Make no mismatch about it!

(1)

Make no mismatch

about it!

Influence of colour incongruity of packaging on purchase intentions

Which milk would you choose?

(2)

Make no mismatch about it!

Influence of colour incongruity of packaging on purchase intentions

S.T.G. Legtenberg

University of Groningen

Faculty of Economics and Business

Master Thesis Business Administration - Marketing Management

August 2012

First supervisor: Dr. J. Liu

Second supervisor: Dr. M.C. Non

Kostersgang 22D, 9711 CX, Groningen, The Netherlands

+316 1396 2745

(3)

i | P a g e

Management summary

As a marketer, it is important to understand how to influence consumer behaviour in such a way that sales will increase. There are many marketing tools available, and packaging of products is one of them. The packaging of a product provides consumers with cues about the characteristics of the product. In times of increased offerings in consumer retail products, it is very important for companies to let their products stand out in the crowd. Implementing a certain degree of mismatches in packages is one way to achieve more arousal around the product. This tactic is often used in practise, but does it actually work? This research investigates the influence of mismatches in packaging cues on consumers’ purchase intentions for consumer retail products. The influence of familiarity with a product and the personality trait openness to experience are taken into account as moderators on the main relationship. In order to investigate the relationships between the variables, a survey was constructed (n = 163) which measures consumers’ purchase intention when confronted with manipulations on both the degree of mismatch and the familiarity with the product.

(4)

ii | P a g e

Preface

This master thesis is the final contribution to the completion of my master in Business Administration – Marketing Management. When I finalized my master in Technology Management, I decided that I wanted to complement my technical and economical knowledge with a study in marketing, because I am really interested in consumer behaviour as I discovered in my job as a salesperson at the MediaMarkt. During the master, the effect of marketing tools on consumer behaviour got my special attention. I was pleased to find out about the opportunity to write my thesis on these effects, since Dr. Liu initiated a research group in placebo effects of marketing. I am very grateful for this opportunity, and I would like to give special thanks to Dr. Liu for her advices and support. My interest in the effect of colour on consumer behaviour has led to a focus on a mismatch in colour cues of packaging. In my every day shopping experiences, I noticed that people have many associations about the colour of packaging, and that many people base their shopping decisions on colour heuristics.

I would like to thank certain people who supported me during the process of writing this thesis. Thanks to Eliza Komen and Victorine Marchesini, with whom I took part in the research group on placebo effects of marketing tools. Working together in the library of the University of Groningen was both inspirational and fun. During the finalization process, I got very useful feedback from my second supervisor, Dr. M.C. Non, about the modelling part of this thesis. She made me rethink several analyses I performed, and hence I came up with a more grounded foundation for the performed analysis. Furthermore, I would like to give special thanks to my boyfriend and my friends, who encouraged me during the past few months, and with whom I got to discuss my findings with. Tineke de Vries mentioned the story about Chocolonely to me, and I would like to thank her for that inspirational case. Finally, I would like to thank Humphrey ter Veer for his help on finding SPSS literature, which enabled me to expand my knowledge about modelling. To conclude both this preface and my study at the University in Groningen, I would like to share a quote that has inspired me during the writing of this thesis: “Learning is acquired by reading books; but the much more

necessary learning, the knowledge of the world, is only to be acquired by reading man, and studying all the various editions of them.” - Lord Chesterfield, 1694-1773

(5)

iii | P a g e

Figures

FIGURE 1:SELECTION OF CHOCOLATE OFFERING AT ALBERT HEIJN ... 1

FIGURE 2:CONCEPTUAL MODEL ... 4

FIGURE 3:MAIN RELATIONSHIP ... 30

FIGURE 4:MODERATING EFFECT OF GENDER ON MAIN RELATIONSHIP ... 34

FIGURE 5: POSSIBLE PACKAGES A PARTICIPANT IS EXPOSED TO... 58

Tables

TABLE 1:OVERVIEW OF POSSIBLE RESOLUTIONS TO INCONGRUITY ... 9

TABLE 2:OUTCOME PER CONDITION PRE-TEST ... 20

TABLE 3:PRE-TEST MANIPULATION CHECK MISMATCH ... 21

TABLE 4:PRE-TEST MANIPULATION CHECK FAMILIARITY ... 21

TABLE 5:OVERVIEW OF PARTICIPANT CHARACTERISTICS PER CONDITION ... 26

TABLE 6:ANOVA RESULTS CONDITIONS ... 26

TABLE 7:MAIN QUESTIONNAIRE MANIPULATION CHECK MISMATCH ... 27

TABLE 8:MAIN QUESTIONNAIRE MANIPULATION CHECK FAMILIARITY ... 27

TABLE 9:COMBINING QUESTIONS INTO NEW VARIABLES ... 27

TABLE 10:CONDITIONS IN QUESTIONNAIRE ... 28

TABLE 11:RESULTS REGRESSION ANALYSIS ... 29

TABLE 12:PEARSON CORRELATION ANALYSIS RESULTS FOR DEMOGRAPHICS ... 36

TABLE 13:POSSIBLE CONDITIONS PRE-TEST ... 51

Matrices

MATRIX 1:2 X 2 X 2 MATRIX OF ALL POSSIBLE CONFIGURATIONS ... 18

(9)

1 | P a g e

1 Introduction

How do you respond when your favourite flavour of chocolate has all of a sudden a completely differently coloured package? Most likely, you will be slightly confused and question yourself whether this is still you favourite flavour of chocolate... will you still buy it, or will you look for an alternative? In the latter case, the expected and actual attributes do not match, and therefore, such a situation is called a mismatch in cues (Spence, 2011). This example is based on a real-life case of Chocolonely1, a producer of fair-trade chocolate. The average Dutch consumer would expect the packaging for the flavours dark, milk and white chocolate to be respectively red, blue and white, due to learned associations, see Figure 1. Aslam (2006) describes that the use of colour can result in strong associations to specific products

or product categories. Chocolonely has altered the colours of their chocolate bars: dark chocolate is wrapped in blue foil and milk chocolate in red foil. How do consumers respond to such a mismatch in packaging cues? In the case of Chocolonely, the mismatch in colours confuses consumers, and many consumers returned to the store to

complain about the fact that they accidently bought the ‘wrong’ bar. The purchase decision is an outcome of a cognitive process. Clydesdale (1993) explains that colour, as packaging cue, influences product acceptability, choice and preference due to the fact that certain cues predict the value of the product (Reimer and Kuehn, 2005 and Bhuian, 1997). The package fulfils a placebo effect in a way that it creates certain expectations about what the product will deliver (Shiv, Carmon and Ariely, 2005). Bloch (1995) and Fenko, Shifferstein & Hekkert (2010) explain that purchase decisions are mainly based on the products’ visual appearances, which is the responsibility of the marketing department of the producing company (Underwood, Klein and Burke, 2001).

As several companies have found out the hard way; it truly matters how a package is designed. Wells, Farley and Armstrong (2007) found that 71% of the consumers they interviewed stated that they rely on packaging to aid their decision-making process at the point of purchase. The use of colour in packaging can determine a product’s desirability (Madden, Hewett and Roth, 2000).

1

http://www.tonyschocolonely.com

(10)

2 | P a g e When consumers do not feel the need to purchase your packaging, they will not buy it. This could result in great losses. Knowledge about associations can both reduce costs and increase profits (Keller, 1991): by acting more efficiently on the needs of consumers, a more specific packaging can be designed, which will result in fewer mismatches. As an illustration, Keller (1991) found prove for a 15% increase in sales due to the adaption of new, positively associated colours in a product packaging of Iglo Products. This illustrates the great potential that many firms can still gain in this area of knowledge. More generally, in the field of marketing, the focus shifts from traditional mass media to point-of-sale communications and promotions, of which packaging is a specific element (Underwood, 2003).

(11)

3 | P a g e things and variety; this could indicate that certain people will be more influenced by mismatches than others (Barrick and Mount, 1991, and Digman, 1990). This might result in different purchase intentions (McEachern and Warnaby, 2008).

The objectives of this research are twofold. First of all, the influence of mismatches in packaging cues on purchasing intention of consumers is studied in order to contribute to the body of knowledge in this field of research. Secondly, the concluded contradictions in literature are discussed by evaluating the results of this research. This provides more insight into the conditions under which the previous findings hold.

1.1 Research questions and conceptual model

This section introduces the main research question and the conceptual model for this research. The introduction discussed the main focus of this research, and it described a short literature review on this topic. The focus of this research lies on consumers’ reaction to a mismatch in packaging cues. The point of purchase, measured with purchase intention, is taken into consideration as the dependent variable of this research, since this variable is increasingly important for firms to make a profit. In studying the relationship between a mismatch in cues and the consumer reaction at the point of purchase, two moderators are taken into consideration. The moderators are

familiarity with product and the personality trait openness to experience which both have two

levels, respectivelylow familiar and high familiarity with the product and low and high openness to experience. The reasoning for the selection of these moderators is provided in the next two paragraphs.

Familiarity with product is chosen as a moderator because familiarity with a product influences the way extrinsic and intrinsic product cues are used to assess the quality of a product (Rao and Monroe, 1988). Several conflicting theories exist on this topic. Some research states that when product familiarity increases, the use of extrinsic cues decreases (Kumar and Gaeth, 1991). They mention that familiarity has an influence on how information is used. Rao and Monroe (1988) mention a different theory: they found an inverted u-shape in the relationship, where low-familiar and high-familiar products had a stronger extrinsic cue-quality relationship than the moderate-familiar products.

(12)

4 | P a g e influence purchase intentions.

The main relationship under investigation of this research is examined with the following research question:

How does a mismatch (or incongruity) in colour-packaging cues influence consumers’ purchase intention for retail consumer goods?

In order to investigate the influence of the moderators on the main relation, two sub questions (SQ) are defined:

SQ1: How does familiarity with product influence the relationship between a mismatch in

colour cues of packaging and purchase intention?

SQ2: How does the personality trait openness to experience influence the relationship

between a mismatch in colour cues of packaging and purchase intention?

The corresponding conceptual model for this research is presented in Figure 2.

1.2 Methodology

In order to find answers to the research questions, several hypotheses are generated based on literature study. In order to test the generated hypotheses, a questionnaire is created. The results are analysed with the use of SPSS in order to find possible relationships. This enables the acceptation or rejection of the hypotheses, which results in the answering of the research questions.

Mismatch expected versus actual colour

package Purchase intention Familiarity with product Openness to experience H1 H2 H3

(13)

5 | P a g e

1.3 Managerial and academic relevance

It has become clear in the introduction of this research, that ‘hitting the right snare’ in the mind of consumers when it comes to packaging, is very important for the sale of products. Underwood (2003) describes that when companies have to reduce their marketing budgets, spending on traditional mass marketing is one of the first budgets to be reduced. More focus is placed on sales promotions and on point of sale communications (Underwood, 2003). Since packaging is a very important medium to achieve sales promotion and point of sale communication, its importance becomes clear. Furthermore, the product offering in our modern day society is ever increasing. As a manager, it is very important that consumers see your product, recognize it, and, even more important, that consumers buy your product (Underwood, 2003). The design of packaging has gained increased attention since the 1930s, when manufacturers started to think more creatively and strategically about the external presentation of their products (Nussbaum, 1988). Ampuero and Vila (2006) describe the essence of packaging as being one of the primary tools in the positioning strategy. The package is used to position the product in the minds of consumers. Ampuero and Vila (2006) explain that the positioning of a product induces its marketing mix. Furthermore, they mention that having the right kind of packaging can help in creating a competitive advantage. When understanding the theory of packaging and the reactions and perceptions in customers’ minds, one can achieve a competitive advantage. Moreover, packaging of competition can be analysed and studied in order to optimize one’s own packaging. Furthermore, as the introduction already explained: delivering the right kind of packaging and cues might result in higher sales and lower costs (lower waste). These finding are important for every company, since efficient profits are increasingly important in the current economic crisis. Therefore, it is important for managers to know the influence of mismatches on purchase intentions of consumers, in order to create effective packages that help to boost sales.

With respect to the academic relevance of this research; there is still a great deal of knowledge to be gained in the field of (in)congruity in associations and cues of packaging. There are some conflicting results in the field of consumer reactions to the product under consideration. Despite these conflicting results, many companies assume that a certain amount of incongruity is a good thing to achieve, while this may not be the case in all scenarios (Lee and Schumann, 2004).

(14)

6 | P a g e

1.4 Structure of the paper

(15)

7 | P a g e

2 Theoretical framework

This chapter discusses the main variables of this research. It provides a solid background that enables the creation of several hypotheses, which formulate an answer to the sub questions as proposed in Chapter 1. The chapter starts with the introduction of the independent variable, which is the mismatch of colour cues in packaging. Hereafter, the dependent variable purchase intention is discussed. When these two topics are introduced, the existing relationship between these variables is discussed by the generation of the first hypothesis. Two moderators, which are familiarity with

product and openness to experience are introduced, and with their introduction follow two additional

hypotheses as a first attempt to answer the sub questions.

2.1 Mismatch of colour cues in packaging – independent variable

The conceptual model, presented in Figure 2, shows the independent variable of this research which is the mismatch of colour cues in packaging. This variable can be defined as: the extent to which the expected and actual colour cues of packaging differ, as perceived by a consumer (based on Spence, 2011).

(16)

8 | P a g e This section first discusses the theory behind the concept of incongruity to get an idea on what can be expected when a consumer is confronted with a mismatch, and afterwards, the different views on the effect of incongruity on consumer behaviour are discussed. This section ends with an elaboration on colour associations in general.

2.1.1 How do consumers process mismatches/incongruities

The processing of incongruity by consumers is described by both Petty and Cacioppo in their classical Elaboration Likelihood Model (ELM) (1981 and 1986) and by Mandler’s Schema Incongruity Processing Theory (1982). Knowledge about how consumer process incongruity is useful in explaining the outcome of this research. It helps to create an understanding into why and how purchase intentions are influenced.

The Elaboration Likelihood Model (ELM) (Petty and Cacioppo, 1981 and 1986) describes what processing strategies are in place when one is faced with a specific cue. ELM provides insight into when certain reactions to incongruity will occur. ELM describes that there are two separate tracks to persuasion, which are the ‘central route’ and the ‘peripheral route’. The former makes use of mostly dominant cognitional processes, in which the attention is focussed on the specific process and in which the influence of the cues are perceived consciously. The latter is a process which takes place more unconsciously. So, the consumer decides how he/she processes certain information, although this choice is often not consciously made. A drawback of ELM is the fact that the model does not describe explicitly how information is processed (Lee and Schumann, 2004). It can be concluded that the processing of colour of packaging would normally happen unconsciously, following the peripheral route, however, when faced with a mismatch, it could switch to the more conscious processing in the central route, due to the arousal created.

(17)

9 | P a g e Lee and Schumann (2004) have integrated both classical theories from Petty and Cacioppo’s (1981 and 1986) and Mandler (1982) into one processing model of incongruity in order to deal with the drawbacks of both models. It helps to predict and explain the influence of a mismatch on the dependent variable – purchase intention. When a person is confronted with an incongruity, he/she can decide whether to process the incongruity or not. This choice can occur both conscious and unconscious. When the choice is made not to process the incongruity, the incongruity can still be processed on a peripheral level (unconscious processing). However, the choice of not processing the incongruity can also result in a complete rejection of the incongruity. The next step in the process is concerned with the actual dealing with the incongruity. This can result in two possible outcomes; a successful resolution and an unsuccessful resolution. Possible resolutions are displayed in Table 1, which can result in either a negative or a positive evaluation.

Resolution Explanation

Assimilation This type of resolution puts the incongruity into an existing schema. Assimilation will most likely only take place when the incongruity is small or low (Sujan and Bettman, 1989).

Alternative schema

When the incongruity cannot be placed into the expected or referred existing schema, an alternative schema may be used. This resolution is concerned with the use of new information retrieved from the incongruity into a (new or existing) alternative schema. This results in self-directed learning (Gregan-Paxton and John, 1997).

Accommodation When the incongruity cannot be placed under the referred schema, or when it cannot be linked or combined with an existing schema, a new schema is required to be formed. This process requires most cognitive energy. Mandler (1982) describes that the creation of a new schema is the accommodation of severe incongruities.

Table 1: Overview of possible resolutions to incongruity

Since colour mismatches are not severe mismatches, it is most likely that either assimilation or alternative schema will be the outcome of the processing of the moderate mismatch.

The type of resolution, together with several moderating influences such as situational influences, individual influences and source- and message influences, determine whether the consumer positively or negatively evaluates the incongruity. This influences whether the consumer will purchase the product or whether the consumer will pick an alternative product. Several moderators on this relationship have yet been investigated, and are shortly discussed here.

Processing time - The amount of time a consumer has influences the way in which incongruity

(18)

10 | P a g e

Perceived risk - Campbell and Goodstein (2001) found that certain perceived risks (social risk

in particular) influence the willingness to deal with incongruities. Incongruent cues which are concerned with high risks are less likely to be evaluated positively than incongruent cues with lower risks.

Mood - Mood can greatly influence how individuals make decisions. Happiness might

negatively influence the process of serious consideration (Asuncion and Lam, 1995), but sadness actually positively influences this process (Mackie, Asuncion and Rosselli, 1992). Furthermore, Asunction and Lam (1995) find that incongruent stimuli can best be processed when a person is in a neutral mode.

Ability to process novelties - This is a personality trait which has been described in many

different theories. Some theories call it the trait of ‘novelty and sensation seeking’ (Zuckerman, 1988) or the trait of rigidity or dogmatism (the avoidance of trying new things) (Peracchio and Tybout, 1996).

Knowledge of product - On this topic, several assumptions exist. Sujan (1985) mentions that

individuals with more knowledge of a product are expected to be more likely to process to incongruent information. This topic is further investigated as a moderator, see section 2.3.

For this research, different, unstudied conditions are selected. Their influence on the main relationship, the impact of incongruity on purchase intentions, is investigated.

2.1.2 Conflicting possible outcomes incongruity on consumer behaviour

Section 2.1.1 has discussed how incongruity is created, and it explained how incongruity is processed by consumers. This section describes into detail the conflicting findings that are present when looking at the theory about consumer behaviour on incongruent cues. Two streams of literature try to predict the relationship of mismatches on purchase intention, but these two streams tend to heavily conflict in their findings.

(19)

11 | P a g e cognitive dissonance, which describes that severe incongruity results in disturbance of the individual.

2.1.2.1 Congruent cues as best predictor of positive evaluations

Sengupta, Goodstein and Boninger (1997) explain that congruent cues deliver more favourable responses than incongruent cues. Furthermore, Peracchio and Tybout (1996) found that congruity in cues results in faster reaction times. Both Hastie and Kumar (1979) and Sengupta et al. (1997) explain that processing congruent information results in better memorizing the information involved. Van Rompay and Pruyn (2011) describe that congruity of different stimuli (or product elements) increases the processing fluency. Processing fluency is concerned with the way the information gained by people is transferred into understandable data. An increase in processing fluency results in a more positive evaluation, since the individual does not feel any tension. Also Lee & Labroo (2004) explain that when stimuli are easy to process, the resulting evaluation is more positive. Sherif, Sherif and Nebergall (1965) mention in their social judgment theory that people reject incongruent information that lies outside their range of acceptance.

Following this stream of literature, it is expected that purchase intentions will be highest when there is no mismatch at all in colour packaging cues.

2.1.2.2 Incongruent cues as best predictor of positive evaluations

Lee and Thorson (2008) mention that a certain amount of incongruity produces more favourable responses to an advertisement than complete matches or severe mismatches. Numerous authors agree on this point of view. Mandler (1982) explains that a certain amount of incongruity is arousing; it heightens processing and produces enjoyment of novelty. Heider (1958) found that incongruity can result in a specific tension, and Latour and Tanner (2003) mention that this arousal might result in positive reactions. Also Goodstein (1993), Hastie and Kumar (1979) and Meyers-Levy and Tybout (1989) describe that the incongruity-arousal might result in positive evaluations. Following this stream of literature, it is expected that purchase intentions will be highest when there is a moderate mismatch in colour packaging cues.

2.1.3 Colour associations

(20)

12 | P a g e several products or product categories (Aslam, 2006). Colour is interpreted differently in each country, as Grossman and Wisenblit (1999) explain, which is probably due to the associations that are linked to specific colour. Williams (1996) reports that, for example, red lingerie sells well in Spain if it is dyed flamenco red. Another example comes from Scandinavia, where lilac is associated with mourning, so underwear in this colour does not sell very well. Grossman and Wisenblit (1999) mention that red stands for extroversion, blue-green for intelligence, sociability and narcissism. Black can be associated with poison. White can be seen as a colour that indicates accuracy and scientific proof, but it is also seen as plain, emotionless and unexciting (Beatty, 1997).

Since this research is performed only in the Netherlands, the results should be used with precaution not to misuse the results in other cultures without taking differences in colour associations into consideration. Grossman and Wisenblit (1999) explain that, regardless of the existing associations that colours might have, other associations of consumers can interfere with a marketer’s intent for the use of a specific colour. Crowley (1993) also explains this dual nature of colour, as he expresses this in both an arousing component and an evaluative component.

2.1.3.1 Colour in packaging

Packaging is the container which holds, protects, preserves and identifies the product as well as facilitates handling and commercialisation (based on Vidales Giovannetti, 1995, translated by Ampuero and Vila, 2006). Peters (1994) makes a bold assumption by stating that packaging is the most important communicator because it reaches all consumers in the category, it is present at the crucial moment of point of sale, and consumers are very involved when they consume the information on the packaging. The expression of Vidales Giovannetti (1995) describes the role of packaging design really well: packaging can be seen as a ‘silent salesman’, since the packaging helps (or counteracts) to deliver the desired message to the consumer.

2.1.4 Summary mismatch of colour cues in packaging

To wrap up this section about the independent variable mismatch of colour cues in packaging, several conclusions can be drawn:

 Colour associations can be created for specific product or product categories, and this can cause specific choices for products;

(21)

13 | P a g e and images);

 The colour of packaging-cue is processed on the peripheral route, since colour tends to influence the consumer unconsciously, however, mismatches might be processed in the central route due to enlarged arousal;

 The evaluation of the incongruity influences the consumer behaviour and thereby might influence the purchase intentions. Conflicting streams in literature predict either an increase in sales (due to positive evaluations) for the presence of a mismatch or for an absence of mismatches.

2.2 Purchase intention – dependent variable

As can be seen in the conceptual model, presented in Figure 2, the dependent variable of this research is purchase intention. Laroche, Kim and Zhou (1996) describe the purchase intention simply as the intention to buy. Hardly any definitions are given on this concept, and this might be due to the fact that it is a commonly known concept.

Chang and Wildt (1994) found that perceived value positively influences the purchase intention. Perceived value is the outcome of the assessment of perceived price and perceived quality. Besides the indirect influence of perceived price and quality through perceived value on purchase intentions, there is also a direct influence of the two concepts on purchase intention, as Chang and Wildt (1994) found. In their directions for further research, they advise to dive deeper into influences on purchase intentions, as is the case with this research. Homburg, Koschate and Hoyer (2005) describe that the purchase intention depends on the level of satisfaction a customer expects from a purchase. The research of Homburg et al. (2005) provides evidence for the claim that consumers are willing to pay more for better expected quality, which is in line with expectations of Shiv et al. (2005).

2.2.1 Summary of dependent and independent variable – creation of H1

(22)

14 | P a g e not meet actual cues and this might result in dissatisfaction. The expected relationship which is investigated is presented in hypothesis 1 (H1). This hypothesis tends to answer the main research

question.

H1: A mismatch in colour-packaging cues results in a lower purchase intention for retail

consumer goods, compared to no mismatch in colour-packaging cues.

2.3 Product familiarity – moderator

Rao and Kent (1988) describe the use of ‘familiarity with product’ as a possible moderator on the relationship incongruity – consumer behaviour. The levels of familiarity can be measured on a continuous scale, however, for this research, only two levels of familiarity are taken into account, since these tend to be most interesting from an academic point of view, due to some conflicting results with these levels (see Chapter 1). These levels are low familiarity with product and high

familiarity with product.

Rao and Kent (1988) concluded that the use of extrinsic- and intrinsic product cues for the assessment of product quality depends on prior knowledge. They found that the more product familiarity increases, the stronger the use of intrinsic cues became for product quality assessments. Furthermore, they describe that the use of extrinsic product cues, such as price, increases when the familiarity with the product is low. So, for a product with low familiarity, extrinsic product cues will have a great influence. When familiarity is high, extrinsic product cues will not have such a great influence, following the reasoning of Rao and Kent (1988).Another advocate of this line of reasoning is Sujan (1985). He mentions that individuals with more knowledge of a product are expected to be more likely to process the incongruent information. Kumar and Gaeth (1991) mention as well that product familiarity has an influence on how information is used. Their research focuses on whether the order of presenting information has an influence on product choice, and whether the familiarity with the product category would influence this relationship. They found an absence of order effects for familiar product categories, but a presence of the order effects for unfamiliar product categories. From this finding, it can be theorized that a consumer, when faced with an unfamiliar product, uses the information provided more intense than when confronted with a familiar product.

(23)

15 | P a g e purchase intention, since the ‘impulse’ for the purchase is no longer available. Wells et al. (2007) found that impulse buying of consumers is often based on extrinsic attributes of the packaging. In their research, it showed that 30% of consumers indicated that familiarity with the product was important when buying on impulse. Impulse buying often happens on the peripheral route, were minimal cognitive effort is involved (see section 2.1.1). The increased awareness caused by a mismatch might cause consumers to rethink the purchase and it might stop them from automatically selecting the product, despite the high familiarity with the product.

It is important to note that familiarity is not the same construct as involvement. Since involvement focuses more on the connection of a consumer with a specific product, it is in essence different than the concept of familiarity, which defines how well the consumer is acquainted with the product. Research on involvement as a moderator is yet available, as partly discussed in section 2.2.

2.3.1 Summary and hypothesis

The findings about the moderator familiarity can be summarized as: in general, it seems that high familiarity with a product decreases the use of extrinsic product cues. But, it might be the case that when a mismatch occurs, the impulsive buying tendency is disrupted because consumers might notice a difference in the packaging. This might result in higher response timeand the processing of information is likely to occur no longer on the peripheral route (see section 2.1.1). More effort is needed to process the incongruity, and this might result in dissatisfaction with the experienced cue, which might result in a lowered purchase intention.

Taking all the existing literature on familiarity with product and the knowledge about consumer processing of information into account, the following hypothesis (H2) can be created as a

first attempt to answer sub question 1 (SQ1):

H2: Familiarity with the product negatively influences the relationship between the mismatch

and purchase intention, thus increasing the negative effect of a mismatch on purchase intentions.

2.4 Openness to experience – moderator

(24)

16 | P a g e Warnaby, 2008). Therefore, this research tries to fill part of this gap by investigating one of the personality traits on consumer behaviour.

All possible personality traits are analysed, categorized and summarized in the Big Five-model (Goldberg, 1971). This model is seen as the most important model in this area of research (Borghans, Duckworth, Heckman and Ter Weel, 2008). The five personality factors are summarized with the acronym “OCEAN”. This acronym stands for Openness to Experience, Conscientiousness, Extraversion,

Agreeableness and Neuroticism. For this research, the trait Openness to Experience is selected as a

moderator on the relationship mismatch – purchase intention.

Openness to experience can be described as “the degree to which a person needs intellectual stimulation, change and variety” (Borghans et al., 2008). Atkinson, Atkinson, Smith, Bem and Nolen-Hoeksema (2000) describe openness to experience as a contradiction on the following scales: inventive and curious versus consistent and cautious. This scale implies that people scoring low on openness to experience are more conventional and traditional in their thinking and behaviour. They prefer routine in their actions, and will go for the most familiar options. Flynn (2005) explains that this might be due to the fear for possible uncertainty. Choosing the most familiar option results in lower uncertainty and this might predict a certain expected preferred quality. Barrick and Mount (1991) and Digman (1990) describe that individuals scoring high on openness are more curious and broad-minded than individuals scoring low on this aspect.

In these definitions, it becomes clear that the need for change and variety can be linked to the possible mismatch, which in definition, is a change as well as a variety from the original product/package. Therefore, it would be interesting to see whether this personality trait will influence the purchase intention when persons are confronted with a mismatch. McEachern and Warnaby (2008) invested how personality traits, combined with knowledge, influence purchase intentions. They found that openness to experience influences purchase intentions in a way that was moderated by knowledge about a product.

(25)

17 | P a g e important for an organization to determine what personality traits their customers have, and especially, how their customers score on the trait openness to experience.

Despite the intuitive reactions that this personality trait might induce, there is still very limited research about the practical implications of the trait openness to experience, as Lepine, Colquitt and Erez (2000) mention. Some research has shown that openness to experience is positively related to higher online purchase intentions (Bosnjak, Galesic and Tuten, 2007). This research contributes to this gap in literature by providing a more practical view on the trait of openness to experience with respect to purchase intentions when confronted with a mismatch.

2.4.1 Summary and hypothesis

It has become clear that the personality trait openness to experience influences individual behaviour. When a person is classified as not open to new experiences, or dislikes variety, it is expected that an encounter with a mismatch results in negative evaluation towards the product (McEachern and Warnaby, 2008). This negative evaluation can lower the purchase intention of a consumer.

The theoretical background helps to formulate the third hypothesis (H3) as an answer to sub

question 2 (SQ2):

H3: Openness to experience as a personality trait positively moderates the relationship

(26)

18 | P a g e

3 Methodology

This chapter describes the specific methodology used in this research in order to provide data to answer the research questions. The research design is discussed and the data collection and the plan of analysis are explained.

3.1 Research design

For this research, a 2 (no mismatch – mismatch) by 2 (low familiarity – high familiarity) by 2 (low openness – high openness) between-subjects design is used. The level of the mismatch and the amount of familiarity of a product are manipulated, and are measured with manipulation checks. Matrix 1 presents the possible conditions for the participants.

Low familiarity High familiarity

Low openness High openness Low openness High openness No mismatch Condition 1 Condition 5 Condition 3 Condition 7

Mismatch Condition 2 Condition 6 Condition 4 Condition 8

Matrix 1: 2 x 2 x 2 matrix of all possible configurations

A survey is held to gain information about consumer behaviour when confronted with a mismatch in packaging. The survey is presented online by using Qualtrics.com, an online software program licensed by the University of Groningen which enables students to build professional questionnaires. The use of a questionnaire results in the generation of data which can be analysed in SPSS in order to find possible relationships among variables.

(27)

19 | P a g e

Low familiarity High familiarity

No mismatch Condition 1 Condition 3

Mismatch Condition 2 Condition 4

Matrix 2: 2 x 2 matrix of manipulating conditions

3.1.1 Manipulated packages

In order to be able to test the reactions of consumers to packages with a mismatch, an example is needed to trigger these reactions. Therefore, several packages are created that contain a mismatch, and that vary on the degree of familiarity to participants. Four packages are designed, one package per condition. The product ‘organic milk’ is chosen as the product under investigation, due to the fact that this product category has highly similar packages with a lot of green elements in them. Furthermore, this product category provides possibilities for the required manipulation in familiarity. Normal organic milk can be classified as highly familiar to most participants, while organic baby milk might not be so familiar to most participants, because it is expected that the group of participants contains a high number of students (due to snowball sampling) who do not yet have babies themselves.

The packages are designed in Photoshop Adobe, and they are based on existing milk packages in order to make them as realistic as possible. A new brand is created in order not to bias the results. The background in the ‘no mismatch’ condition is green, and the background in the ‘mismatch’ condition is black, since this colour is very unfamiliar for organic products, and it might predict very unhealthy associations (Grossman and Wisenblit, 1999). The pre-test and main questionnaire both ensure whether black is an adequate colour for the mismatch condition.

The created packages can be found in Figure 5.

3.2 Pre-test

(28)

20 | P a g e

Condition 1 Condition 2 Condition 3 Condition 4

Familiarity High High Low Low

Mismatch Low High Low High

N (total = 74) 18 18 18 20 Buying frequency Never – Frequently (1 – 7) Mean Std Deviation 2,17 1,098 3,56 2,640 1,44 1,149 1,65 1,565 Familiarity Low – High (1 – 5) Mean Std Deviation 3,06 2,363 3,61 3,256 1,83 1,654 2,20 2,262 Congruity I Mismatch – Match (1 – 7) Mean Std Deviation 5,22 1,215 4,06 2,045 4,67 1,572 4,15 1,814 Congruity II

Bad fit - Good fit (1 – 7) Mean Std Deviation 4,89 1,132 3,94 1,924 4,89 1,323 3,95 1,905

Table 2: Outcome per condition pre-test

3.2.1 Reliability check

The first step in analysing the data is the investigation whether the questions about the experienced amount of incongruity presented in the packaging (question 3 and 4), can be combined into a new variable. Therefore, Cronbach’s Alpha is determined, and its value is 0.898 (see Appendix 2), which is sufficient to combine both questions. A new variable is calculated, and named “Congruity”.

3.2.2 Manipulation check

The next step in the process to determine if the created packages results in sufficient manipulations, consists of the execution of independent t-tests to check for significance in differentiating groups. Appendix 2 presents the SPSS output from both t-tests.

(29)

21 | P a g e

Condition Mean Std. Deviation Sig. (2-tailed) Mismatch no 4,9167 1,21008 0,013

Mismatch yes 4,0000 1,82698 0,013

Table 3: Pre-test manipulation check mismatch

Table 4 shows that also the manipulation of familiarity succeeded. The mean of the low familiar condition (M = 2.03) significantly differed from the mean score in the high familiarity condition (M = 3.33) (t (62.5) = -2.297, p < .05). The Levene’s test is significant, and therefore the option under “Equal variances not assumed” is read.

Condition Mean Std. Deviation Sig. (2-tailed) Familiarity low 2,03 1,979 0,025

Familiarity high 3,33 2,818 0,025

Table 4: Pre-test manipulation check familiarity

With the outcomes of the pre-test, the final questionnaire is constructed, which can be found in Appendix 3.

3.3 Participants

Shopping for groceries is something that many people do. Differences in ages, gender and income level are not expected to have an influence on whether or not people do their groceries. Since all products have individual product packages nowadays, all products can be included into this experiment. In order to get a representative picture of the consumer behaviour per condition, at least 40 respondents are required per condition. This means that at least 160 (4 * 40) respondents are necessary in order to be able to process the results of the questionnaire.

3.4 Variables

The variables included in this research are introduced in Chapter 2. This section provides more information about how the variables are manipulated and tested in the main questionnaire, as presented in Appendix 3.

3.4.1 Independent variable: level of incongruity

(30)

22 | P a g e the designed packages.

To check again whether the manipulation of the mismatch worked, just as in the pre-test, participants are asked whether they experienced a certain mismatch or ‘bad fit’ in the package between colours and product. This question is based on Keller and Aaker (1992). To check again for the appropriateness of green as the matching colour, and black as the mismatching colour, participants are asked to score several colours on their appropriateness for an organic milk product. Furthermore, the participants are asked whether they are colour blind, in order to check whether the colour-manipulation as a match/mismatch could have worked.

3.4.2 Dependent variable: purchase intention

The influence of the independent variable on the dependent variable ‘purchase intention’ is investigated. In order to make sure that participants understand the manipulated situation, a small instruction is given at the beginning of the questionnaire which states that participants should imagine themselves looking for an organic milk product. Next up, participants are asked to indicate their purchase intention after seeing the package, based on research of Dodds, Monroe and Grewal (1991). The purchase intention is double checked, by asking participants to fill out the likelihood of buying the product (adopted from Baker and Churchill, 1977). People might not be paying full attention when filling in a question, or they might not understand a certain question and therefore accidently filling out the wrong answer. Asking multiple questions on the same topic can solve this issue.

3.4.3 Moderator 1: Familiarity

The first moderator under investigation on the main relationship is familiarity with the product. Since the participants are mainly students, the products ‘milk’ and ‘baby milk’ are selected, as explained in section 3.1.1. Average students drink milk every day, but they often do not have children yet, and therefore, baby milk is less familiar to them.

(31)

23 | P a g e whether the participant has at least a minimal level of familiarity. The participant is asked where he/she usually does their groceries, in order to check the familiarity with the product in another way. Budget supermarkets tend to carry different brands than high quality supermarkets. And finally, there is the possibility that participants never do groceries their selves, which might also indicate a non-familiarity with the product packages.

3.4.4 Moderator 2: Openness to experience

The second moderator that is investigated on its influence on the main relationship is openness to experience. Participants are asked to fill in questions about their personality, thus presenting self-reported personality traits. Since it is easier for participants to use their natural personality traits, rather than manipulated personality traits, the openness to experience is not manipulated. Furthermore, no literature was found where personality traits were actively manipulated.

In order to find out how a participant scores on the variable ‘openness to experience’, participants are asked to comment on several statements. These are based on the research of Steenkamp and Baumgartner (1992) into change seeking behaviour. The amount of change seeking behaviour is an indicator of the openness to experience. High change seeking behaviour indicates high openness to experience. The statements of Steenkamp and Baumgartner (1992) are complemented with questions about participants’ brand preferences, switching intentions and variety seeking (based on Raju, 1980). People with higher switching intentions and higher variety seeking traits tend to have a higher openness to experience, as Raju (1980) explains.

3.4.5 Evaluation questions to gain information about participants

Several demographic questions are asked as well besides the specific questions about each variable. These help to clarify the outcomes of the purchasing behaviour. Participants are asked to fill out their gender and age. It might be the case that certain age groups have a higher familiarity with specific products, or that certain gender differences exist.

(32)

24 | P a g e

3.5 Procedure

The questionnaire is constructed online at Qualtrics.com. With a specific hyperlink, participants are invited by e-mail or through a Facebook message. The software of Qualtrics.com enabled the randomization of the questionnaire, thus resulting in an even distribution of conditions among participants.

3.6 Analysis

(33)

25 | P a g e

4 Results: analyses and discussion empirical data

This chapter presents the main results of this research. The data generated with the questionnaire is used in order to test the hypotheses. The representativeness, reliability and validity are discussed as well. Huizingh (2006) advises the following routine when collecting data for a survey: collecting information from surveys, creating a data overview, checking data, editing data, analysing data, interpreting results of analysis and finally: the creating of the report. This sequence of steps is used during this thesis.

4.1 Sample characteristics

Qualtrics.com was used to present the questionnaire online to respondents. Snowball sampling online resulted in a total number of 164 participants, which is more than the required 160 participants (see section 3.3). All participants successfully completed the questionnaire; no questions were left blank due to the forced answering option implemented in the questionnaire. One respondent was found to be severely colour blind, and therefore, this participation was deleted from the dataset. This resulted in a total number of 163 participants (n=163) for this research.

A series of demographic questions provides information about the sample population. A small overview is presented in this section. All corresponding SPSS output can be found in Appendix 4. Of the total sample, 46% of the participants is male, and 54% is female. Most participants are between 21 and 40 years old (74.8%), which can be explained by the snowball sampling among students. Students have an active network of other students, and therefore, it is very likely that the recruitment among students resulted in such a large group of participants in the age group of 21 – 40 years. However, in the Dutch society, the student population (HBO & WO) is only 3.9% of the entire Dutch population (658,600/16,733,7272). Therefore, the representativeness could be increased by including more differentiated people in the research. Furthermore, 96.9% of all participants have lived in the Netherlands for more than five years thus resulting in a high expected knowledge about the Dutch food market. Most participants get their groceries at a high quality supermarket (84%).

In order to check whether these demographics are random assigned, the following analysis is performed. The variables gender and age are included, since these variables might have some influence on the possible answers provided by participants. Table 5, which is based on Appendix 5, section 5.1, shows an overview of characteristics of participants per condition.

2

(34)

26 | P a g e

Condition 1 Condition2 Condition 3 Condition 4

Male 21 50% Male 21 51% Male 16 40% Male 17 43%

Female 21 50% Female 20 49% Female 24 60% Female 23 58%

Total 42 100% Total 41 100% Total 40 100% Total 40 100%

0 - 20 years 8 19% 0 - 20 years 8 20% 0 - 20 years 7 18% 0 - 20 years 5 13%

61+ years 0 0% 61+ years 0 0% 61+ years 2 5% 61+ years 1 3%

Total 42 100% Total 41 100% Total 40 100% Total 40 100%

Table 5: Overview of participant characteristics per condition

To check whether the differences in random assignment are significant in the various conditions, a one-way ANOVA test is performed, see section 5.2 in Appendix 5. Besides gender and age, also the variable ‘Openness to Experience’ is included, to check whether there are major differences of this personality trait in participants per condition. The confidence interval is set at 95%, thereby, the p-value has to be higher van .05 in order to achieve at random assignment for a variable.

Age Gender Openness_Experience

Sig. 0,304 0,689 0,281

Table 6: ANOVA results conditions

With an ANOVA test, the H0 claims that there is no significant difference between the means

of the groups. Table 6 shows that Age (F (3) = 1.222, p > 0.05), Gender (F (3) = 0.491, p > 0.05) and Openness_Experience (F (3) = 1.286, p > 0.05) are equally distributed within the conditions since their p-value is > .05, thus resulting in an acceptation of every H0.

4.1.1 Manipulation checks

Besides the manipulation check that is performed in the pre-test, it is important to perform a manipulation check on the main questionnaire as well to check whether the created manipulations are perceived in the right way by participants. The images of the packages used are manipulated on two different aspects: familiarity and congruity. Each of these constructs is analysed. A one-way ANOVA test enables the manipulation check. The SPSS output is shown in section 5.3 in Appendix 5.

4.1.1.1 Mismatch manipulation check

(35)

27 | P a g e

Condition Mean Std. Deviation Sig.

Mismatch no 4,9939 1,53859 0,000

Mismatch yes 3,6296 1,95381

Table 7: Main questionnaire manipulation check mismatch

4.1.1.2 Familiarity manipulation check

Table 8 shows the manipulation of familiarity. It demonstrates that both groups differ significantly when looking at their means (M = 2.83) and (M = 4.51) (F (1, 162) = 50.629, p < .05). Thus, it can be concluded that the familiarity manipulation is successful.

Condition Mean Std. Deviation Sig.

Familiarity low 2,8313 1,56423 0,000

Mismatch yes 4,5124 1,44733

Table 8: Main questionnaire manipulation check familiarity

4.2 Reliability - aggregated questions

In order to check whether several questions about the same topic did actually deliver the same answers, Cronbach’s Alpha is calculated. Hair, Black, Babin and Anderson (2010) describe that Cronbach’s Alpha has to be at least 0.60 for summated scales to be reliable. Table 9 shows the combined questions and their corresponding Cronbach’s Alpha. The SPSS output of these tests is shown in Appendix 6.

Construct Questions Cronbach’s Alpha

Purchase Intention Q7_1, Q7_2, Q7_3 and Q8_1, Q82 0,899

Openness to Experience Q1_1 - Q1_8 0,605

Familiarity Q10 & 11 0,862

Congruity Q12A 0,913

Table 9: Combining questions into new variables

The Cronbach’s Alpha is 0.899 for the questions about purchase intention, which is over 0.60. Therefore, these questions are taken together into a new variable which is called ‘Purchase_Intention’.

(36)

28 | P a g e The first eight questions about the trait Openness to experience have a Cronbach’s Alpha of 0.605. Excluding certain variables did not result in a higher Cronbach’s Alpha (see section 6.2 in Appendix 6), and therefore, all variables are used. A new variable is created and named “Openness_Experience”.

Questions 10 and 11 aim to measure the familiarity of the participant with the product. The Cronbach’s Alpha for these two variables is 0.862, which is sufficient to combine the variables into a new variable called “Familiarity”. Adding the question about buying frequency resulted in a lower Cronbach’s Alpha, and this question is therefore excluded (see section 6.2 in Appendix 6).

Question 12, which has two sub questions, measures the amount of mismatch experienced by the participant. These two questions have a Cronbach’s Alpha of 0.913. The new variable is called “Congruity” (see section 6.4 in Appendix 6).

Now that both moderators and the independent variable are transformed into single variables, the relationships between the variables can be studied. Section 4.3 discusses the main results of the questionnaire by using a regression analysis to test the possible relationships.

4.3 Regression analysis: testing hypotheses

This section determines whether the manipulations resulted in either confirmation or rejection of the hypotheses. Section 7.1 in Appendix 7 shows the mean purchase intention per condition. The main outcomes are presented in Table 10.

Condition 1 Condition 2 Condition 3 Condition 4

Mismatch Low High Low High

Familiarity Low Low High High

Mean Purchase intention 3,9000 3,5366 4,0500 3,6500

Std. Deviation 1,1970 0,9864 1,2159 1,2896

Table 10: Conditions in questionnaire

In section 7.2 in Appendix 7, the regression analysis is presented. The results are discussed in this section. The main relationship of the influence of a mismatch on purchase intention is investigated, as well as the influence of both moderators openness to experience and familiarity. The confidence interval is set at 95%, i.e. in order to be significant, the condition p < .05 has to be fulfilled.

(37)

29 | P a g e Mismatch is coded as “0 = No mismatch” and “1 = Mismatch”. Furthermore, the variable Familiarity is coded as “0 = Low familiarity” and “1 = High familiarity”. Before the regression analysis can be performed, it is necessary to perform multiple transitions to the data. The variables that made use of a Likert scale are centered. Centering helps to reduce possible multicollinearity, which can occur due to the fact that two moderators (interaction variables) are included. Also the variable Openness_Experience is centered. Next, interaction variables are created to measure the possible moderating effect of the variables Openness to experience and Familiarity:

 Interaction Mismatch * Familiarity

 Interaction Mismatch * Openness to experience

These interaction variables enable the testing of possible influence of the moderators on the main effect. The regression equation for this research is:

: purchase intention : intercept : mismatch : familiarity : openness to experience

: interaction Mismatch * Familiarity

: interaction Mismatch * Openness to experience

: error term : respondent i

Table 11 shows the results of the regression analysis.

Variable Corresponding hypothesis B t Sig.

Constant 3,886 21,639 ,000 Main effect H1 -,392 -1,530 ,128 Moderator Familiarity (interaction) H2 -,057 -,154 ,878 Moderator Openness (interaction) H3 ,060 ,241 ,810

Table 11: Results regression analysis

(38)

30 | P a g e

4.3.1 Main effect: influence of mismatch on purchase intention

Hypothesis 1 states that the presence of a mismatch in packaging cues results in a lowered purchase intention, compared to the absence of a mismatch. The regression analysis, as shown in Appendix 7.2, presents the results of the analysis. The influence of mismatch on purchase intention is not significant (p > .05), but the direction of the relationship is correctly predicted. The coefficient of influence is negative (B = -.392), indicating the following relationship: an increase in mismatch negatively influences the purchase intention, thus reducing the purchase intention. Figure 3 shows the main relationship of the presence of a mismatch on the purchase intention graphically.

Despite the correctly predicted negative relationship that was found in the regression analysis, the non-significance of the variable in the regression analysis results in the rejection of H1.

Figure 3: Main relationship

4.3.2 Moderator Familiarity

The second hypothesis discusses the influence of the familiarity of the participant with the product on the main relationship. It is expected that higher familiarity with the product negatively influences the main relationship between a mismatch and purchase intention. The influence of the moderator familiarity on the main relationship is not significant (p > .05), but the influence is negative (B = -.057) which indicates that the directional influence of the hypothesis is correctly predicted.

To conclude, H2 cannot be supported, although the directional influence on the dependent

(39)

31 | P a g e

4.3.3 Moderator Openness to Experience

The influence of the personality trait openness to experience on the main relationship is predicted in hypothesis 3. Higher openness to experience of consumers is expected to have a positive impact on the main relationship, thus positively influencing the purchase intention when confronted with a mismatch.

The influence of the moderator is not significant (p > .05). However, the direction of the coefficient is positive (B = .060), thus the direction of the influence on the main effect is correctly presented in the hypothesis. It can be concluded that H3 has to be rejected due to the insignificance

of the results, despite the correct direction of the relationship.

4.4 Explaining non-significance hypotheses

In this section, explanations are given for the non-significance of the hypotheses.

4.4.1 Random assignment

As section 4.1 showed, there are no major differences in gender per condition. Also the variation in age of participants per group did not differ significantly. The personality trait openness to experience did not differ significantly per condition as well. Therefore, it is very unlikely that these variables have influenced the results of the regression analysis.

4.4.2 Manipulation successfulness

As the manipulation checks showed in section 4.1.1, both the manipulation of familiarity and congruity succeeded. However, it could be argued that the difference between the manipulated conditions should be even larger to evoke more ‘extreme’ scores. This could benefit the outcome in a way that influences of mismatches on participants’ purchase intentions become clearer. Participants often indicated that they were not really familiar with organic products, but they did actually were familiar with milk itself. Therefore, the organic aspect could have confused participants about their familiarity with the product.

Furthermore, it could be the case that the chosen products (organic milk and organic baby milk) are more familiar to a certain gender type, or that there are differences in liking of these products per gender.

Make no mismatch about it!

Make no mismatch

about it!

Influence of colour incongruity of packaging on purchase intentions

Which milk would you choose?

Make no mismatch about it!

Influence of colour incongruity of packaging on purchase intentions

S.T.G. Legtenberg

University of Groningen

Faculty of Economics and Business

Master Thesis Business Administration - Marketing Management

August 2012

First supervisor: Dr. J. Liu

Second supervisor: Dr. M.C. Non

Kostersgang 22D, 9711 CX, Groningen, The Netherlands

+316 1396 2745

Management summary

Preface

Table of contents

Figures

Tables

Matrices

1 Introduction

1.1 Research questions and conceptual model

1.2 Methodology

1.3 Managerial and academic relevance

1.4 Structure of the paper

2 Theoretical framework

2.1 Mismatch of colour cues in packaging – independent variable

2.2 Purchase intention – dependent variable

2.3 Product familiarity – moderator

2.4 Openness to experience – moderator

3 Methodology

3.1 Research design

3.2 Pre-test

3.3 Participants

3.4 Variables

3.5 Procedure

3.6 Analysis

4 Results: analyses and discussion empirical data

4.1 Sample characteristics

4.2 Reliability - aggregated questions

4.3 Regression analysis: testing hypotheses

4.4 Explaining non-significance hypotheses