of two clusters as an appropriate one for the aims of this study, hierarchical analysis that conducted suggested two clusters as well. The Agglomerate coefficient change from 761,82 to 1268,30 (increase of 66,48%) between the one and the two clusters solution. The percentage change is the biggest one.

42

Table 12: Reduced PII average score for the juices market segments ANOVA test Cluster 1 (N=46) Cluster 2 (N=27) Important - unimportant* Boring - interesting Relevant - irrelevant* Exciting - unexciting* means nothing - means a lot to me Appealing - unappealing* Fascinating - mundane* Worthless - valuable Involving - uninvolving* not needed - needed 1,63 6,00 1,84 2,06 6,26 2,15 2,58 6,45 1,65 6,56 3,22 4,51 3,74 3,70 4,25 3,85 3,96 4,81 3,85 5,00

F 30,797 33,671 42,804 42,428 92,973 41,872 34,594 48,242 73,205 51,644

Sig. 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00

*Items that are reversed scored, i.e. higher rating scores pond to low level of involvement with the product.

Table 13: Demographic profile of the two defined juices involvement segments. Cluster 1: Cluster 2: Involved in juices Uninvolved in juices 46 (63,0%) 27 (37,0%) N % N % Gender Males: 18 39,1 12 44,4 Females: 28 60,9 15 55,6 Age 18 25: 26 30: 31 35: 36 40: 41 45: 46 50: >50: Secondary educ. 2nd level: Post-secondary educ. : Technical higher educ.: University: Postgraduate studies: Me: Me and another: Another: 2 20 16 5 3 0 0 1 2 8 10 25 17 29 0 4,3 43,5 34,8 10,9 6,5 0,0 0,0 2,2 4,3 17,4 21,7 54,3 37,0 63,0 0,0 4 9 10 1 1 1 1 2 2 3 10 10 8 19 0 14,8 33,3 37,0 3,7 3,7 3,7 3,7 7,4 7,4 11,1 37,0 37,0 29,6 70,4 0,0

Chisquare

Sign.

0,198

0,656

7,450

0,281

Education

4,387

0,356

Food shopping

0,406

0,534

43

Cluster 1 consists of consumers who evaluate with higher score the positive meanings of the PII about juices, while Cluster 2 by them who evaluate with higher score negative meanings. This fact leads us to name Cluster 1: Involved in juices and Cluster 2: Uninvolved in juices. Mean scores for each item and cluster are shown in Table 12. Table 12 presents also the results of the ANOVA test that indicates that both clusters are discrete in terms of all PII items. After the two segments have been indentified, the profile of them in terms of demographics was created. These profiles are shown in Table 13. The two defined segments do not differ significantly each other in any demographic variable. As the two juice involvement segments have been denitrified in terms of size as well as in demographic profile, we can run conjoint analysis to each one of them to investigate the particular preferences of the consumers of them about juice package design and the healthiness expectations that it generates. Thus Conjoint analysis indicates that the most important package designing factor for juices involved consumers in terms of healthiness expectations is the Shape at a rate of 56,77%, followed by Visibility 23,96%. Graphics is at the thisrd place with 17,19% and \colour has a rather low importance with 2,08%. Juices uninvolved consumers assess also Shape as the most important attribute with 56,82%, followed by Visibility and Graphics which have equal importance at 21,59%. Colour appears to have no importance at all for uninvolved consumers (0,00%). Table 14 presents the importance of each attribute for each cluster. Also the impotence levels for each cluster as well as for the juices case as a whole are presented graphically in Chart 2.

Table 14: Attribute importance and preferred levels of juices packages for the two defined segments Attributes Visibility Shape Graphics Colour Cluster 1: Involved in juices Importance Preferred level Part-worth of pref. level 23,96% 56,77% 17,19% 2,08% Transparent Curvy/ round (bottle) Landscape Red 0,749 1,774 0,537 0,065 Cluster 2: Uninvolved in juices Importance Preferred level Part-worth of pref. level 21,59% 56,82% 21,59% 0,00% Transparent Curvy/ round (bottle) Landscape Indifferent 0,669 1,762 0,669 0,000

44

Chart 2: Importance of juice package design attributes for each segment and the market as a whole

Consumers preferences for particular designing levels appear to be almost the same for both segments; therefore they are almost the same as in the unsegmented case that examined before. Consumers of both segments prefer to see the product itself through a transparent part of the package. The existence of this transparent part has a positive partworth of 0,749 and 0,669 in the two clusters respectively. Also consumers prefer the Curvy/ round shape for the juice packages since the part-worth of it is 1,774 for the involved segment and 1,762 for the uninvolved. The image of a landscape has also positive partworths for both segments too (0,537 and 0,669 respectively). While involved consumers appear to have a slight preference for the red colour (part-worth 0,065), uninvolved ones are indifferent between green and red colour since the part-worth is 0,000. These results are shown in Table 14.

Table 15: Mean score for each juice package design attribute level and cluster - ANOVA test Attribute Levels Mean score F Cluster 1 Cluster 2 Colour Green 5,44 5,48 0,010 Red 5,48 5,46 0,003 Shape Square Curvy/round Visibility Transparent Not Transparent Graphics Image of the product Landscape 4,86 6,05 5,71 5,21 5,28 5,64 5,00 5,93 5,63 5,30 5,29 5,64 0,090 0,081 0,028 0,039 0,001 0,000

Sig. 0,91 0,96 0,76 0,77 0,86 0,84 0,97 0,98

45

In order statistically significant differences between the preferences of the two segments to be identified, an ANOVA test was employed on the data of the average evaluation score for each level (see Table 14). The test indicates that there are no significant differences between the two segments in terms of design levels evaluation, neither at a significant level of 95% nor at 95%.

6.3 The effect of healthy eating lifestyle The fourth research question was about the effect of different healthy eating habits on consumers preferences for food package design in terms of the healthiness expectations that it generates. The approach that be employed to answer this question was the same as in the product involvement case. We used the health factor items of the Food Related Lifestyle research tool in order to segment the market in terms of healthy eating habits. Cluster analysis that based on data from these items provided a two cluster solution with a cluster of 69 consumers and another of just 4. (see Table 16). Thus it was considered meaningless to proceed further, conducting the conjoint analysis on such unbalanced segments. Almost all the participants in the survey belong to the cluster of healthy eating fans. The mean scores of this cluster for all the three items reach the upper end of the seven point Likert scale. The other cluster consists of just the 5% of the sample. Actually participants appear not to be differentiated in terms of their healthy eating lifestyle.

Table 16: Segmentation based on healthy eating habits Cluster 1 (N=69) I prefer to buy natural products, ie products without preservatives. To me the naturalness of the food that I buy is an important quality. I try to avoid food products with additives. 6,44 6,40 6,10

Cluster 2 (N=4) 2,00 1,75 3,25

F 96,491 122,791 15,259

Sig. 0,00 0,00 0,00

46

7. DISCUSSION This study aimed to investigate the linkage between healthiness expectations of food products and the design of food packages. Particular research question were set and an experiment with real prototypes was conducted in order the research questions to be answered. The whole attempt was based on the Total Food Quality Model that links intrinsic and extrinsic food quality cues with quality expectations. The first question is about the importance of the different package design attributes in terms of the healthiness expectations that they generate to consumers. The experiment that conducted with two different products, a food cereals and a beverage fruit juice indicates that the most important attributes for the design of the package of a healthy food product are the shape of it as well as the visibility of the product through the package. The shape received an importance rate of 37,44% for the cereals and 56,79% for the juice, while visibility received 39,81% for the cereals and 23,21% for the juice. The other two attributes that examined appear as less important in both cases. The type of image/graphics takes the third place for cereals package design as well as for juices, with an importance of 12,80% and 18,57% respectively. Surprisingly, colour appears to be the least important, it received an important rate of 9,95% for the cereals and 1,43% for juices. The literature review suggested colour as a very important factor for food package design, but this study indicates the opposite. The reason for this could be the way that stimuli were designed and the relative importance of the other attributes that examined. At the previous research section of this study the special issue of colour was mentioned. Now the special issue of shape should be mentioned. Previous research indicates that the shape is an important package design attribute, but as it seems at least in this survey, it acts as a super attribute. This can be resulted not only by the high importance that it receives in this experiment but also by the discussion with the participants. Most of them made a comment for the package shape not always about the core meaning of shape but in general for the package type. It can be said that in this survey the shape attribute actually represented also the package type as well as the construction materials of it. This attribute is the one that make the stimulus touchable in the eyes of consumers, is the one that acts in three dimensions.

47

The shape attribute also is the only one that differs significantly between the two products for its both levels. Consumers prefer square/ box shape for the cereals while they prefer curvy/round shape for the juices. This highlights also the fact the shape acts as a super attribute .Actually consumers probably do not prefer the square shape of the cereals package but the carton box of it. Likewise they do not like the fact that the juice is in a curvy package but the fact that it is within a bottle rather within a carton box. Another issue that is raised here is the issue of habit. Consumers are used to buy particular food products in particular package types that dominate the market. So there choice may be have to do a lot with what they think as a realistic one. The most expensive and well established cereals brands are sold within square carton boxes, while corresponding fruit juices are sold within round bottles. Thus the preferences about the shape, which actually is not just shape, are affected by the product type and consumers habits. . Consumers seem to rely on traditional shapes that they are familiar with (Silayoi & Speece 2007). So we can claim that for the shape attribute a more effective distinction could be between classical and traditional ones on one hand and contemporary and innovative on the other. We just have seen the preferences of consumers about the shape attribute. The second research question had to do exactly with that; how the variance between the different levels of package design attributes affects consumers expectations of food product healthiness. As have been mentioned above the visibility attribute is the most important in the case of cereals and the second one in the case of juices. So it is obvious that consumers prefer to be able to see the product itself through its package. This has been hypothesized but still it is considered as an important finding of this survey since this attribute was under little research till now. Also the evaluation of the transparent part on food package differs significantly between the two products. It receives higher score in the case of cereals rather in the case of juices. This may be has to do with the type of product. Consumers may be are more suspicious and desire to have direct optical contact with a heterogeneous food product like cereals. On the other hand they may be more confident about homogeneous liquid products like juices. Another reason for this difference may be the beliefs about the protecting features of particular packages types. They may like a transparent bottle but may believe that it cannot protect effectively its fresh content. The issue that is raised here is the one of what someone likes and what he/she believes to be

48

right. Something may be aesthetically nice for someone, but he/she is not confident about its real value. About the colour attribute, although it seems to be of low importance, consumers preferences differ between the two products. This is because of the different evaluation of the green colour; consumers prefer it for cereals packages much more than the red one. They have a slight preference for the red colour in the case of juices. Thus as it seems colour is another product oriented designing attribute. This does not mean that green miss its value as a calming colour, relate with nature and health, but that particular products may have been associated with particular colours. Fresh juices are known for the energy and the vitality that they offer so in their case the colour of healthiness may be a warm one like red. As we have seen the graphics attribute has a moderate importance in both products. The evaluation of its levels does not differs significantly between them, in both cases consumers prefer on the package the image of a natural landscape instead of an image of the product in use. Our third research question had to do with effects of the product involvement on the preferences of food package design in terms of healthiness expectations. The findings of this survey indicate that there is no significant difference in the way that consumers of different product involvement levels evaluate the different designing features. This is common for both products, cereals as well as juices. Nevertheless it can be said that in general less involved consumers attach more importance to simple aesthetic attributes like the colour and the graphics than the involved consumers do. The last research question of this study was about the effects of healthy eating habits on consumers preferences of food package design in terms of healthiness expectations that it generates. To define consumers healthy eating lifestyle we used items of the Food Related Lifestyle tool. But the analysis of data that collected did not provide us with any effective segmentation. Almost all participants fall into one segment, they appear as healthy eating fans. Thus it does not make sense to proceed with analysis of difference of the different levels of healthy eating habits. It was not possible to answer the fourth research question, either because the research tool that used was not the appropriate one or you can claim that the lifestyle of consumers have been changed rapidly and massively towards a healthy eating habits.

49

8. CONCLUSIONS AND RECOMMENDATIONS This objective of this study was to investigate the effects of food package design on the expectations of food healthiness that it generates. The study based on the Total Food Quality Model, a theoretical framework that suggests the linkage of different intrinsic and extrinsic cues of food quality with different dimensions of food quality expectations and experiences. For the purposes of this study only the part of the TFQM that link food package as extrinsic cue with expected food healthiness was used. For the investigation of any possible effect an experiment with images of real packages stimuli was conducted. This experiment aimed to examine how different package design attributes affect the healthiness expectations.

8.1 Key results The key results of this survey as well as of the experiment included in it are the followings; The most important package design attribute in terms of healthiness expectations is the packages shape. The Shape must be seen as a super attribute, because as it seems, this attribute summarizes the three dimension status of the package thus it includes also its type as well as its construction materials. Consumers preferences for the different variance levels of the shape are strongly product oriented. Consumers prefer the square box in the case of cereals but they prefer the round bottle in the case of fruit juices. This along with previous research means that packages with classic and traditional shape for their product category make consumers to feel more confident for the product and generate higher healthiness expectations. Another also very important attribute is the visibility of the product through the package. The existence of a transparent part on the package that allows the visual contact with the food product is something that has not been researched systematical in terms of consumer behavior. This study suggests that the existence of a transparent part increases the consumers healthiness expectations of the product, especially in cases of heterogeneous food products like cereals.

50

The image that illustrated on the front side of the package has a moderate effect on healthiness expectations. Consumers seem to prefer images of nature like a landscape, independently of the product kind that the package contains.

Surprisingly, colour seems to not be of high importance. Actually it is the least important factor for every kind of product. Nevertheless the preferences for the colour seem to be slightly product oriented. But calming colours like green still have its value as indicators of healthiness.

The preferences of product involved consumers do not differ significantly from those of product uninvolved ones. Nevertheless it can be said that uninvolved consumers assess more carefully simple aesthetic attributes like graphics and colour than the involved ones.

8.2 Conclusion The conclusion of this study is that the visual and aesthetic dimension of the food packages should be taken seriously into account because it transfers undercover messages; healthiness is one of them. Of course package design is not the main vechicle to communicate the healthiness of a food product, but it plays a very important supportive role in this effort. As it seems the most important package design attributes are those which are not just aesthetic but also play a functional role. Thus package shape which is not just shape but the package itself is a very important attribute since it contains the product, it protects the product and it is touchable. So someone evaluate the shape not just according his/her personal taste but also according his/her believes for the functionality of it. The same is happened with the visibility attribute; it plays the role of the show case of the product. Some people like to be able to see the product, some others not, but also some people believe that it is safe the product to be exposure in light while some others do not. The same does not happen with simple aesthetic attributes like colour and graphics.

8.3 Strategic and tactical marketing recommendations The competition in market for healthy food products has become very intensive.. New healthy products, new healthy features, new brands enter the market continuously. Food companies should manage and overcome the competition. They have to find and 51

promote their own competence in terms of healthy food product. An important barrier in their effort to promote healthy claims is the tough regulations that aim to protect consumers against unsubstantiated or untruthful statements. Thus the situation becomes harder for food marketers. They have to use carefully every single tool that they have in order to overcome this situation and promote their product successfully. Package design is such a tool. The design attributes are not under tough regulations and the creativity that can be applied on it is unlimited. Food package design should be used by marketers as a carrier of undercover messages. Healthiness is such a message that partly could reach consumers through food package design. The results of this study could be useful for this. Positioning foods as healthy ones usually offers a price premium for the companies that manage to do so. It must not be forgotten the package is a part of the augmented product, furthermore for food products where its functionality is critical for the core product. Therefore when the positioning is healthy food, the product is a healthy food, the price is for healthy food, and promotion communicates the product as healthy food, the package owes to be the costume of a healthy food.

8.4 Research implications For the purposes of this study particular food package design attributes were examined for the healthiness expectations that they generate to consumers. These attributes are: the shape, the colour, the visibility of the product and the graphics. The survey indicates that two of them are very important; these are te shape and the visibility. We strongly believe that further research should be conducted for these two attributes. The reason is that as it is mentioned at least in this survey the shape acted as super attitude that contains also package type and materials, so further research is suggested splitting this attribute. About the visibility very little research has been done till now. Also it would be useful next studies on the topic to examine separately simple aesthetic attributes like colour and graphics from attribute that plays also a functional role like shape, visibility, packing materials etc.

(Grunert, K et al. 1996)

52

REFERENCES Ahmed, A, Ahmed, N & Salman, A 2005, 'Critical issues in packaged food business', British Food Journal, vol. 107, no. 10, pp. 760 - 80. Andersen, ES 1994, The evolution of credence goods: A transaction approach to product specification and quality control, Working paper no 21, MAPP. Ares, G, Besio, M, Gimnez, A & Deliza, R 2010, 'Relationship between involvement and functional milk desserts intention to purchase. Influence on attitude towards packaging characteristics', Appetite, vol. 55, no. 2, pp. 298-304. Ares, G & Deliza, R 2010a, 'Identifying important package features of milk desserts using free listing and word association', Food Quality and Preference, vol. 21, no. 6, pp. 621-8. Ares, G & Deliza, R 2010b, 'Studying the influence of package shape and colour on consumer expectations of milk desserts using word association and conjoint analysis', Food Quality and Preference, vol. 21, no. 8, pp. 930-7. Bech-Larsen, T & Grunert, KG 2003, 'The perceived healthiness of functional foods: A conjoint study of Danish, Finnish and American consumers' perception of functional foods', Appetite, vol. 40, no. 1, pp. 9-14. Bech-Larsen, T & Scholderer, J 2007, 'Functional foods in Europe: consumer research, market experiences and regulatory aspects', Trends in Food Science & Technology, vol. 18, no. 4, pp. 231-4. Bellizzi, JA & Hite, RE 1992, 'Environmental colour, consumer feelings, and purchase likelihood', Psychology and Marketing, vol. 9, no. 5, pp. 347-63. Bone, PF & France, KR 2001, 'Package Graphics and Consumer Product Beliefs', Journal of Business and Psychology, vol. 15, no. 3, pp. 467-89. Bruns, K, Fjord, TA & Grunert, K 2002, 'CONSUMERS FOOD CHOICE AND QUALITY PERCEPTION', The Aarhus School of Business, vol. Working paper no 77. Chrysochou, P, Askegaard, S, Grunert, KG & Kristensen, DB 2010, 'Social discourses of healthy eating. A market segmentation approach', Appetite, vol. 55, no. 2, pp. 288-97. Deliza, R & MacFie, H 1996, 'THE GENERATION OF SENSORY EXPECTATION BY EXTERNAL CUES AND ITS EFFECT ON SENSORY PERCEPTION AND HEDONIC RATINGS: A REVIEW', Journal of Sensory Studies, vol. 11, no. 2, pp. 103-28. Grimes, A & Doole, I 1998, 'Exploring the Relationships Between Colour and International Branding: A Cross Cultural Comparison of the UK and Taiwan', Journal of Marketing Management, vol. 14, no. 7, pp. 799 - 817.

53

Grossman, RP & Wisenblit, JZ 1999, 'What we know about consumers colour choices', Journal of Marketing Practice: Applied Marketing Science, vol. 5, no. 3, pp. 78 - 88. Grunert, K, Larsen, H, Madsen, TK & Baadsgaard, A 1996, Market orientation in food and agriculture, Kluwer Academic, Boston. Grunert, KG, Bruns, K & Bisp, S 1993, Food-related life style: Development of a crossculturally valid instrument for market surveillance, MAPP working paper no 12. Hair, JF, Black, WC, Babin, BJ, Anderson, RE & Tatham, RL 2006, Multivariate Data Analysis, 6th edn, Pearson Prentice Hall, New Jersey. Imram, N 1999, 'The role of visual cues in consumer perception and acceptance of a food product', Nutrition & Food Science, vol. 99, no. 5, pp. 224 - 30. Lhteenmki, L, Lampila, P, Grunert, K, Boztug, Y, Ueland, , Astrm, A & Martinsdttir, E 2010, 'Impact of health-related claims on the perception of other product attributes', Food Policy, vol. 35, no. 3, pp. 230-9. Madden, TJ, Hewett, K & Roth, MS 2000, 'Managing Images in Different Cultures: A Cross-National Study of Colour Meanings and Preferences', Journal of International Marketing, vol. 8, no. 4, pp. 90 - 107. Madzharov, AV & Block, LG 2010, 'Effects of product unit image on consumption of snack foods', Journal of Consumer Psychology, vol. 20, no. 4, pp. 398-409. Marshall, D, Stuart, M & Bell, R 2006, 'Examining the relationship between product package colour and product selection in preschoolers', Food Quality and Preference, vol. 17, no. 7-8, pp. 615-21. Meziane, Z 2007, Future Innovations in Food and Drinks to 2012. NPD, Trend Convergence and Emerging Growth Opportunities, Business Insights Ltd. Mizutani, N, Okamoto, M, Yamaguchi, Y, Kusakabe, Y, Dan, I & Yamanaka, T 2010, 'Package images modulate flavor perception for orange juice', Food Quality and Preference, vol. 21, no. 7, pp. 867-72. Moriarity, MB 1991, Creative advertising: theory and practice, Prentice-Hall, Englewood Cliffs, N.J. Nancarrow, C, Wright, LT & Brace, I 1998, 'Gaining competitive advantage from packaging and labelling in marketing communications', British Food Journal, vol. 100 no. 2, pp. 110 - 8. Niva, M 2007, 'All foods affect health': Understandings of functional foods and healthy eating among health-oriented Finns', Appetite, vol. 48, no. 3, pp. 384-93. 54

Petzoldt, M, Joiko, C & Menrad, K 2008, Factors and their impacts for influencing food quality and safety in the value chains, D4.3, University of Applied Science Weihenstephan - AG2020. Poulsen, CS, Juhl, HJ, Kristensen, K, Bech, AC & Engelund, E 1996, 'Quality guidance and quality formation. ', Food Quality and Preference, vol. 7, pp. 127-35. Raghubir, P & Krishna, A 1999, 'Vital Dimensions in Volume Perception: Can the Eye Fool the Stomach?', Journal of Marketing Research, vol. 36, no. 3, pp. 313-26. Rettie, R & Brewer, C 2000, 'The verbal and visual components of package design', Journal of Product & Brand Management, vol. 9, no. 1, pp. 56 - 70. Saba, A, Vassallo, M, Shepherd, R, Lampila, P, Arvola, A, Dean, M, Winkelmann, M, Claupein, E & Lhteenmki, L 2010, 'Country-wise differences in perception of healthrelated messages in cereal-based food products', Food Quality and Preference, vol. 21, no. 4, pp. 385-93. Silayoi, P & Speece, M 2004, 'Packaging and purchase decisions: An exploratory study on the impact of involvement level and time pressure', British Food Journal, vol. 106, no. 8, pp. 607 - 28. Silayoi, P & Speece, M 2007, 'The importance of packaging attributes: a conjoint analysis approach', European Journal of Marketing, vol. 41 no. 11/12, pp. 1495 - 517. Steenkamp, J-BEM & van Trijp, HCM 1996, 'Quality guidance: A consumer-based approach to food quality improvement using partial least squares.', European Review of Agricultural Economics, vol. 23, pp. 195-215. Underwood, RL & Klein, NM 2002, 'PACKAGING AS BRAND COMMUNICATION EFFECTS OF PRODUCT PICTURES ON CONSUMER RESPONSES TO THE PACKAGE AND BRAND', Journal of Marketing Theory and Practice, vol. Fall 2002, pp. 58 - 68. Underwood, RL, Klein, NM & Burke, RR 2001, 'Packaging communication: attentional effects of product imagery', Journal of Product & Brand Management, vol. 10 no. 7, pp. 403 - 22. USDA 2011, Global Food Markets: International Consumer and Retail Trends, Economic Research Service, USDA, <http://www.ers.usda.gov/Briefing/GlobalFoodMarkets/consumer.htm>. van Trijp, HC & van der Lans, IA 2007, 'Consumer perceptions of nutrition and health claims', Appetite, vol. 48, no. 3, pp. 305-24. Verbeke, W, Scholderer, J & Lhteenmki, L 2009, 'Consumer appeal of nutrition and health claims in three existing product concepts', Appetite, vol. 52, no. 3, pp. 684-92. 55

APPENDIX I: The questionnaire

We are going to show you 8 pictures each one of which presents a package of wholegrain cereals. Please after examine carefully all the packages, answer to the following:

How healthy it appears to you the cereals you see in the picture?
For each code that represents each one of the products you have seen, note your preference circling a number in the scale 1 to 10, where 1 = Not healthy at all and 10 = Absolutely healthy.

Product

1 2 3 4 5 6 7 8

Not healthy at all 1 1 1 1 1 1 1 1

2 2 2 2 2 2 2 2

3 3 3 3 3 3 3 3

4 4 4 4 4 4 4 4

5 5 5 5 5 5 5 5

6 6 6 6 6 6 6 6

7 7 7 7 7 7 7 7

8 8 8 8 8 8 8 8

9 9 9 9 9 9 9 9

Absolutely healthy 10 10 10 10 10 10 10 10

Please refer to wholegrain cereals generally in terms of the following opposite meanings. Circle a number in the scale 1 to 7 that is more appropriate for you, where 1 = I absolutely agree with the meaning in left and 7 = I absolutely agree with the meaning in right
For me wholegrain cereals are: important boring relevant exciting means nothing appealing fascinating worthless involving not needed 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 unimportant interesting irrelevant unexciting means a lot to me unappealing mundane valuable uninvolving needed

56

We are going to show you 8 pictures each one of which presents a package of fresh fruit juice. Please after examine carefully all the packages, answer to the following:

How healthy it appears to you the fruit juice you see in the picture?
For each code that represents each one of the products you have seen, note your preference circling a number in the scale 1 to 10, where 1 = Not healthy at all and 10 = Absolutely healthy.

Product

X1 X2 X3 X4 X5 X6 X7 X8

Not healthy at all 1 1 1 1 1 1 1 1

2 2 2 2 2 2 2 2

3 3 3 3 3 3 3 3

4 4 4 4 4 4 4 4

5 5 5 5 5 5 5 5

6 6 6 6 6 6 6 6

7 7 7 7 7 7 7 7

8 8 8 8 8 8 8 8

9 9 9 9 9 9 9 9

Absolutely healthy 10 10 10 10 10 10 10 10

Please refer to fresh fruit juices generally in terms of the following opposite meanings. Circle a number in the scale 1 to 7 that is more appropriate for you, where 1 = I absolutely agree with the meaning in left and 7 = I absolutely agree with the meaning in right
For me fresh fruit juices are: important boring relevant exciting means nothing appealing fascinating worthless involving not needed 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 unimportant interesting irrelevant unexciting means a lot to me unappealing mundane valuable uninvolving needed

57

How do you agree or disagree with the following sentences. Give your answer circling a number in the scale 1 to 7, where 1 = Totally disagree and 7 = Totally agree
.
Totally disagree Totally agree

I prefer to buy natural products, ie products without preservatives. To me the naturalness of the food that I buy is an important quality. I try to avoid food products with additives.

1 1 1

2 2 2

3 3 3

4 4 4

5 5 5

6 6 6

7 7 7

Who is responsible for food shopping inyour household? What is your age? What is your gender? What is your education level?

Me

Someone else

Me and someone else

Male No school Primary school

Female Post-secondary education Technical higher education University Postgraduate studies

Secondary educ. 1st level Secondary educ. 2nd level

Thank you for your participation!

58

APPENDIX II: The stimuli cards

59

A1

A2

A3

A4

60

A5

A6

A7

A8

61

X1

X2

X3

X4

62

X5

X6

X7

X8

63