effects of package design on consumer expectations of food product healthiness
TRANSCRIPT
Master’s thesis
EFFECTS OF PACKAGE DESIGN ON CONSUMER EXPECTATIONS OF FOOD PRODUCT HEALTHINESS
UNIVERSITY OF AARHUS AARHUS SCHOOL OF BUSINESS
Department of Marketing and Statistics
Author: Thomas Sioutis Advisor: Joachim Scholderer
July 2011
2
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.
3
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.
4
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
5.3 MEASURES .............................................................................................................................................. 31 5.4 ANALYSIS ............................................................................................................................................... 33
5.4.1 Conjoint analysis ........................................................................................................................................ 33 5.4.2 Cluster analysis .......................................................................................................................................... 34 5.4.3 Statistical tests ............................................................................................................................................ 35
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
5
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
6
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
7
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 & Lähteenmäki 2009) and sometimes
they could have negative effects on other attributes of a food product like the perceived
naturalness (Lähteenmäki 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 product’s characteristics.
Package graphics, however, can be used to strengthen or weaken the marketer’s 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.
8
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.
9
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
10
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
(Brunsø, 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 (Brunsø, 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.
11
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 product’s 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
Figure 1: The Total Food Quality Model
12
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 consumer’s 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 (Brunsø, 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).
13
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 package’s 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.
14
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
15
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).
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 consumer’s buying behaviour.
This effect is valid for both higher motivation and lower motivation subjects (Bone &
France 2001).
Fingure 2: Part of the TFQM that is used for this study
16
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).
17
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 consumer’s 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
18
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
19
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 Marshall’s, Stuart’s & Bell’s 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
20
(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 brand’s
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
21
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)
22
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.
23
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 author’s intuition suggests the following hypothesis for
the research questions that have been raised:
24
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.
25
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 author’s
social network. The other age categories are underrepresented (Table 1).
Table 1: Basic data of the sample
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%.
Frequency % Frequency %
Gender Age Male 30 41,1 18 – 25 6 8,2 Female 43 58,9 26 – 30 29 39,7 Total 73 100,0 31 – 35 26 35,6 36 – 40 6 8,2 Education level 41 – 45 4 5,5 No school 0 0,0 46 – 50 1 1,4 Primary school 0 0,0 >50 1 1,4 Secondary educ. 1st level 0 0,0 Total 73 100,0 Secondary educ. 2nd level 3 4,1 Post-secondary educ. 4 5,5 Responsibility for food shopping Technical higher educ. 11 15,1 Me 25 34,2 University 20 27,4 Me and another 48 65,8 Postgraduate studies 35 47,9 Another 0 0,0 Total 73 100,0 Total 73 100,0
26
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.
27
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
28
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
29
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 Green Square (carton box) Image of the product Transparent
A2 Red Square (carton box) Landscape Transparent
A3 Green Square (carton box) Landscape Not transparent
A4 Red Square (carton box) Image of the product Not transparent
A5 Green Curvy (nylon bag) Landscape Transparent
A6 Red Curvy (nylon bag) Image of the product Transparent
A7 Green Curvy (nylon bag) Image of the product Not transparent
A8 Red Curvy (nylon bag) Landscape Not transparent
Juice
X1 Red Square (carton box) Image of the product Transparent
X2 Green Square (carton box) Landscape Transparent
X3 Red Square (carton box) Landscape Not transparent
X4 Green Square (carton box) Image of the product Not transparent
X5 Red Curvy/ round (bottle) Landscape Transparent
X6 Green Curvy/ round (bottle) Image of the product Transparent
X7 Red Curvy/ round (bottle) Image of the product Not transparent
X8 Green Curvy/ round (bottle) Landscape 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.
30
Figure 3: The 8 packages of cereals of the experiment
Figure 4: The 8 juices of the experiment
31
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
32
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, Brunsø and Bisp (1993) when they developed the Food Related
Lifestyle (FRL), a tool that measures the lifestyle about food (Grunert, KG, Brunsø & 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 Mean St. D. How healthy it appears to you the cereals you see in the picture?
How healthy it appears to you the juice you see in the picture?
A1 6,67 1,77 X1 4,21 2,32 A2 6,74 1,82 X2 4,62 2,41 A3 6,05 2,23 X3 5,62 2,25 A4 5,36 2,11 X4 5,23 2,04 A5 6,08 1,94 X5 7,15 2,07 A6 5,47 1,95 X6 6,77 1,93 A7 4,42 1,99 X7 4,93 2,09 A8 4,52 2,08 X8 5,19 2,25
To me wholegrain cereals are: To me fresh packaged juices are: important/ unimportant* 2,77 1,58 important/ unimportant* 2,22 1,41
boring/ interesting 4,48 1,59 boring/ interesting 5,45 1,27 relevant/ irrelevant* 3,29 1,80 relevant/ irrelevant* 2,55 1,50
exciting/ unexciting* 3,42 1,47 exciting/ unexciting* 2,67 1,30 means nothing/ means a lot to me 4,97 1,45 means nothing/ means a lot to me 5,52 1,29
appealing/ unappealing* 3,52 1,51 appealing/ unappealing* 2,78 1,36 fascinating/ mundane* 4,11 1,47 fascinating/ mundane* 3,10 1,17
worthless/ valuable 5,53 1,28 worthless/ valuable 5,85 1,25 involving/ uninvolving* 3,04 1,69 involving/ uninvolving* 2,47 1,50
not needed/ needed 5,78 0,98 not needed/ needed 5,99 1,17
I prefer to buy natural products, ie products without preservatives. 6,21 1,34 To me the naturalness of the food that I buy is an important quality. 6,15 1,34
I try to avoid food products with additives. 5,95 1,55
33
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 (Brunsø, 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
34
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
35
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 Ward’s
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
36
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.
37
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 37,44 % Square (carton box) 1,314
Graphics 12,80 % Landscape 0,449
Colour 9,95 % Green 0,349
100,00 %
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 23,21 % Transparent 0,714
Graphics 18,57 % Landscape 0,579
Colour 1,43 % Red 0,045
100,00 %
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.
38
Table 7: Mean score for each design attribute level and product – Paired-sample T-test Attribute Levels Mean score in
Cereals Mean score in
Juice t Sig. (2-tailed)
Colour Green 5,81 5,46 2,070 0,04
Red 5,52 5,48 0,269 0,79
Shape Square 6,21 4,92 6,840 0,00
Curvy/round 5,12 6,01 -4,079 0,00
Visibility Transparent 6,24 5,68 2,958 0,00
Not Transparent 5,09 5,25 -0,813 0,42
Graphics Image of the product 5,48 5,29 1,148 0,26
Landscape 5,85 5,64 1,220 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
39
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) F Sig
Important - unimportant* 1,57 3,86 82,315 0,00
Boring - interesting 5,54 3,50 50,839 0,00
Relevant - irrelevant* 2,42 4,07 19,218 0,00
Exciting - unexciting* 2,45 4,31 48,104 0,00
means nothing - means a lot to me 5,88 4,13 41,511 0,00
Appealing - unappealing* 2,74 4,23 23,368 0,00
Fascinating - mundane* 3,31 4,84 26,817 0,00
Worthless - valuable 6,34 4,78 42,066 0,00
Involving - uninvolving* 1,85 4,13 60,488 0,00
not needed - needed 6,42 5,18 49,724 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 (chi-
square 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.
40
Table 9: Demographic profile of the two defined cereals involvement segments. Cluster 1:
Involved in cereals 35 (47,9%)
Cluster 2: Uninvolved in cereals 38 (52,1%)
Chi-square
Sign.
N % N % Gender Males: 8 22,9 22 57,9 9,240 0,002 Females: 27 77,1 16 42,1 Age 18 – 25: 2 5,7 4 10,5 11,561 0,073 26 – 30: 19 54,3 10 26,3 31 – 35: 7 20,0 19 50,0 36 – 40: 4 11,4 2 5,3 41 – 45: 2 5,7 2 5,3 46 – 50: 1 2,9 0 0,0 >50: 0 0,0 1 2,6 Education Secondary educ. 2nd level: 0 0,0 3 7,9 5,874 0,209 Post-secondary educ. : 2 5,7 2 5,3 Technical higher educ.: 8 22,9 3 7,9 University: 10 28,6 10 26,3 Postgraduate studies: 15 42,9 20 52,6 Food shopping
Me: 15 42,9 10 26,3 2,214 0,137 Me and another: 20 57,1 28 73,7
Another: 0 0,0 0 0,0
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.
41
Table 10: Attribute importance and preferred levels of cereals packages for the two defined segments Cluster 1: Involved in cereals Cluster 2: Uninvolved in cereals
Attributes Importance Preferred level
Part-worth of pref. level
Importance Preferred level
Part-worth of pref. level
Visibility 40,29% Transparent 1,274 39,35% Transparent 1,472
Shape 48,06% Square (carton box)
1,519 27,31% Square (carton box)
1,022
Graphics 6,80% Landscape 0,215 18,52% Landscape 0,693
Colour 4,851% Green 0,153 14,81% Green 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
42
Table 11: Mean score for each cereals package design attribute level and cluster - ANOVA test Attribute Levels Mean score F Sig. Cluster 1 Cluster 2 Colour Green 5,91 5,71 0,370 0,54
Red 5,77 5,28 1,983 0,16
Shape Square 6,55 5,88 3,510 0,06
Curvy/round 5,13 5,11 0,003 0,95
Visibility Transparent 6,43 6,05 1,078 0,30
Not Transparent 5,25 4,94 0,591 0,44
Graphics Image of the product 5,74 5,23 2,222 0,14
Landscape 5,94 5,76 0,271 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.
43
Table 12: Reduced PII average score for the juices market segments – ANOVA test Cluster 1 (N=46) Cluster 2 (N=27) F Sig.
Important - unimportant* 1,63 3,22 30,797 0,00
Boring - interesting 6,00 4,51 33,671 0,00
Relevant - irrelevant* 1,84 3,74 42,804 0,00
Exciting - unexciting* 2,06 3,70 42,428 0,00
means nothing - means a lot to me 6,26 4,25 92,973 0,00
Appealing - unappealing* 2,15 3,85 41,872 0,00
Fascinating - mundane* 2,58 3,96 34,594 0,00
Worthless - valuable 6,45 4,81 48,242 0,00
Involving - uninvolving* 1,65 3,85 73,205 0,00
not needed - needed 6,56 5,00 51,644 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:
Involved in juices 46 (63,0%)
Cluster 2: Uninvolved in juices 27 (37,0%)
Chi-square
Sign.
N % N % Gender Males: 18 39,1 12 44,4 0,198 0,656 Females: 28 60,9 15 55,6 Age 18 – 25: 2 4,3 4 14,8 7,450 0,281 26 – 30: 20 43,5 9 33,3 31 – 35: 16 34,8 10 37,0 36 – 40: 5 10,9 1 3,7 41 – 45: 3 6,5 1 3,7 46 – 50: 0 0,0 1 3,7 >50: 0 0,0 1 3,7 Education Secondary educ. 2nd level: 1 2,2 2 7,4 4,387 0,356 Post-secondary educ. : 2 4,3 2 7,4 Technical higher educ.: 8 17,4 3 11,1 University: 10 21,7 10 37,0 Postgraduate studies: 25 54,3 10 37,0 Food shopping
Me: 17 37,0 8 29,6 0,406 0,534 Me and another: 29 63,0 19 70,4
Another: 0 0,0 0 0,0
44
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 Cluster 1: Involved in juices Cluster 2: Uninvolved in juices
Attributes Importance Preferred level Part-worth of pref. level
Importance Preferred level Part-worth of pref. level
Visibility 23,96% Transparent 0,749 21,59% Transparent 0,669
Shape 56,77% Curvy/ round (bottle)
1,774 56,82% Curvy/ round (bottle)
1,762
Graphics 17,19% Landscape 0,537 21,59% Landscape 0,669
Colour 2,08% Red 0,065 0,00% Indifferent 0,000
45
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 part-
worth 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 part-
worths 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 Sig. Cluster 1 Cluster 2 Colour Green 5,44 5,48 0,010 0,91
Red 5,48 5,46 0,003 0,96
Shape Square 4,86 5,00 0,090 0,76
Curvy/round 6,05 5,93 0,081 0,77
Visibility Transparent 5,71 5,63 0,028 0,86
Not Transparent 5,21 5,30 0,039 0,84
Graphics Image of the product 5,28 5,29 0,001 0,97
Landscape 5,64 5,64 0,000 0,98
46
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) Cluster 2 (N=4) F Sig.
I prefer to buy natural products, ie products without preservatives.
6,44 2,00 96,491 0,00
To me the naturalness of the food that I buy is an important quality.
6,40 1,75 122,791 0,00
I try to avoid food products with additives. 6,10 3,25 15,259 0,00
47
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.
48
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
49
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.
50
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 package’s 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.
51
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
52
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)
53
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, Giménez, 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. Brunsø, 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.
54
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, Brunsø, K & Bisp, S 1993, Food-related life style: Development of a cross-culturally 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. Lähteenmäki, L, Lampila, P, Grunert, K, Boztug, Y, Ueland, Ø, Aström, A & Martinsdóttir, 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.
55
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 & Lähteenmäki, L 2010, 'Country-wise differences in perception of health-related 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 & Lähteenmäki, L 2009, 'Consumer appeal of nutrition and health claims in three existing product concepts', Appetite, vol. 52, no. 3, pp. 684-92.
56
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 Not healthy at all
Absolutely healthy
Α1 1 2 3 4 5 6 7 8 9 10
Α2 1 2 3 4 5 6 7 8 9 10
Α3 1 2 3 4 5 6 7 8 9 10
Α4 1 2 3 4 5 6 7 8 9 10
Α5 1 2 3 4 5 6 7 8 9 10
Α6 1 2 3 4 5 6 7 8 9 10
Α7 1 2 3 4 5 6 7 8 9 10
Α8 1 2 3 4 5 6 7 8 9 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 1 2 3 4 5 6 7 unimportant
boring 1 2 3 4 5 6 7 interesting
relevant 1 2 3 4 5 6 7 irrelevant
exciting 1 2 3 4 5 6 7 unexciting
means nothing 1 2 3 4 5 6 7 means a lot to me
appealing 1 2 3 4 5 6 7 unappealing
fascinating 1 2 3 4 5 6 7 mundane
worthless 1 2 3 4 5 6 7 valuable
involving 1 2 3 4 5 6 7 uninvolving
not needed 1 2 3 4 5 6 7 needed
57
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 Not healthy at all
Absolutely healthy
X1 1 2 3 4 5 6 7 8 9 10
X2 1 2 3 4 5 6 7 8 9 10
X3 1 2 3 4 5 6 7 8 9 10
X4 1 2 3 4 5 6 7 8 9 10
X5 1 2 3 4 5 6 7 8 9 10
X6 1 2 3 4 5 6 7 8 9 10
X7 1 2 3 4 5 6 7 8 9 10
X8 1 2 3 4 5 6 7 8 9 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 1 2 3 4 5 6 7 unimportant
boring 1 2 3 4 5 6 7 interesting
relevant 1 2 3 4 5 6 7 irrelevant
exciting 1 2 3 4 5 6 7 unexciting
means nothing 1 2 3 4 5 6 7 means a lot to me
appealing 1 2 3 4 5 6 7 unappealing
fascinating 1 2 3 4 5 6 7 mundane
worthless 1 2 3 4 5 6 7 valuable
involving 1 2 3 4 5 6 7 uninvolving
not needed 1 2 3 4 5 6 7 needed
58
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.
1 2 3 4 5 6 7
To me the naturalness of the food that I buy is an important quality.
1 2 3 4 5 6 7
I try to avoid food products with additives. 1 2 3 4 5 6 7
Who is responsible for food shopping inyour household?
Me Someone else
Me and someone else
What is your age?
What is your gender?
Male Female
What is your education level?
No school
Primary school
Secondary educ. 1st level
Secondary educ. 2nd level
Post-secondary education
Technical higher education
University
Postgraduate studies
Thank you for your participation!
59
APPENDIX II: The stimuli cards
60
A1 A2
A3 A4
61
A5 A6
A7 A8
62
X1 X2
X3 X4
63
X5 X6
X7 X8